Neither the Texas Transportation Institute (TTI), the National CooperativeHighway Research Board, nor anyone acting on behalf of any of the foregoingnamed entities, assume any liability with respect to the use of (or liability fordamages resulting from the use of) any data, information, apparatus, method, orprocess disclosed in this document and its supporting spreadsheet calculations.The foregoing named entities (individually and collectively) expressly disclaim allwarranties, including the implied warranties of merchantability, fitness for aparticular purpose, and good and workmanlike services. Further, TTI does notauthorize any person to create a warranty on TTI's behalf.

DISCLAIMER OF WARRANTY

The electronic version of this document includes linksto several Excel spreadsheets. Each spreadsheetautomates the calculations for the associated figure ortable. Figures and tables that are automated in thismanner are identified in the document with textfor the table or figure title. Click on this text toimplement the associated spreadsheet. Excel 2000or later is required.

©

©

green

INTERACTIVE FEATURES

The electronic version of this document was developed by: Mr. Ivan Lorenz and Mr. Daniyar Satarov, Texas Transportation Institute, College Station, Texas

SPINE = 5/16"Job No._NCHRP#457 NCHRP Green

TRANSPORTATION RESEARCH BOARD

Evaluating IntersectionImprovements:

An Engineering Study Guide

NATIONALCOOPERATIVE HIGHWAYRESEARCH PROGRAMNCHRP

REPORT 457

NATIONAL RESEARCH COUNCIL

TRANSPORTATION RESEARCH BOARD

Natio

nal R

ese

arc

h C

ouncil

210

1 Co

nstitu

tion A

ve

nue

, N.W

.

Washin

gto

n, D

.C. 2

04

18

AD

DR

ESS SERVICE R

EQUESTED

TRBN

CHR

PR

eport 457Evaluating Intersection Im

provements: An Engineering Study Guide

Program Staff

ROBERT J. REILLY, Director, Cooperative Research Programs

CRAWFORD F. JENCKS, Manager, NCHRP

DAVID B. BEAL, Senior Program Officer

HARVEY BERLIN, Senior Program Officer

B. RAY DERR, Senior Program Officer

AMIR N. HANNA, Senior Program OfficerEDWARD T. HARRIGAN, Senior Program Officer

TIMOTHY G. HESS, Senior Program Officer

RONALD D. McCREADY,

CHARLES W. NIESSNER, Senior Program Officer

Senior Program Officer

EILEEN P. DELANEY, Managing Editor

JAMIE FEAR, Associate Editor

HILARY FREER, Associate Editor

ANDREA BRIERE, Assistant EditorBETH HATCH, Editorial AssistantCHRISTOPHER HEDGES, Senior Program Officer

TRANSPORTATION RESEARCH BOARD EXECUTIVE COMMITTEE 2001

OFFICERSChair: John M. Samuels, Senior Vice President-Operations Planning & Support, Norfolk Southern Corporation, Norfolk, VAVice Chair: Thomas R. Warne, Executive Director, Utah DOT Executive Director: Robert E. Skinner, Jr., Transportation Research Board

MEMBERSWILLIAM D. ANKNER, Director, Rhode Island DOTTHOMAS F. BARRY, JR., Secretary of Transportation, Florida DOTJACK E. BUFFINGTON, Associate Director and Research Professor, Mack-Blackwell National Rural Transportation Study Center, University of ArkansasSARAH C. CAMPBELL, President, TransManagement, Inc., Washington, DCE. DEAN CARLSON, Secretary of Transportation, Kansas DOTJOANNE F. CASEY, President, Intermodal Association of North AmericaJAMES C. CODELL III, Transportation Secretary, Transportation Cabinet, Frankfort, KYJOHN L. CRAIG, Director, Nebraska Department of RoadsROBERT A. FROSCH, Senior Research Fellow, John F. Kennedy School of Government, Harvard UniversityGORMAN GILBERT, Director, Oklahoma Transportation Center, Oklahoma State UniversityGENEVIEVE GIULIANO, Professor, School of Policy, Planning, and Development, University of Southern California, Los AngelesLESTER A. HOEL, L. A. Lacy Distinguished Professor, Department of Civil Engineering, University of VirginiaH. THOMAS KORNEGAY, Executive Director, Port of Houston AuthorityBRADLEY L. MALLORY, Secretary of Transportation, Pennsylvania DOTMICHAEL D. MEYER, Professor, School of Civil and Environmental Engineering, Georgia Institute of TechnologyJEFFREY R. MORELAND, Executive Vice President-Law and Chief of Staff, Burlington Northern Santa Fe Corporation, Fort Worth, TXSID MORRISON, Secretary of Transportation, Washington State DOTJOHN P. POORMAN, Staff Director, Capital District Transportation Committee, Albany, NYCATHERINE L. ROSS, Executive Director, Georgia Regional Transportation AgencyWAYNE SHACKELFORD, Senior Vice President, Gresham Smith & Partners, Alpharetta, GAPAUL P. SKOUTELAS, CEO, Port Authority of Allegheny County, Pittsburgh, PAMICHAEL S. TOWNES, Executive Director, Transportation District Commission of Hampton Roads, Hampton, VAMARTIN WACHS, Director, Institute of Transportation Studies, University of California at BerkeleyMICHAEL W. WICKHAM, Chairman and CEO, Roadway Express, Inc., Akron, OHJAMES A. WILDING, President and CEO, Metropolitan Washington Airports AuthorityM. GORDON WOLMAN, Professor of Geography and Environmental Engineering, The Johns Hopkins University

MIKE ACOTT, President, National Asphalt Pavement Association (ex officio)EDWARD A. BRIGHAM, Acting Deputy Administrator, Research and Special Programs Administration, U.S.DOT (ex officio)BRUCE J. CARLTON, Acting Deputy Administrator, Maritime Administration, U.S.DOT (ex officio)JULIE A. CIRILLO, Assistant Administrator and Chief Safety Officer, Federal Motor Carrier Safety Administration, U.S.DOT (ex officio)SUSAN M. COUGHLIN, Director and COO, The American Trucking Associations Foundation, Inc. (ex officio)ROBERT B. FLOWERS (Lt. Gen., U.S. Army), Chief of Engineers and Commander, U.S. Army Corps of Engineers (ex officio)HAROLD K. FORSEN, Foreign Secretary, National Academy of Engineering (ex officio)JANE F. GARVEY, Federal Aviation Administrator, U.S.DOT (ex officio)EDWARD R. HAMBERGER, President and CEO, Association of American Railroads (ex officio)JOHN C. HORSLEY, Executive Director, American Association of State Highway and Transportation Officials (ex officio)S. MARK LINDSEY, Acting Deputy Administrator, Federal Railroad Administration, U.S.DOT (ex officio)JAMES M. LOY (Adm., U.S. Coast Guard), Commandant, U.S. Coast Guard (ex officio)WILLIAM W. MILLAR, President, American Public Transportation Association (ex officio)MARGO T. OGE, Director, Office of Transportation and Air Quality, U.S. Environmental Protection Agency (ex officio)VALENTIN J. RIVA, President and CEO, American Concrete Pavement Association (ex officio)VINCENT F. SCHIMMOLLER, Deputy Executive Director, Federal Highway Administration, U.S.DOT (ex officio)ASHISH K. SEN, Director, Bureau of Transportation Statistics, U.S.DOT (ex officio)L. ROBERT SHELTON III, Executive Director, National Highway Traffic Safety Administration, U.S.DOT (ex officio)MICHAEL R. THOMAS, Applications Division Director, Office of Earth Sciences Enterprise, National Aeronautics Space Administration (ex officio)HIRAM J. WALKER, Acting Deputy Administrator, Federal Transit Administration, U.S.DOT (ex officio)

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAMTransportation Research Board Executive Committee Subcommittee for NCHRP

ROBERT E. SKINNER, JR., Transportation Research BoardMARTIN WACHS, Institute of Transportation Studies, University of California at

BerkeleyTHOMAS R. WARNE, Utah DOT

JOHN M. SAMUELS, Norfolk Southern Corporation, Norfolk, VA (Chair)LESTER A. HOEL, University of VirginiaJOHN C. HORSLEY, American Association of State Highway and Transportation

OfficialsVINCENT F. SCHIMMOLLER, Federal Highway Administration

ROBERT H. WORTMAN, University of Arizona (Chair)

RICHARD E. BENNETT, Missouri DOT

THOMAS H. CULPEPPER, Wilbur Smith Associates, Falls Church, VA

BETH P. DE ANGELO, Parsons Brinckerhoff-FG, Inc., Princeton, NJ

GLENN M. GRIGG, Sunnyvale, CA

PARA M. JAYASINGHE, City of Boston Public Works Department

STEVEN G. JEWELL, City of Columbus Traffic Engineering Division, Columbus, OH

KEN KOBETSKY, AASHTO

PAWAN MAINI, Fitzgerald & Halliday, Inc., Midlothian, VA

HENRY LIEU, FHWA Liaison Representative

RICHARD A. CUNARD, TRB Liaison Representative

Project Panel G3-58 Field of Traffic Area of Operations and Control

T R A N S P O R T A T I O N R E S E A R C H B O A R D — N A T I O N A L R E S E A R C H C O U N C I L

NATIONAL ACADEMY PRESSWASHINGTON, D.C. — 2001

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

NCHRP REPORT 457

Research Sponsored by the American Association of State Highway and Transportation Officials in Cooperation with the Federal Highway Administration

SUBJECT AREAS

Highway Operations, Capacity, and Traffic Control

Engineering Study Guide for Evaluating Intersection Improvements

JAMES A. BONNESON, P.E. AND

MICHAEL D. FONTAINE

Texas Transportation Institute

Texas A&M University

College Station, TX

Published reports of the

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

are available from:

Transportation Research BoardNational Research Council2101 Constitution Avenue, N.W.Washington, D.C. 20418

and can be ordered through the Internet at:

http://www.national-academies.org/trb/bookstore

Printed in the United States of America

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

Systematic, well-designed research provides the most effectiveapproach to the solution of many problems facing highwayadministrators and engineers. Often, highway problems are of localinterest and can best be studied by highway departmentsindividually or in cooperation with their state universities andothers. However, the accelerating growth of highway transportationdevelops increasingly complex problems of wide interest tohighway authorities. These problems are best studied through acoordinated program of cooperative research.

In recognition of these needs, the highway administrators of theAmerican Association of State Highway and TransportationOfficials initiated in 1962 an objective national highway researchprogram employing modern scientific techniques. This program issupported on a continuing basis by funds from participatingmember states of the Association and it receives the full cooperationand support of the Federal Highway Administration, United StatesDepartment of Transportation.

The Transportation Research Board of the National ResearchCouncil was requested by the Association to administer the researchprogram because of the Board’s recognized objectivity andunderstanding of modern research practices. The Board is uniquelysuited for this purpose as it maintains an extensive committeestructure from which authorities on any highway transportationsubject may be drawn; it possesses avenues of communications andcooperation with federal, state and local governmental agencies,universities, and industry; its relationship to the National ResearchCouncil is an insurance of objectivity; it maintains a full-timeresearch correlation staff of specialists in highway transportationmatters to bring the findings of research directly to those who are ina position to use them.

The program is developed on the basis of research needsidentified by chief administrators of the highway and transportationdepartments and by committees of AASHTO. Each year, specificareas of research needs to be included in the program are proposedto the National Research Council and the Board by the AmericanAssociation of State Highway and Transportation Officials.Research projects to fulfill these needs are defined by the Board, andqualified research agencies are selected from those that havesubmitted proposals. Administration and surveillance of researchcontracts are the responsibilities of the National Research Counciland the Transportation Research Board.

The needs for highway research are many, and the NationalCooperative Highway Research Program can make significantcontributions to the solution of highway transportation problems ofmutual concern to many responsible groups. The program,however, is intended to complement rather than to substitute for orduplicate other highway research programs.

Note: The Transportation Research Board, the National Research Council,the Federal Highway Administration, the American Association of StateHighway and Transportation Officials, and the individual states participating inthe National Cooperative Highway Research Program do not endorse productsor manufacturers. Trade or manufacturers’ names appear herein solelybecause they are considered essential to the object of this report.

NCHRP REPORT 457

Project G3-58 FY’99

ISSN 0077-5614

ISBN 0-309-06705-7

Library of Congress Control Number 2001-131981

© 2001 Transportation Research Board

Price $40.00

NOTICE

The project that is the subject of this report was a part of the National Cooperative

Highway Research Program conducted by the Transportation Research Board with the

approval of the Governing Board of the National Research Council. Such approval

reflects the Governing Board’s judgment that the program concerned is of national

importance and appropriate with respect to both the purposes and resources of the

National Research Council.

The members of the technical committee selected to monitor this project and to review

this report were chosen for recognized scholarly competence and with due

consideration for the balance of disciplines appropriate to the project. The opinions and

conclusions expressed or implied are those of the research agency that performed the

research, and, while they have been accepted as appropriate by the technical committee,

they are not necessarily those of the Transportation Research Board, the National

Research Council, the American Association of State Highway and Transportation

Officials, or the Federal Highway Administration, U.S. Department of Transportation.

Each report is reviewed and accepted for publication by the technical committee

according to procedures established and monitored by the Transportation Research

Board Executive Committee and the Governing Board of the National Research

Council.

FOREWORDBy Staff

Transportation ResearchBoard

This guide describes the engineering study process for evaluating the operationaleffectiveness of various intersection improvements. It also shows how capacity analy-sis and traffic simulation models can be used to assess the operational impacts of thoseimprovements. Use of this guide, particularly by junior traffic engineers, shouldenhance the decision-making process and reduce inappropriate installations of trafficcontrol signals. An enhanced version of this report is available on the world wide webat http://trb.org/trb/publications/nchrp/esg.pdf.

The Manual on Uniform Traffic Control Devices-Millennium Edition (MUTCD2000) states as a standard that an “engineering study of traffic conditions, pedestriancharacteristics, and physical characteristics of the location shall be performed to deter-mine whether installation of a traffic control signal is justified at a particular location.”The MUTCD 2000 furthers states as guidance that “a traffic control signal should notbe installed unless an engineering study indicates that installing a traffic control signalwill improve the overall safety and/or operation of the intersection.” The MUTCD 2000describes some aspects of the engineering study, including the traffic signal warrants,but does not attempt to fully describe the decision-making process.

Capacity analysis (e.g., the methods in the TRB Highway Capacity Manual) and traf-fic simulation models can be beneficially used in these engineering studies to assess theoperational impact of a traffic control signal and other intersection improvements. Some-times these tools may show that, while the warrants are met at a particular location, a lesscostly improvement would operate more effectively than a traffic control signal.

Under NCHRP Project 3-58, the Texas Transportation Institute analyzed difficultiescommonly faced when using traffic signal warrants to determine the appropriatenessof a traffic control signal. They also identified operational measures of effectivenessthat should be considered in the assessment of intersection improvements. They thendeveloped a guide to conducting an engineering study of an intersection.

Evaluating Intersection Improvements: An Engineering Study Guide defines thesteps involved in an engineering study of a problem intersection, beginning with iden-tifying the problem and viable alternative improvements to address the problem. It alsoillustrates how to use capacity analysis and traffic simulation models to determine themost effective operational improvement. The guide does not assist in the analysis ofthe safety or other impacts of the alternative improvements, although these must beconsidered when determining the most appropriate improvement. References to othersources of information on these types of analysis are provided.

The reader may be interested in an enhanced version of the guide available on theweb. That version includes internal hyperlinks between different parts of the report andexternal links to source material. This web version also includes 17 interactive work-sheets that can be helpful in using the guide.

1 CHAPTER 1 IntroductionOverview of the Assessment Process, 1

Alternative Identification and Screening, 1Engineering Study, 1Alternative Selection, 2

Objective and Scope, 2Objective, 2Scope, 2

Overview of the Guide, 2Other Impacts, 4

5 CHAPTER 2 Alternative Identification and ScreeningProcess, 5

Overview, 5Step 1. Define Problem and Cause, 5

1-a. Gather Information, 51-b. Define Problem and Cause, 7

Step 2. Select Candidate Alternatives, 82-a. Identify Potential Alternatives, 82-b. Organize and Select Alternatives, 8

Step 3. Select Viable Alternatives, 103-a. Gather Information, 113-b. Assess and Select Alternatives, 12

Guidelines, 13Guidelines for Use of Selected Geometric and Traffic Control Alternatives, 13

Overview, 13Add Flash Mode to Signal Control, 13Convert to Traffic Signal Control, 14Convert to Multi-Way Stop Control, 15Convert to Two-Way Stop or Yield Control, 16Prohibit On-Street Parking, 18Prohibit Left-Turn Movements, 19Convert to Roundabout, 19Add a Second Lane on the Minor Road, 20Add a Left-Turn Bay on the Major Road, 21Add a Right-Turn Bay on the Major Road, 22Increase Length of Turn Bay, 23Increase the Right-Turn Radius, 25

Conditions Affecting the Accuracy of Conclusions from the Signal Warrant Check, 26

Overview, 26Right-Turn Volume on the Minor Road, 26Heavy Vehicles on the Minor Road, 27Pedestrian Volume, 27Progressive Traffic Flow on the Major Road, 28Three-Leg Intersection, 28Added Through Lane on the Major Road, 29Left-Turn Bay, 29Right-Turn Bay on the Minor Road, 29Wide Median on the Major Road, 30Combination of Factors, 30

31 CHAPTER 3 Engineering StudyProcess, 31

Overview, 31Step 1. Determine Study Type, 31

1-a. Determine Type of Operation, 311-b. Determine Type of Evaluation, 32

Step 2. Select Analysis Tool, 332-a. Identify Desired Capabilities, 332-b. Evaluate and Select Analysis Tool, 33

CONTENTS

Step 3. Conduct Evaluation, 363-a. Gather Information, 363-b. Evaluate Operational Performance, 383-c. Determine Alternative Effectiveness, 38

Guidelines, 39Guidelines for Designing the Signalized Intersection Alternative, 39

Overview, 39Intersection Geometry, 40Controller Operation, 41Controller Phase Sequence, 42Basic Controller Settings, 44Controller Settings for Isolated Operation, 45Controller Settings for Coordinated Operation, 52

Guidelines for Use of Stochastic Simulation Models, 56Overview, 56Random Arrivals, 56Random Number Seed, 56Minimum Simulation Initialization Time, 56Simulation Period and Run Duration, 56

60 CHAPTER 4 Alternative SelectionProcess, 60

Overview, 60Step 1. Identify Impacts, 60

1-a. Identify Decision Factors, 601-b. Assess Degree of Impact, 61

Step 2. Select Best Alternative, 61Step 3. Document Study, 63

64 BIBLIOGRAPHY

65 REFERENCES

A-1 APPENDIX A Case Study Applications

B-1 APPENDIX B MUTCD Traffic Control Signal Warrant Worksheets

C-1 APPENDIX C Data for Guideline Evaluation

AUTHOR ACKNOWLEDGMENTSThe research team would like to recognize the support and guid-

ance provided by the project panel. The project panel membersinclude Dr. Robert Wortman (chair), University of Arizona; Mr.Rick Bennett, Missouri Department of Transportation; Dr. ThomasH. Culpepper, Wilbur Smith and Associates; Ms. Beth P. DeAngelo, Parsons Brinkerhoff—FG, Inc.; Mr. Glenn M. Grigg, con-sultant; Mr. Para M. Jayasinghe, City of Boston, Massachusetts;Mr. Steve G. Jewell, City of Columbus, Ohio; Mr. Ken Kobetsky,AASHTO; Mr. Pawan Maini, Fitzgerald and Halliday, Inc.; and Mr.Henry Lieu, Federal Highway Administration.

Finally, the research team would like to acknowledge the engi-neers who participated in the focus group assembled for the researchproject. These individuals include Mr. W. Martin Bretherton, Jr.,Gwinnett County Department of Transportation, Georgia; Mr. RonNorthouse, City of San Jose, California; Mr. Michael R. Scanlon,City of Tampa, Florida; Mr. Richard Filkins, Washington StateDepartment of Transportation; and Mr. Chad J. Smith, Iowa Depart-ment of Transportation. Their thoughtful advice and comment onthe procedures and guidelines are greatly appreciated.

Abbreviations used without definitions in TRB publications:

AASHO American Association of State Highway OfficialsAASHTO American Association of State Highway and Transportation OfficialsASCE American Society of Civil EngineersASME American Society of Mechanical EngineersASTM American Society for Testing and MaterialsFAA Federal Aviation AdministrationFHWA Federal Highway AdministrationFRA Federal Railroad AdministrationFTA Federal Transit AdministrationIEEE Institute of Electrical and Electronics EngineersITE Institute of Transportation EngineersNCHRP National Cooperative Highway Research ProgramNCTRP National Cooperative Transit Research and Development ProgramNHTSA National Highway Traffic Safety AdministrationSAE Society of Automotive EngineersTCRP Transit Cooperative Research ProgramTRB Transportation Research BoardU.S.DOT United States Department of Transportation

Advisers to the Nation on Science, Engineering, and Medicine

National Academy of SciencesNational Academy of EngineeringInstitute of MedicineNational Research Council

The Transportation Research Board is a unit of the National Research Council, which serves the National Academy of Sciences and the National Academy of Engineering. The Board’s mission is to promote innovation and progress in transportation by stimulating and conducting research, facilitating the dissemination of information, and encouraging the implementation of research results. The Board’s varied activities annually draw on approximately 4,000 engineers, scientists, and other transportation researchers and practitioners from the public and private sectors and academia, all of whom contribute their expertise in the public interest. The program is supported by state transportation departments, federal agencies including the component administrations of the U.S. Department of Transportation, and other organizations and individuals interested in the development of transportation.

The National Academy of Sciences is a private, nonprofit, self-perpetuating society of distin-guished scholars engaged in scientific and engineering research, dedicated to the furtherance of science and technology and to their use for the general welfare. Upon the authority of the charter granted to it by the Congress in 1863, the Academy has a mandate that requires it to advise the federal government on scientific and technical matters. Dr. Bruce M. Alberts is president of the National Academy of Sciences.

The National Academy of Engineering was established in 1964, under the charter of the National Academy of Sciences, as a parallel organization of outstanding engineers. It is autonomous in its administration and in the selection of its members, sharing with the National Academy of Sciences the responsibility for advising the federal government. The National Academy of Engineering also sponsors engineering programs aimed at meeting national needs, encourages education and research, and recognizes the superior achievements of engineers. Dr. William A. Wulf is president of the National Academy of Engineering.

The Institute of Medicine was established in 1970 by the National Academy of Sciences to secure the services of eminent members of appropriate professions in the examination of policy matters pertaining to the health of the public. The Institute acts under the responsibility given to the National Academy of Sciences by its congressional charter to be an adviser to the federal government and, upon its own initiative, to identify issues of medical care, research, and education. Dr. Kenneth I. Shine is president of the Institute of Medicine.

The National Research Council was organized by the National Academy of Sciences in 1916 to associate the broad community of science and technology with the Academy’s purpose of furthering knowledge and advising the federal government. Functioning in accordance with general policies determined by the Academy, the Council has become the principal operating agency of both the National Academy of Sciences and the National Academy of Engineering in providing services to the government, the public, and the scientific and engineering communities. The Council is administered jointly by both the Academies and the Institute of Medicine. Dr. Bruce M. Alberts and Dr. William A. Wulf are chairman and vice chairman, respectively, of the National Research Council.

1

CHAPTER 1

INTRODUCTION

The Manual on Uniform Traffic Control Devices-Millennium Edition (MUTCD 2000) (1) includes warrantsthat set minimum thresholds for considering the installationof a traffic signal. The threshold values used in these war-rants were established many years ago. They are based largelyon engineering judgment and reflect sensitivity to two vari-ables: approach volume and number of lanes. However,their subjective basis and limited sensitivity can make theminaccurate predictors of the need for signal control at someintersections and could result in the unnecessary installa-tion of traffic signals. In fact, two separate evaluations ofthe MUTCD warrants found that they do not always yieldconclusions that agree with engineering judgment (2,3).This finding is a concern because a recent literature reviewby Bonneson and Fontaine (4) revealed that delays and stopscan increase 100 to 200 percent when marginally warrantedsignals are installed at an intersection.

To eliminate the unnecessary installation of traffic signals,the text of the MUTCD was revised for its 1988 publication.The purpose of the revisions was to indicate clearly that thedecision to install a traffic signal should be based on the find-ings from an engineering study. The 1998 MUTCD statedthat “satisfaction of a warrant or warrants is not in itself jus-tification for a signal.” This text was retained for the MUTCD2000 and further states that “A traffic control signal shouldnot be installed unless an engineering study indicates thatinstalling a traffic control signal will improve the overallsafety and/or operation of the intersection.” The intent of thisrecommendation is to encourage consideration of the fullrange of intersection improvement alternatives and selectionof the most effective alternative for implementation.

This guide documents the steps involved in the formalengineering study of improvement alternatives and focuseson the use of capacity analysis procedures and simulationmodels (i.e., analysis tools) to evaluate the operational im-pacts of improvement alternatives. Case studies illustrate howthese analysis tools can sometimes be used to show that anintersection would operate more effectively without a trafficsignal, even though a signal warrant is met.

OVERVIEW OF THE ASSESSMENT PROCESS

Generally, the engineering assessment process is initiatedwhen operational or safety problems are identified at an un-signalized intersection. The goal of this process is to identify

the most effective solution to the identified problem. Theassessment process is often integral to a larger, more com-prehensive system management process in which all prob-lematic transportation facilities are considered for improve-ment. The stages of the system management process areshown in Figure 1-1.

The assessment process represents three specific stages ofthe system management process. In the first stage, viable al-ternatives are identified. In the second stage, an engineeringstudy is conducted to evaluate the effectiveness of each viablealternative. Finally, the best alternative is selected on the basisof its effectiveness and its other, non-motorist-related effects.

Following the assessment process, the best alternative fora given intersection is pooled with other improvement proj-ects for funding consideration. If funding is available, thealternative is constructed, and the intersection is monitoredto confirm that the original problem has been solved.

Alternative Identification and Screening

During the alternative identification and screening stage,the traffic engineer makes a preliminary assessment of theproblem and identifies viable intersection improvement alter-natives. Alternatives may include changes in traffic control,intersection geometry, or both. Traffic control alternativescan include two-way stop control, multi-way stop control, or signal control. Geometric alternatives that are sometimesidentified include the addition of left or right-turn bays on themajor or minor roadways.

Initially, alternatives are identified that could solve theobserved problem. Then, this list of candidate alternatives isreduced, using formal guidelines or warrants, to include onlythe most viable alternatives. Formally recognized guidelinesexist for traffic signals, multi-way stop control, and somemajor geometric improvements. Guidelines for control deviceapplications are provided in the MUTCD 2000 (1). Similarly,guidelines that describe when a left-turn bay is needed on themajor-road approach to an unsignalized intersection are pro-vided in A Policy on Geometric Design of Highways andStreets (Green Book) (5).

Engineering Study

During the engineering study stage, the viable alternativesare evaluated in terms of their effect on road users and the

immediate environment. The accurate assessment of theseeffects typically requires the application of formally recog-nized analysis procedures. However, in some less-complexsituations, experience with similar alternatives and operatingconditions may be sufficient to enable the engineer to estimatean alternative’s effectiveness.

Alternative Selection

During the alternative selection stage, the engineer assessesthe effects of each alternative and then selects the “best” onefor implementation. Effects considered may include improve-ments in traffic operations or safety and disruption to area aes-thetics, the environment, or adjacent property. Selection ofthe best alternative may also reflect consideration of the con-struction cost of each alternative. The method of selection canvary from a complicated life-cycle, benefit-cost analysis to asimple identification of the alternative that yields the leasttraffic delay. The methods used and the effects consideredwill depend on the conditions present at the problem location.

OBJECTIVE AND SCOPE

Objective

This guide has two objectives: (1) to define the steps in-volved in an engineering study of a problem intersection and(2) to provide guidelines for using capacity analysis or simu-lation to determine the most effective alternative on the basisof operational considerations. To achieve these objectives,the coverage of the guide is expanded to include the threestages of the assessment process, as shown in Figure 1-1. Thisapproach provides a comprehensive treatment of problemintersections by preceding the engineering study stage with

2

an alternative identification and screening stage and follow-ing it with an alternative selection stage. Application of allthree stages will ensure that a wide range of alternatives isconsidered and that the alternative selected is effective.

Scope

The guide is intended to describe how to evaluate the oper-ational effects of alternative geometrics and control modes ata problem intersection. This description includes guidelineson (1) the alternative selection process and (2) the use ofcapacity analysis procedures and simulation models for alter-native evaluation. The assessment of an alternative’s safety(or other) effect is beyond the scope of the guide. The analystis encouraged to consult the References and Bibliography forguidance on safety evaluations.

The target audience of this guide is the junior traffic engi-neer. Procedures and guidelines provided herein are intendedto identify critical decision points and problem-solving tech-niques for engineers with only a few years of experience. Tomeet the needs of this intended audience, each topic is coveredthoroughly. Sometimes, the material herein may contradict thepolicies and procedures of the reader’s agency. In such in-stances, agencies may substitute their policies or practices forthose in this guide.

This guide is intended to be applicable to intersectionswith a wide range of control modes. However, the three modescommonly found at U.S. intersections are given greater em-phasis in some sections of the document. These modes are asfollows:

1. Two-way stop control,2. Multi-way stop control, and3. Signal control.

Although the three control modes listed above are empha-sized, the procedures in the guide can be used to evaluateother control modes (e.g., no control, two-way yield control,and roundabouts).

OVERVIEW OF THE GUIDE

This guide consists of four chapters and three appendixes.Chapters 2, 3, and 4 describe the three stages of the assess-ment process, as shown in Figure 1-1. A flowchart diagram-ming the assessment process is shown in Figure 1-2.

Chapter 2 of the guide describes the alternative identifica-tion and screening stage. The chapter describes how to iden-tify the problems at the subject intersection, how to identifycandidate improvement alternatives, and how to screen thesealternatives so that only the most viable alternatives are eval-uated in the engineering study.

Chapter 3 of the guide describes the engineering studystage. This chapter describes the process for evaluating the

Figure 1-1. Stages of the system management process.

3

Figure 1-2. Flow chart of the assessment process.

effectiveness of the selected alternatives. The process focuseson how to evaluate traffic operations by using capacity analy-sis procedures or simulation models. Guidelines are providedto help the analyst determine the most appropriate analysistool (i.e., capacity analysis procedure or simulation model),the data needed for the evaluation, and whether or not analternative operates satisfactorily.

Chapter 4 of the guide describes the alternative selectionstage, including how to identify the best alternative for agiven intersection. Alternative selection is described in gen-eral, rather than precise, terms, so as to allow the analystsome flexibility in selecting the types of effects to considerwhen selecting an alternative and the relative weight to beapplied to each effect.

Guidelines that support the assessment process are pro-vided in the Guidelines section of Chapters 2 and 3. Thematerials provided include (1) a list of traffic control andgeometric design alternatives, (2) warrants and guidelines de-scribing conditions suitable for selected alternatives, (3) tech-niques for designing the traffic signal control alternative, and(4) guidelines for using stochastic simulation models. Use ofthis guide should help agencies to maximize their return oninfrastructure investment and that they will avoid the unnec-essary installation of traffic signals.

The appendixes to the guide are intended to provide sup-plementary information that makes the guide a more versatiledocument. The use of the guide is demonstrated in severalcase study situations in Appendix A. Appendix B provides

4

worksheets for documenting the MUTCD signal warrant check.Appendix C describes techniques that can be used to gatheror derive the data needed for the engineering study.

Although the coverage in the guide is thorough, it is as-sumed that the reader will have access to two reference doc-uments: (1) the current edition of the MUTCD and (2) theManual of Transportation Engineering Studies (6). This lat-ter document describes procedures for collecting data for thedirect evaluation of intersection performance and for collect-ing the input data needed for a capacity analysis procedure orsimulation model.

Other Impacts

Although the focus of the guide is on evaluating the oper-ational effects of improvement alternatives, the analyst shouldalso consider safety and other effects during the engineeringstudy. Safety impact assessment may range from an informalsubjective assessment to a formal quantitative evaluation,depending on the types of problems being experienced (oranticipated) at the subject intersection. Often, the best alter-native will be the one that improves traffic operation andenhances safety. However, this may not always be the case—some alternatives may improve operation but degrade safetyand vice versa. Therefore, the analyst should carefully eval-uate and weigh all relevant effects when selecting the bestalternative.

5

CHAPTER 2

ALTERNATIVE IDENTIFICATION AND SCREENING

Intersections that are inefficient or unsafe are typicallysubjected to some type of engineering assessment to identifythe underlying problem and its solution. The engineeringassessment process consists of three stages: (1) alternativeidentification and screening, (2) the engineering study, and(3) alternative selection. The first stage of the assessmentprocess is the subject of this chapter. The steps involved inthe alternative identification and screening stage are definedin the first section of this chapter. Guidelines are provided inthe second section. These guidelines describe (1) conditionswhere selected improvement alternatives may be helpful and(2) conditions that might mislead the signal warrant check.

PROCESS

Overview

The alternative identification and screening stage consistsof three steps. These steps are as follows:

1. Define problem and cause,2. Select candidate alternatives, and3. Select viable alternatives.

In the first step, the problem is defined and its cause is identi-fied through the conduct of a site visit and the collection of rel-evant existing data. Then, several alternatives are selected forfurther consideration. Finally, these alternatives are screenedusing available engineering guidelines so that a subset list ofviable alternatives is identified.

The objective of the alternative identification and screen-ing stage is to determine if a problem exists and, if it does, toidentify one or more viable alternatives that can eliminate or mitigate the problem. This objective is achieved throughan evaluation of existing conditions and consideration of arange of improvement alternatives. The steps in this processare described in the remainder of this section.

Step 1. Define Problem and Cause

The first step in the alternative identification and screen-ing stage is to determine the nature of the problem at an inter-section and to define its potential causes. The tasks involvedin making this determination are to

a. Gather information andb. Define the problem and cause.

In the first task, data are collected that describe the inter-section’s history and its present condition in terms of trafficvolume, safety record, and geometric layout. Then, these dataare used to define the intersection’s operational or safetyproblems and to identify their causes.

The assessment process typically begins when the engi-neer is notified of a problem at an intersection or series ofintersections. This notification can come from sources thatare either internal or external to the agency. Some possiblesources of problem reports include

• Complaints from local residents about existing inter-sections,

• Developers seeking traffic control for proposed inter-sections,

• Observations from field crews,• Data obtained from consultant studies,• Information obtained from regional traffic counts,• Potential problematic conditions identified through

regional traffic projections, and• Annual safety analysis of high-crash locations.

1-a. Gather Information

Overview. Following notification of a problem, the engi-neer makes a preliminary assessment to determine the extentof the problem and whether it requires a solution. This as-sessment involves gathering readily available informationabout the problem intersection so that the problem can bedefined and its cause identified. Information sources includehistoric file data, firsthand observation, and a site survey.

Completion of this task should not require a rigorous datacollection effort. A visit to the site should provide sufficientinformation to determine if a problem exists. Traffic volume,conflict, or delay data should not be needed to make this deter-mination. In Step 3 of the process, Select Viable Alternatives,additional data will be collected to allow a more detailedassessment of the problem and its solution.

Historic Data. The analyst should gather existing engi-neering studies, corridor studies, traffic studies, and crash

data summaries that may help to identify potential problemsat the subject intersection. Crash data summarized in a colli-sion diagram may indicate where problems are occurring atthe intersection. (Details on the construction of a collisiondiagram are provided in Appendix C.)

Observational Study. An important component of theproblem-cause identification process is the firsthand obser-vation of traffic operations. A visit to observe operationsshould be scheduled to coincide with the occurrence of thereported problems (e.g., p.m. peak traffic demand hour). Ini-

6

tially, the engineer should drive through the subject intersec-tion and attempt to experience the problem. Then, the engi-neer should observe intersection operation from a curbsidevantage point. An onsite observation report, such as thatshown in Figure 2-1, should be completed during the site visit.(Details regarding the completion of this report and a blankreport form are provided in Appendix C.)

Site Survey. The engineer should also have a survey ofsite conditions conducted. The survey data should be recordedon a condition diagram. The diagram represents a plan-view,

Figure 2-1. Sample onsite observation report.

scale drawing of the subject intersection. Conditions recordedmay include road width, pavement markings, speed limits,and traffic control devices. (Details of the site survey and ablank condition diagram are provided in Appendix C.)

Traffic Projections. Traffic volume projections are im-portant when a new intersection is being proposed. Projectedturning movement counts may provide the best data avail-able for the analyst to determine the most appropriate type ofcontrol for the intersection. The analyst should also deter-mine if major changes in traffic patterns or land use are antic-ipated in the vicinity of the subject intersection. Improvementalternatives should be able to accommodate future trafficpatterns.

1-b. Define Problem and Cause

Overview. During this task, the analyst will define theproblem and identify its cause. The information gathered inthe preceding task is used for this purpose. Initially, the ana-lyst will review this information and determine if sufficientevidence of a problem exists. If it is determined that a prob-lem exists, then the problem is formally defined and its causeis identified.

7

Assess Evidence. After reviewing the available historicalinformation and information obtained from the site visit, theengineer should determine if sufficient evidence of a prob-lem exists to proceed further in the assessment process. Ifsuch evidence exists, then further study may be necessary.

Define Problem and Identify Cause. Problems can gen-erally be categorized as operational or safety-related. Opera-tional problems are typically associated with excessive delayto one or more traffic movements. Safety-related problemsare typically associated with frequent conflicts, erratic maneu-vers, non-compliance with control devices, and collisions.The information obtained during Task 1-a should be used todetermine which (if any) of these two categories of problemsexist for each of the intersection traffic movements.

Once the problem has been defined, the engineer shouldidentify what is causing the problem. Problems and potentialcauses are listed in Table 2-1. The possible causes for eachproblem category (i.e., delay or conflict) are identified bycheckmark (✓ ). When using Table 2-1, each intersection ap-proach should be individually evaluated. The movementsexperiencing a problem should be identified and comparedwith the checked combinations shown. The terms “delay”and “conflict” refer broadly to operational and safety prob-lems observed during the site visit (they are not meant to

TABLE 2-1 Common operational problems and possible causes

imply that a delay or conflict study must be conducted beforeTable 2-1 can be consulted).

To illustrate the use of Table 2-1, consider a minor-roadapproach at a four-leg intersection. All three movements onthis approach were observed to have excessive delay (but noconflicts). Table 2-1 indicates that three possible causes areassociated with the observed minor-road delays: (1) inade-quate capacity for minor-road movements, (2) inadequate sep-aration of minor-road movements, and (3) staggered arrival ofmajor-road platoons. These possible causes are advanced tothe next step to determine candidate improvement alternatives.

Define Influence Area. Designation of the intersection“influence area” is an important part of this task. This area mayinclude the subject intersection only or it may include the net-work of roads that surround the subject intersection. As a min-imum, the influence area should extend from the subject inter-section sufficiently far so as to include the queues associatedwith the intersection’s traffic movements.

The extent of the influence area’s coverage is based pri-marily on the proximity of the subject intersection to otherintersections. The inclusion of nearby intersections in theinfluence area would be based in part on the answers to Ques-tions A, B, C, and D on the observation report (Figure 2-1).If one or more of these questions are answered “Yes” or “NotSure,” then the subject intersection may be influenced by oneor more nearby intersections.

Efforts to improve conditions at the subject intersectionshould be sensitive to the operation of all intersections in theinfluence area. All intersections in the influence area shouldbe included in the evaluation conducted in the engineeringstudy stage (see Chapter 3).

Step 2. Select Candidate Alternatives

After the problem is defined and its likely cause is identified,one or more candidate alternatives should be selected. Thissection provides a procedure for selecting candidate improve-ment alternatives. The procedure consists of the followingtwo tasks:

a. Identify potential alternatives andb. Organize and select alternatives

During the first task, a list of alternatives that could solve theproblems is identified. Then, these alternatives are organizedand screened to obtain a subset of candidate alternatives basedon site constraints and institutional preferences.

2-a. Identify Potential Alternatives

Overview. Tables 2-2 and 2-3 are provided in this sec-tion to assist in the identification of potential improvementalternatives. The first table applies to problems that require

8

intersection-specific solutions and the second table appliesto intersections requiring system-related solutions. The alter-natives listed in these tables emphasize operational improve-ments; the analyst is referred to Chapter 11 of the Manual ofTransportation Engineering Studies (6 ) for a complete list ofsafety problems, causes, and possible solution alternatives.

Intersection-Specific Alternatives. Table 2-2 identifiesproblems commonly encountered at intersections. This tablealso lists potential corrective strategies and candidate alter-natives. Problems and causes other than those listed may alsobe known to exist at a specific intersection. In these situa-tions, judgment should be used to identify the appropriateimprovement alternatives.

The characteristics of the potential alternatives listed inTable 2-2 are indicated by underline and italic font. Alter-natives that affect traffic operations (e.g., motorist delay)are underlined. Alternatives shown in italics should be eval-uated for viability in the next step (i.e., Step 3 - SelectViable Alternatives).

To illustrate the use of Table 2-2, consider a minor-roadintersection approach observed to have excessive delay causedby inadequate separation of its traffic movements. Table 2-2indicates that one corrective strategy is to separate the con-flicting movements by using one of the following alternatives:(1) add a second lane on the approach or (2) increase right-turnradius. This latter alternative would widen the throat of theapproach and effectively separate right-turning vehicles fromthrough and left-turning vehicles.

System-Related Alternatives. Table 2-3 lists the prob-lems that may be found at non-isolated intersections. Theseintersections have an influence area that extends beyond thelimits of the subject intersection and includes the adjacentintersections (either unsignalized or signalized). The improve-ment alternatives listed in both Table 2-2 and Table 2-3 shouldbe considered when the intersection is not isolated. The alter-natives that affect traffic operations are underlined; thosealternatives whose viability can be evaluated in the next stepare italicized.

2-b. Organize and Select Alternatives

Overview. During this task, the potential alternatives arescreened to determine if they are suitable candidates for thesubject location and whether their immediate implementa-tion is more cost-effective than the conduct of a formal engi-neering study. If no alternatives are selected for immediateimplementation, the alternatives that remain after screeningrepresent the “candidate” improvement alternatives. Thesealternatives are then advanced to Step 3.

Identify Candidate Alternatives. Initially, the list ofpotential improvement alternatives should be screened for

9

TABLE 2-2 Potential intersection-specific, engineering improvement alternatives

suitability on the basis of firsthand knowledge of site con-straints and agency preferences. For example, if the site beingstudied is in a downtown area where right-of-way is limited,the analyst may choose to eliminate options that will requirethe acquisition of significant amounts of right-of-way. Thisscreening is performed without the collection of additionaldata, so only options that would clearly not be acceptableshould be eliminated.

Identify Alternatives Suitable for Immediate Imple-mentation. All low-cost alternatives should be considered forimmediate implementation and evaluation. An alternative’s“cost” includes the direct cost of its implementation as well asany indirect cost to adjacent land users and the environment.Low-cost alternatives are defined as those alternatives thathave a cost that is significantly less than that of the formalengineering study. Typical low-cost alternatives include ad-visory signing, revised pavement markings, and vegetationremoval (to improve sight lines).

If any low-cost alternatives have been identified, the engi-neer can proceed directly to the implementation stage. In thisstage, the low-cost alternative(s) would be programmed for

10

implementation and monitored for effectiveness. If a follow-up observational study reveals that the alternatives imple-mented in this manner have not substantially improved thereported problems, then a new assessment process should be initiated.

Step 3. Select Viable Alternatives

The third step in the alternative identification and screen-ing stage is to determine the viability of the candidatealternatives identified in Step 2. The tasks in making thisdetermination are to

a. Gather information andb. Assess and select alternatives

In the first task, relevant guidelines are identified, and corre-sponding data are collected to facilitate the evaluation of thecandidate alternatives identified in Step 2. Then, in the sec-ond task, the guidelines are used to determine which of thecandidate alternatives are more viable than the others.

TABLE 2-3 Potential system-related, engineering improvement alternatives

The purpose of Step 3 is to screen out those alternativesthat are not likely to have a significant positive effect on inter-section operations. By screening out these alternatives, it ishoped that the list of alternatives to be evaluated during theengineering study (described in Chapter 3) will be reduced sothat it includes only the most promising alternatives.

3-a. Gather Information

Overview. The procedure for selecting viable alternativesis based on a review of guidelines that indicate when animprovement alternative is likely to be effective. These guide-lines are provided in the Guidelines section of this chapter. Ifdesired, other guidelines may be added or substituted by theresponsible agency. The Guidelines section addresses thefollowing alternatives:

• Add flash mode to signal control,• Convert to traffic signal control,• Convert to multi-way stop control,• Convert to two-way stop control,• Convert to two-way yield control,• Prohibit on-street parking,• Prohibit left-turn movements,• Convert to roundabout,• Add a second lane on the minor road,• Add a left-turn bay on the major road,• Add a right-turn bay on the major road,• Increase the length of the turn bay, and• Increase right-turn radius.

(The order in which these alternatives are listed is arbitrary andis not intended to convey any sense of priority or importance.)

Alternative Categories. Some alternatives offered inTables 2-2 and 2-3 are not included in the list above. Theseomissions occurred because (1) there is no formal guidanceavailable in the literature, or (2) the alternative’s effect cannotbe quantified in terms of delay. Therefore, it is possible thatonly a subset of the candidate alternatives can be evaluated inthis step.

Table 2-4 indicates the action to be taken based on thealternative’s effect and guideline availability. At the onset ofthis task, Table 2-4 should be consulted for each candidate

11

alternative to determine the appropriate action. Alternativesin Category I should be assessed using the process describedin this section.

Identify Data. The next activity to be undertaken for thistask is to gather the data needed to apply the guidelines appli-cable to each Category-I alternative. The specific types of dataneeded to evaluate each guideline are listed in Table 2-5. Thelist is generally complete for all of the guidelines shown;however, additional data may be needed in specific situa-tions. In all cases, the analyst should review the guideline, asdescribed in the Guidelines section, to determine if additionaldata are needed. Data needed for one guideline, ProhibitLeft-Turn Movements, are not listed in the table because theyare unique; however, they are described in the Guidelinessection of this chapter.

Before gathering any data, the analyst should identify thespecific data needed for the collective set of guidelines tobe evaluated. Frequently, data needed for one guideline canbe used for another guideline. The data collection plan shouldtake advantage of such overlap to avoid redundancy.

If the traffic signal or multi-way stop control alternative isbeing considered, the analyst should note that only one com-ponent of the respective warrants or criteria need to be satis-fied to designate the alternative as “viable.” Thus, the analystshould determine which of the signal warrants and the multi-way stop control criteria will be evaluated and collect dataonly for these selected warrants and criteria. Often, the prob-lems and causes identified in Step 2 will direct the selectionof the warrants or criteria to be evaluated. When this is notthe case, Hawkins and Carlson (7 ) recommend evaluation ofMUTCD 2000 Warrants 1, 2, and 3 as a first step, because theassociated data require the least effort to collect.

Collect Data. The procedures for collecting the data neededto evaluate the selected guidelines will vary depending onwhether the subject intersection exists or is proposed for con-struction. For existing intersections, appropriate data collec-tion procedures are described in the Manual of Transporta-tion Engineering Studies (6). For proposed intersections,techniques described in Appendix C can be used to estimateturn movement volumes from forecast average daily trafficdemands. Regardless of the source, the data should representtraffic conditions occurring on an “average day” (i.e., a day

TABLE 2-4 Categories of candidate alternatives

representing traffic volumes normally and repeatedly found ata location).

3-b. Assess and Select Alternatives

Overview. During this task, the Category-I alternatives areevaluated in order to develop a subset list of viable alterna-tives. The information gathered in Task 3-a is used to evalu-ate each alternative using the appropriate guideline. The stepsinvolved in the application of each guideline are listed in theGuidelines section. Those alternatives that satisfy their corre-sponding guideline should be considered “viable” and evalu-ated more formally during the engineering study (described inChapter 3).

At this point, some alternatives may appear less promisingthan others do in terms of their ability to solve a problem, andthe analyst may be tempted to drop the less promising alter-natives. However, these alternatives may provide the best bal-

12

ance between implementation impact and improvement ef-fectiveness. The most appropriate alternative can only beaccurately identified after the effectiveness of each alterna-tive has been evaluated during the engineering study. In short,all viable alternatives should be advanced to the engineeringstudy stage.

Traffic Signal Alternative Issues. Two issues should beaddressed when the traffic signal alternative is found to beviable (i.e., one or more warrants are satisfied). Considerationof these issues is important because of the potential for thesignal to affect intersection operation or safety negatively.The first issue relates to the belief of some engineers that thetraffic signal is the most direct means of solving all intersec-tion problems. The second issue relates to the presence ofatypical (or problematic) conditions that may reduce the util-ity of the warrant check.

With regard to the first issue, the analyst might be inclinedto drop all other alternatives when traffic signal control is

TABLE 2-5 Data needed to evaluate guidelines

found to be viable. However, the consequences of droppingthe other alternatives can be significant. For example, stud-ies show that when stop control is converted to signal controland the intersection volume “just” satisfies a warrant, theresulting overall delay often increases. Specifically, Williamsand Ardekani (8) found that delays increased by as much as 113 percent; Kay et al. (9) found that delays increased 200 percent; and Bissell and Neudorff (10) found that delaysincreased by 10 seconds per vehicle (s/veh) when traffic sig-nals were used to replace two-way stop control. Therefore,when the signal control alternative is found to be viable, theanalyst is strongly encouraged to identify and advance otheralternatives to the engineering study stage.

Regarding the second issue, atypical or unusual traffic,signalization, or geometric conditions may reduce the accu-racy of the results of the warrant check. Some of the morefrequently encountered problematic conditions include thefollowing:

• Right-turn volume on the minor road,• Heavy vehicles on the minor road,• Pedestrian volumes,• Progressive traffic flow on the major road,• Three-leg intersection,• Added through lane on the minor road,• Left-turn bay,• Right-turn bay, and• Wide median on the major road.

(The order in which the conditions are listed is arbitrary andis not intended to convey any sense of priority or importance.)

With regard to the second issue, guidance provided in theGuidelines section can be used to determine if a problematiccondition is likely to affect the results of the warrant check. Ifa problematic condition exists, then the effect of the conditionon intersection operations should be carefully evaluated in theengineering study stage.

GUIDELINES

This section provides guidelines that can be used during thealternative identification and screening stage. These guide-lines are presented in two separate sections with the followingtitles:

1. Guidelines for Use of Selected Geometric and TrafficControl Alternatives.

2. Conditions Affecting the Accuracy of Conclusions fromthe Signal Warrant Check.

The first section describes guidelines that can be used toevaluate the viability of 13 alternatives. The second sectiondescribes nine problematic traffic, signalization, or geomet-ric conditions that reduce the accuracy of the results of thewarrant check.

13

Guidelines for Use of Selected Geometric and Traffic Control Alternatives

Overview

This section provides guidelines that can be used to deter-mine when various traffic control devices and geometric ele-ments may be helpful in improving intersection operations orsafety. These guidelines were obtained primarily from refer-ence documents that are generally recognized as authoritativeguides on engineering practice. Other sources for the guide-lines include reports documenting significant research effortswhose recommendations are intended for nationwide appli-cation. The alternatives for which guidelines are provided inthis section include

• Add flash mode to signal control,• Convert to traffic signal control,• Convert to multi-way stop control,• Convert to two-way stop or yield control,• Prohibit on-street parking,• Prohibit left-turn movements,• Convert to roundabout,• Add a second lane on the minor road,• Add a left-turn bay on the major road,• Add a right-turn bay on the major road,• Increase the length of the turn bay, and• Increase the right-turn radius.

(The order in which the alternatives are listed is arbitrary andis not intended to convey any sense of priority or importance.)

The guidelines described in this section tend to be conser-vative such that they indicate only when an alternative maybe helpful (i.e., viable). If an alternative is found to satisfy theguideline threshold conditions, then the effectiveness of thealternative should be verified through the conduct of an engi-neering study (as described in Chapter 3). The results of theengineering study should form the basis for any recommen-dation to implement an alternative. In contrast, if the guide-line is not satisfied, then it should be assumed that the corre-sponding alternative is not viable and should be droppedfrom further consideration.

Finally, the guidelines provided in this section do not in-clude all possible improvement alternatives. The guidelinesprovided are believed to correspond to the more commonlyimplemented alternatives. Judgment may be needed to iden-tify viable alternatives for intersections that have unique oper-ating or geometric conditions or when several alternatives areproposed for use in combination at a specific intersection.

Add Flash Mode to Signal Control

Introduction. Intersections with light-to-moderate trafficdemands may benefit from traffic signal control during the

hours of peak demand but may not derive benefit during off-peak hours. When this benefit results from a reduction inmotorist delay, it may be useful to operate the signal in a flashmode during the off-peak hours. Flash mode may consist ofa yellow/red combination or a red/red combination. For theyellow/red combination, the major road is flashed yellow andthe minor road is flashed red.

Pusey and Butzer (11) indicate that flash mode operationhas several benefits. These benefits can include

• Reduced stops and delays to major-road traffic,• Reduced delay to minor-road traffic,• Reduced electrical consumption, and• Reduced vehicular fuel consumption and traffic noise.

A review of the literature by Kacir et al. (12) revealed thatyellow/red flash mode may be associated with higher crashrates, especially when the ratio of major-road-to-minor-roadvolume is less than 2.0.

Guidance. Benioff et al. (13) define conditions in whichflash mode is not likely to cause safety problems. Specifi-cally, they recommend using flashing yellow/red when (1) thetotal major-road volume is less than 200 vehicles per hour(veh/h), or (2) when the ratio of major-road-to-minor-roadvolume is greater than 3.0. This recommendation is illustratedin Figure 2-2 as the unshaded area.

Kacir et al. (12) compared the delays produced by flashingyellow/red with those produced by traffic signal control. Theyfound that delays were significantly reduced when (1) themajor-road-to-minor-road volume ratio is greater than 3.0, (2) the major-road volume is less than 250 vehicles per hourper lane (veh/h/ln), and (3) the higher approach minor-roadvolume is less than 85 veh/h/ln. These recommendationswere incorporated in Figure 2-2 with dashed lines (they are

14

based on an assumed directional distribution of 55/45 percentfor both roadways).

Application. The guidelines stated in the preceding sec-tion (and shown in Figure 2-2) are intended to minimize theoperational impact of a traffic signal and maintain a reason-able degree of safety. This guideline assumes that traffic sig-nal control is a viable alternative (i.e., that one or more signalwarrants have been satisfied). Flash mode should be consid-ered during each hour of the average day for which the majorand minor volumes fall in the unshaded region. If flash modeis used, it should be used during extended periods (i.e., severalhours per period) each day to minimize driver confusion.

Application of this guideline requires two types of data:

1. Major-road and minor-road approach volumes for eachhour of the average day and

2. Major-road and minor-road approach through-lanecount.

These traffic demands can be measured, or they can be esti-mated as a fraction of the average daily traffic demand.Kacir et al. (12) suggest the values listed in Table 2-6 canbe used to estimate demands during the late-night hours.Also, the stopped drivers’ unobstructed view of approach-ing unstopped vehicles should be verified before flash modeis implemented.

Convert to Traffic Signal Control

Introduction. An intersection with a properly designedand operated traffic control signal will have one or more ofthe following benefits (relative to an intersection without atraffic signal):

0

50

100

150

200

250

300

0 200 400 600 800 1000

Major R oadw ay (total o f b oth ap proaches), veh /h

Min

or

Ro

ad

wa

y(t

ota

l of

bo

th a

pp

roa

ch

es

), v

eh

/h

Consider yellow/red flash mode.

2/1 & 2/2

2/2

Yellow/red

safe but

may cause

delay.

Do not cons ider

yellow/red flash mode.

1/1, 2/1, 2/2

Legend: x /y - x = major road lanes, y = minor road lanes

Figure 2-2. Guideline for use of signal flash mode.

• More orderly movement of traffic,• Increased intersection capacity,• Reduced frequency of certain types of collisions (e.g.,

right-angle),• Continuous or nearly continuous movement of traffic

along the through route, and• Reduced delay to minor vehicular and pedestrian

movements by interrupting heavy traffic at periodicintervals.

If the traffic signal is not properly designed or operated, oneor more of the following problems may result:

1. Excessive delay to all traffic movements,2. Excessive disobedience of the signal indication,3. Increased frequency of diversion through neighbor-

hoods, and4. Increased frequency of certain types of collisions (e.g.,

rear-end).

These four problems may also be evident at a signalizedintersection whose traffic signal is no longer needed.

Guidance. The MUTCD 2000 describes eight warrantsthat define conditions in which a traffic control signal is likelyto improve intersection safety, operations, or both. Thesewarrants are listed in Table 2-7. If a warrant is met, then sig-nal control may be appropriate. However, the appropriatenessof the signal should be confirmed through the conduct of anengineering study (as described in Chapter 3).

Application. Collectively, the warrants listed in Table 2-7address a wide range of factors that can affect safety and oper-ations at an intersection. These factors include traffic vol-

15

umes, volume variations during the day, approach geometry,major-road speed, crash frequency, motorist delay, and gapfrequency in the major-road traffic stream. Often, only a sub-set of these warrants are evaluated because only one or two warrants are likely to be sensitive to the problem beingexperienced at the subject intersection.

Once the warrants to be evaluated are identified, the nec-essary data are collected using appropriate field study tech-niques or estimation methods. The data needed to evaluateeach warrant are listed in Table 2-8. Procedures for collect-ing these data are described in the Manual of TransportationEngineering Studies (6 ). Data collection procedures are alsodescribed in “Traffic Signal Warrants: Guidelines for Con-ducting a Traffic Signal Warrant Analysis” (7 ). For pro-posed intersections, techniques described in Appendix C canbe used to estimate turn movement volumes from forecasttraffic demands.

Convert to Multi-Way Stop Control

Introduction. Multi-way stop control is most useful whenintersection traffic volume is high enough to create frequentconflicts and the traffic volume is evenly split between theintersecting roads. Multi-way stop control may reduce thenumber of crashes more than will other, less-restrictive formsof control. Multi-way stop control can also result in less delaythan other types of control when approach demands are nearlybalanced and do not satisfy one or more of the MUTCD 2000Warrants 1, 2, or 3.

Guidance. The MUTCD 2000 (1, p. 2B-10) describes fourcriteria that define conditions in which a multi-way stop islikely to improve intersection safety, operations, or both. If one

TABLE 2-6 Hourly volume levels expressed as a percentage of average daily traffic demand

TABLE 2-7 Signal warrants in the MUTCD 2000

criterion is met, multi-way stop control may be appropriate.However, this finding should be confirmed through the con-duct of an engineering study (as described in Chapter 3). Thefour guidance criteria are as follows:

A. Where traffic control signals are justified, the multi-way stop is an interim measure that can be installedquickly to control traffic while arrangements arebeing made for the installation of the traffic controlsignal.

B. A crash problem, as indicated by five or more reportedcrashes in a 12-month period that are susceptible to cor-rection by a multi-way stop installation. Such crashesinclude right- and left-turn collisions as well as right-angle collisions.

C. Minimum volumes:1. The vehicular volume entering the intersection

from the major-street approaches (total of both ap-proaches) averages at least 300 vehicles per hourfor any 8 hours of an average day, and

2. The combined vehicular, pedestrian, and bicyclevolume entering the intersection from the minorstreet approaches (total of both approaches) aver-ages at least 200 units per hour for the same 8 hours,with an average delay to minor-street vehicular traf-fic of at least 30 seconds per vehicle during thehighest hour, but

3. If the 85th percentile approach speed of the major-street traffic exceeds 65 km/h (40 mph), the mini-mum vehicular volume warrants are 70 percent ofthe above values.

D. Where no single criterion is satisfied, but where Crite-ria B, C.1, and C.2 are all satisfied to 80 percent of theminimum values. Criterion C.3 is excluded from thiscondition.

Option:

16

Other criteria that may be considered in an engineeringstudy include:

A. The need to control left-turn conflicts.B. The need to control vehicle/pedestrian conflicts near

locations that generate high pedestrian volumes.C. Locations where a road user, after stopping, cannot

see conflicting traffic and is not able to safely nego-tiate the intersection unless conflicting cross traffic isalso required to stop.

D. An intersection of two residential neighborhood col-lector (through) streets of similar design and oper-ating characteristics where multi-way stop controlwould improve traffic operational characteristics ofthe intersection.

Application. Collectively, the criteria listed in the pre-ceding section address a wide range of factors that can affectsafety and operations at an intersection. These factors in-clude traffic volume, volume variations during the day, ap-proach geometry, major-road speed, crash frequency, andmotorist delay.

Once the criteria to be evaluated are identified, the nec-essary data are collected using appropriate field study tech-niques or estimation methods. The data needed for each cri-terion are listed in Table 2-9. Procedures for collectingthese data are described in the Manual of TransportationEngineering Studies (6 ).

Convert to Two-Way Stop or Yield Control

Introduction. The Stop sign is intended for intersectionapproaches on which drivers need to stop before proceeding

TABLE 2-8 Data needed to evaluate the MUTCD 2000 signal warrants

into the intersection. A benefit of stop control (over no con-trol) is that it can improve overall intersection safety byclearly defining which movements have to yield the right-of-way. On the other hand, stop control has some disadvantagesthat include the following:

• Increased frequency of some collisions (e.g., rear-endcollision),

• Increased road-user costs through increased fuel con-sumption and delay, and

• Increased air and noise pollution.

The Yield sign is used to inform drivers approaching anintersection that they do not have priority at the intersection butthat stopping is not necessary if they can verify that the inter-section will be clear when they reach it. Yield control is lessrestrictive than stop control, but more restrictive than “no con-trol.” Yield control tends to be associated with more collisionsthan stop control, but has a significantly lower road-user cost.

Guidance. The MUTCD 2000 (1, p. 2B-8) presents fourconditions where stop control may be appropriate:

A. Intersection of a less important road with a main roadwhere application of the normal right-of-way rule wouldnot be expected to provide reasonably safe operation.

B. Street entering a through highway or street.C. Unsignalized intersection in a signalized area.D. High speeds, restricted view, or crash records indicates

a need for control by the STOP sign.

The MUTCD 2000 (1, 2B-12) also offers several condi-tions where yield control may be appropriate. The conditionsapplicable to intersections are as follows:

A. When the ability to see all potentially conflicting traf-fic is sufficient to allow a road user traveling at theposted speed, the 85th percentile speed, or the statutoryspeed to pass through the intersection or to stop in asafe manner.

. . . . . .

17

C. At the second crossroad of a divided highway, wherethe median width is 9 m (30 ft) or greater. A STOP signmay be installed at the entrance to the first roadway ofa divided highway and a YIELD sign may be installedat the entrance to the second roadway.

D. At an intersection where a special problem exists andwhere engineering judgment indicates the problem tobe susceptible to correction by the use of the Yield sign.

With regard to the second Condition “A” above (i.e.,“When the ability. . . .”), the minor-road driver’s view ofthe major road should not be obstructed by curvature,grade, or objects in one or both of the quadrants adjacent tothe minor-road approach. Guidelines for determining theminor-road driver’s sight distance needs are described inChapter 5 of the Manual of Transportation EngineeringStudies (6 ).

Box (14) developed guidelines for use of traffic controlsigns at low-volume urban intersections. He recommendedconsideration of roadway classification, crash history, andthe safe approach speed in determining the most appropriatecontrol mode. The recommendations made by Box have beenincorporated in Table 2-10.

Box (14) indicates that Table 2-10 should only be used forintersections with a total entering traffic volume of 300 veh/hor less during the peak hour. He also cautions that the no-control or yield-control options may not work well when thetotal entering volume exceeds 100 veh/h.

Application. The guidance stated in the preceding sec-tion indicates the conditions suitable to the use of two-waystop or yield control. Using this guidance requires five typesof data:

1. Major- and minor-road approach volumes for the peakhour of the average day;

2. Major-road 85th percentile speed (posted speed can besubstituted if data are unavailable);

3. Minor-road safe approach speed (based on availableintersection sight distance);

TABLE 2-9 Data needed to evaluate the MUTCD 2000 multi-way stop control criteria

4. Major- and minor-road classification; and5. Crash history for the previous 12-months as a minimum,

but preferably for the previous 3 years.

These data would be used with Table 2-10 to determine themost appropriate minor-road control mode. The conditionsfrom the MUTCD 2000 listed (see the preceding Guidancesection) should be assessed in combination with Table 2-10.If any of these conditions appear satisfied or if the conclusionreached from the use of Table 2-10 indicates that stop or yieldcontrol is appropriate, then stop or yield control may be aviable alternative for controlling the minor-road intersectionapproach.

Prohibit On-Street Parking

Introduction. Parking maneuvers into or out of on-streetparking stalls can affect the operation and safety of thethrough traffic lane adversely. The stalls can be either angleor parallel with the curb. The time required for the parkingmaneuver is shorter for the angled stall than for a parallelparking stall. However, when leaving the stall, it is faster toleave a parallel parking stall than to leave an angled stall.Measurements indicate that the total blockage for both man-euvers is about 36 s. Chapter 16 of the Highway CapacityManual (15) indicates that such maneuvers momentarilyblock the adjacent through lane and can significantly reduce

18

intersection capacity if they occur within 76 m (250 ft) of thestop line.

Parking on the intersection approach can also have safetyconsequences. Cleveland et al. (16) report two studies indi-cating that the removal of parking can reduce crash frequencyby 16 to 32 percent. Parked vehicles along one road can alsocreate safety problems for drivers on the intersecting road byblocking the driver’s view of conflicting traffic.

Guidance. Special Report 125: Parking Principles (17)describes guidelines for determining when to prohibit on-street parking. These guidelines indicate maximum flow ratesthat can be associated with on-street parking. Table 2-11summarizes the guidance provided.

Application. The guidance described in the precedingsection describes conditions that may justify the prohibitionof on-street parking. Use of this guidance requires two typesof data:

1. Major- and minor-road approach volumes for 8 or morehours of the average day.

2. Major- and minor-road approach through-lane count.

Table 2-11 should be consulted once for each hour of inter-est on a given intersection approach. Each approach is indi-vidually evaluated. If the combination of volume and lanes

TABLE 2-10 Candidate control for the minor-road approach1

TABLE 2-11 Guidelines for determining when to prohibit on-street parking

exceeds the maximum value listed in the table, then parkingprohibition should be considered a viable alternative for thesubject approach during the specified hour. If it is determinedthat parking should be prohibited during any 1 hr, it wouldbe preferable to prohibit parking during several hours (e.g.,7:00 a.m. to 6:00 p.m.) to minimize enforcement problems.However, parking prohibition during just the peak trafficdemand hours can be considered.

Prohibit Left-Turn Movements

Introduction. Left-turning vehicles at an unsignalizedintersection can create numerous operational and safety prob-lems. The left-turn maneuver tends to require a longer servicetime than the through maneuver. In fact, even modest left-turn volumes can cause safety or operational problems, espe-cially when there is inadequate storage for left-turn vehicles.When the right-of-way needed to provide this storage is notavailable, left-turn restriction (through regulation or chan-nelization) is a means of eliminating these problems. How-ever, the potential benefits of turn restriction should be care-fully weighed against the increased travel time and trip lengththat is likely to be incurred by redirected motorists.

Guidance. Turn restrictions at an intersection should becarefully considered because they can cause traffic to divertto other, local roads. In general, the analyst should ensurethat turn restriction is part of a larger program intended toidentify the ultimate cause of the left-turn problem. Causesof such problems may include inadequate left-turn capacityat adjacent signalized intersections, inadequate left-turn stor-age at the subject intersection, and inadequate arterial travelspeed (such that traffic diverts through neighborhoods via aleft-turn).

The following guidelines can be used to identify the con-ditions suitable for left-turn restriction at existing intersec-tions. These guidelines are based on the criteria offered byKoepke and Levinson (18).

• Left-turn-related delay, conflicts, or crash frequencyshould be at unacceptable levels.

• An alternative route is available for the redirected left-turn vehicles.

• The alternative route is not expected to add more than afew minutes to the redirected motorist’s travel time.

• The intersection is in an urban or suburban area. (Note: insuburban settings, turn restriction is generally not foundexcept where such treatments are part of an areawidecirculation plan.)

If operational problems are only experienced during thepeak hours, then left-turn restriction during these hours maybe a viable alternative. If operational problems exist through-out the day and provision of a left-turn bay (of adequatelength) is not an option, then full-time left-turn restriction

19

may be a viable alternative. Regardless of the duration of therestriction, all four of the above criteria should be satisfiedbefore turn restriction is given further consideration.

Application. The guidance described in the precedingsection can be used to determine if left-turn restriction mightimprove operations or safety at an intersection. Evaluationof this guidance requires three types of data (as measuredduring the morning peak hour, afternoon peak hour, and onerepresentative off-peak hour of the average day):

1. Delay resulting from left-turn vehicles queued in athrough lane because of nonexistent or inadequate baystorage;

2. Left-turn-related traffic conflicts (or left-turn-relatedcrash history for the previous year); and

3. Travel time for the likely alternative route(s).

The left-turn-related delay, conflict, and crash data shouldbe used to determine if turn restriction is needed and, ifneeded, for what hours of the day. When turn restrictions onlyneed to be in place during certain times of the day, the turnrestriction should be shown by (1) a variable message sign,(2) an internally illuminated sign whose legend is visibleonly during the hours where the prohibition is applicable, or(3) permanently mounted signs with a supplementary legendstating the hours when the prohibition is in effect. If turnrestriction is needed throughout the day, then raised-curbchannelization should be considered to maximize compliance.

Convert to Roundabout

Introduction. The modern roundabout can offer opera-tional and safety benefits that exceed those offered by otherforms of intersection control for certain conditions. Robin-son et al. (19) indicate that roundabouts have the followingadvantages:

1. They are effective traffic-calming devices because theyreduce speeds.

2. They can reduce the frequency and severity of somecrashes (e.g., right-angle, head-on, and turn-related).

3. They result in less delay than multi-way stop control.4. They result in less delay than two-way stop control when

volumes are sufficient to cause operational problems atthe two-way stop-controlled intersection.

5. They result in less delay than signal control when vol-umes do not exceed the roundabout’s capacity.

The roundabout also has some characteristics that may pre-clude its use at some locations. These characteristics are asfollows:

1. Roundabouts may require more right-of-way than aconventional intersection.

2. They may not be easily traversed by large or over-sizedtrucks.

3. Roundabouts may not yield efficient operation if themajor-road volume greatly exceeds the minor-roadvolume.

4. They may not be conducive to through movement pro-gression in coordinated signal networks.

5. They may not be conducive to serving pedestrian orbicycle traffic.

Guidance. The report prepared by Robinson et al. (19)provides some information on the conditions where modernroundabouts are well suited. These conditions have been re-stated in Table 2-12 in terms of questions to be answered whenconsidering a roundabout. The questions have been worded sothat the roundabout becomes a more viable alternative as thenumber of questions answered “Yes” increases.

The maximum daily service volume needed to answerQuestion 2 in Table 2-12 can be obtained from Figure 2-3.This figure was developed by Robinson et al. (19) and as-sumes that (1) the peak hour has 10 percent of the daily vol-ume, (2) one direction of flow on each road has 58 percent ofthe total two-way flow, (3) 10 percent of each approach vol-ume is turning right, and (4) the maximum service volumeequates to 85 percent of capacity. Figure 2-3 applies to a four-leg roundabout; if applied to a three-leg roundabout, the val-ues obtained from the figure should be multiplied by 0.75. Anequation is offered in the footnote to Table 2-12 to facilitatethe computation of maximum service volume.

Application. The guidance stated in the preceding sec-tion indicates the conditions in which a roundabout may bedesirable. Application of this guidance requires four typesof data:

1. Major- and minor-road approach volumes for theaverage day;

2. Major- and minor-road turn movement volumes forthe average day (used to compute average left-turnpercentage);

20

3. Major- and minor-road approach sight distance; and4. Major- and minor-road pedestrian, bicycle, and heavy

vehicle volumes for the average day.

Any evaluation of operational benefits derived as a resultof the roundabout should consider operations throughout theday because the roundabout offers significant operationalbenefit (over stop or signal control) during off-peak hours.

Add a Second Lane on the Minor Road

Introduction. The quality of service provided to the minor-road movements at a two-way stop-controlled intersection isdependent on the number of lanes provided. Typically, onelane is shared by all movements on the minor-road approach;however, sometimes a turn bay or second through lane isadded to reduce the delay to selected movements. A secondlane can be added as a right-turn bay, a left-turn bay, or a sec-ond shared lane (i.e., through plus left in the inside lane andthrough plus right in the outside lane). Of these options,adding a right-turn bay is often the most effective as it cansignificantly reduce the delay to the right-turn movement; it

TABLE 2-12 Worksheet to determine the suitability of modern roundabout

Figure 2-3. Maximum daily service volumes for a four-leg roundabout.

can also indirectly reduce the delay to the left-turn or throughmovements by lessening their need to compete for servicewith the right-turn movement.

One disadvantage of adding a lane to the minor-road ap-proach is that it may require reallocating the existing pave-ment or widening of the approach cross section. Sometimesthe pavement width needed for the additional lane is availablewithin the existing roadway cross section. In this instance, theonly impact is a reallocation of the paved surface throughmodification of the pavement markings. However, in down-town settings this reallocation may require the removal ofsome curb parking stalls and can affect adjacent business sig-nificantly. Occasionally, the cross section must be widened toprovide for the additional lane. If the needed lane width canbe provided within the available right-of-way, the cost maybe limited to that of construction. However, if additionalright-of-way is needed, the costs of acquiring this property inurban settings can be high.

Guidance. The literature does not offer guidance regard-ing conditions where a second approach lane would benefitfrom the operation of a minor-road approach. However, theprocedures in Chapter 17 of the Highway Capacity Manual2000 (15) can be used to identify major- and minor- road vol-ume combinations that would benefit operationally from theprovision of a second approach lane or bay. Bonneson andFontaine (20) developed Figure 2-4 using these proceduresand an assumed upper limit of 0.7 for the shared-lane, minor-road volume-to-capacity ratio.

Application. Figure 2-4 indicates the conditions that mayjustify the use of two approach lanes. Use of the informationin this figure requires two types of data:

1. Major-road approach volume for the peak hour of theaverage day and

2. Minor-road turn movement volume for the peak hour ofthe average day (used to compute right-turn percentage).

21

Figure 2-4 would be used once for each minor-road ap-proach to the intersection. The appropriate trend line wouldbe identified on the basis of the percentage of right-turns onthe subject minor-road approach. If the volume combinationfor the major and minor roads intersects above or to the rightof this trend line, a second traffic lane should be consideredfor the subject minor-road approach. If a bay is selected foraddition to the intersection, it should be long enough to storevehicles 95 percent of the time (i.e., the bay should not over-flow more than 5 percent of the time). Techniques for esti-mating the 95th percentile storage length are provided in thesection, Increase the Length of the Turn Bay.

Add a Left-Turn Bay on the Major Road

Introduction. Provision of a left-turn bay on the majorroad to a two-way stop-controlled intersection can signifi-cantly improve operations and safety at the intersection. Aleft-turn bay effectively separates those vehicles that areslowing or stopped to turn from those vehicles in throughtraffic lanes. This separation minimizes turn-related crashesand eliminates unnecessary delay to through vehicles. Datareported by Neuman (21) indicate that the crash rate forunsignalized intersections can be reduced by 35 to 75 percentthrough the provision of a left-turn bay.

One disadvantage of adding a bay to the major-road ap-proach is that it may require reallocating the existing pave-ment or widening of the approach cross section. Sometimesthe pavement width needed for the additional lane is availablewithin the existing roadway cross section. However, in down-town settings this reallocation may require the removal ofsome curb parking stalls and can affect adjacent business sig-nificantly. Occasionally, the cross section must be widened toprovide for the turn bay. If the needed width can be providedwithin the available right-of-way, the cost may be limited tothat of construction. However, if additional right-of-way isneeded, the costs of acquiring this property in urban settingscan be high.

Guidance. Neuman (21) suggests that the followingguidelines should be used to determine when to provide aleft-turn bay on the major road of a two-way stop-controlledintersection:

1. A left-turn lane should be considered at any mediancrossover on a divided, high-speed road.

2. A left-turn lane should be provided on the unstoppedapproach of a high-speed rural highway when it inter-sects with other arterials or collectors.

3. A left-turn lane is recommended on the unstoppedapproach of any intersection when the combination ofintersection volumes intersect above or to the right ofthe appropriate trend line shown in Figure 2-5.

Figure 2-4. Guideline for determining minor-road ap-proach geometry at two-way stop-controlled intersections.

Application. The guidance stated in the preceding sec-tion defines the conditions that may justify the provision of aleft-turn bay. Application of this guidance requires two typesof data:

1. Major-road turn movement volume for the peak hourof the average day and

2. Major-road 85th percentile speed (posted speed can besubstituted if data are unavailable).

Use of Figure 2-5 requires determination of the opposingvolume, the advancing volume, and the operating speed. Theopposing volume should include only the right-turn andthrough movements on the approach across from (and head-ing in the opposite direction of) the subject major-road ap-proach. The advancing volume should include the left-turn,right-turn, and through movements on the subject approach.The operating speed can be estimated as the 85th percentilespeed. If the operating speed does not coincide with 60, 80,or 100 km/h (i.e., 40, 50, or 60 mph), then interpolation can

22

be used or, as a more conservative approach, the operatingspeed can be rounded up to the nearest speed for which afigure is provided.

In application, Figure 2-5 is used once for each major-roadapproach to the intersection. The appropriate trend line isidentified on the basis of the percentage of left-turns on thesubject major-road approach. If the advancing and opposingvolume combination intersects above or to the right of thistrend line, a left-turn bay should be considered for the subjectapproach. If a bay is included at the intersection, it should belong enough to store left-turn vehicles 99.5 percent of thetime (i.e., the bay should not overflow more than 0.5 percentof the time). Techniques for estimating this storage length areprovided in the section, Increase the Length of the Turn Bay.

Add a Right-Turn Bay on the Major Road

Introduction. Provision of a right-turn bay on the majorroad to a two-way stop-controlled intersection can signifi-

Figure 2-5. Guideline for determining the need for a major-road left-turn bay at a two-way stop-controlled intersection.

cantly improve operations and safety at the intersection. Aright-turn bay effectively separates those vehicles that areslowing or stopped to turn from those vehicles in the throughtraffic lanes. This separation minimizes turn-related colli-sions (e.g., angle, rear-end, and same-direction-sideswipe)and eliminates unnecessary delay to through vehicles.

One disadvantage of adding a bay to the major-road ap-proach is that it may require reallocating the existing pave-ment or widening of the approach cross section. Sometimesthe pavement width needed for the additional lane is availablewithin the existing roadway cross section. However, in down-town settings this reallocation may require the removal ofsome curb parking stalls and can affect adjacent business sig-nificantly. Occasionally, the cross section must be widened toprovide for the turn bay. If the needed width can be providedwithin the available right-of-way, the cost may be limited tothat of construction. However, if additional right-of-way isneeded, the costs of acquiring this property in urban settingscan be high.

Guidance. Hasan and Stokes (22) developed guidelinesfor determining when to provide a right-turn bay on the majorroad of a two-way stop-controlled intersection. These guide-lines were based on an evaluation of the operating and colli-sion costs associated with the right-turn maneuver relative tothe cost of constructing a right-turn bay. The operating costsincluded those of road-user fuel and delay. Separate guide-lines were developed for two-lane and four-lane roadways.These guidelines are shown in Figure 2-6.

Application. The guidance described in the preceding sec-tion defines conditions that may justify the provision of aright-turn bay. Application of this guidance requires two typesof data:

1. Major-road turn movement volume for the peak hourof the average day and

2. Major-road 85th percentile speed (posted speed can besubstituted if data are unavailable).

Figure 2-6 should be consulted once for each major-roadapproach. If the combination of major-road approach volumeand right-turn volume intersects above or to the right of thetrend line corresponding to the major-road operating speed,then a right-turn bay is a viable alternative.

23

Increase Length of Turn Bay

Introduction. Turn bay length can affect the safety andoperation of the intersection approach significantly. Thiseffect becomes more negative as the frequency with whichvehicles exceed the available storage increases. Also, forunstopped approaches, this effect becomes more negative asmore of the turning vehicle’s deceleration occurs in thethrough lane, prior to the bay. The need to provide adequatestorage length, deceleration length, or both is dependent onthe type of approach control used and whether the vehicle isturning left or right. Table 2-13 identifies the appropriate bay

Figure 2-6. Guideline for determining the need for amajor-road right-turn bay at a two-way stop-controlledintersection.

TABLE 2-13 Turn-bay length components at unsignalized intersections

length components for typical combinations of approachcontrol and turn movement.

Guidance. A turn bay should be long enough to storewaiting vehicles almost all of the time and, in doing so, pre-vent conflict between the through and turning movements. Abay equal in length to the 95th percentile queue (which allowsfor a 5-percent probability of overflow) is typically used formost stop- or signal-controlled approaches. A bay length cor-responding to a 0.5-percent probability of overflow is rec-ommended by Harmelink (23) for unstopped approaches inrecognition of the greater speed differential between throughand turning movements.

Figures 2-7 and 2-8 illustrate the combination of volumeand storage length that limit overflows to 0.5 and 5 percent,respectively. The trends in these figures are based on thequeue length equation reported by Harmelink (23). The ca-pacity needed for this equation was estimated using the pro-cedures in Chapter 17 of the Highway Capacity Manual2000 (15). The development of Figures 2-7 and 2-8 is de-scribed elsewhere by Bonneson and Fontaine (20). A prac-tical minimum storage length of 8 m (25 ft) is reflected inboth figures.

As indicated in Table 2-13, a bay on the unstopped ap-proach to an intersection should also provide sufficient lengthfor the turning vehicle to decelerate once it is clear of theadjacent through traffic lane. A report by Koepke (24) pro-vides an indication of the bay length needed for turn vehicledeceleration. However, the lengths cited by Koepke includeboth deceleration length and bay taper length. Subtraction ofthe taper length [estimated at 35 m (120 ft)] from the reportedlengths leaves a conservative estimate of the length of full-width turn bay needed to facilitate the deceleration process.This length is shown in Figure 2-9.

Application. The guidelines described in the precedingsection can be used to determine when turn-bay length may

24

need to be increased to minimize operational or safety prob-lems associated with short turn bays. Application of thisguideline requires three types of data:

1. Major- and minor-road turn movement volumes for thepeak hour of the average day;

2. Major-road 85th percentile speed (posted speed can besubstituted if data are unavailable); and

3. Major- and minor-road bay lengths (taper length shouldbe excluded).

In application, each turn movement is considered sepa-rately. Table 2-13 is consulted first to determine if storagelength, deceleration length, or both should be provided withinthe bay for the subject approach. Next, Figures 2-7, 2-8, and2-9 are consulted. The subject movement volume and con-flicting volumes should be used to determine the minimumlength(s) needed from the appropriate figure(s). The operat-

Figure 2-7. Guideline for determining if the bay storagelength for an unstopped approach is adequate.

Figure 2-8. Guideline for determining if the bay storagelength for a stopped approach is adequate.

Figure 2-9. Guideline for determining if the decelerationlength component of a bay is adequate.

ing speed can be estimated as the 85th percentile speed. Theconflicting volume should reflect those traffic streams thathave priority over the subject movement and that cross itstravel path within the intersection [see Chapter 17 of theHighway Capacity Manual 2000 (15) for a more detailed def-inition of conflicting volume]. If the available bay length isless than the computed minimum length, then an increase inbay length is a viable alternative.

Increase the Right-Turn Radius

Introduction. The radius of the right-turn path can affectthe operation and safety of an intersection. Larger radii havethe following advantages:

• On the major-road approach, they allow for higher speedturns and by that, reduce delay to following throughvehicles.

• On the minor-road approach, they widen the throat ofthe approach and effectively act as a short right-turn bay.

• They allow large trucks and buses to turn withoutencroachment into adjacent lanes or onto adjacentproperty.

Larger radii also have disadvantages. These disadvantagesinclude the following:

• They require more right-of-way.• They may reduce pedestrian safety through increased

crossing distance and higher turn speeds.

The second disadvantage noted for large radii can be min-imized through the inclusion of island channelization onthe outside of the right-turn path. This island would have atriangular shape and serve as a pedestrian refuge.

Guidance. The AASHTO document, A Policy on Geo-metric Design of Highways and Streets (Green Book) (5) pro-vides general guidance on the minimum radii needed at inter-sections in suburban or urban areas. These guidelines arereproduced in Table 2-14.

Larger radii should be used when right-of-way is availableand pedestrian volumes are minimal. If larger radii are desired

25

and right-of-way is restricted, then a simple-curve-radius-with-taper or a 3-centered curve may be used to increase the“effective” radius without using as much right-of-way as asimple radius. Desirable geometrics for these curves are de-scribed in Tables IX-1 and IX-2 of the Green Book (5). Iflarger radii are desired and pedestrian volumes are sig-nificant, then island channelization should be provided. TheGreen Book lists several 3-centered curve designs in TableIX-4 that produce islands meeting the minimum island sizerequirements.

Hasan and Stokes (22) developed guidelines for deter-mining when to provide a simple-curve-radius-with-taperdesign on the major road of a two-way stop-controlled inter-section. These guidelines were based on an evaluation of theoperating and collision costs associated with the right-turnmaneuver relative to the cost of constructing a right-turnbay. The operating costs included those of road-user fueland delay. Separate guidelines were developed for two-laneand four-lane roadways. These guidelines are shown inFigure 2-10.

Application. The guidance provided in the preceding sec-tion can be used to determine if an intersection has adequateright-turn radius design. Application of the Green Book (5)guidance requires three types of data:

1. Heavy vehicle volume during the peak hour of theaverage day;

2. Major- and minor-road classification; and3. Major- and minor-road right-turn radius, measured to

the edge of the traveled way.

As a first step, the engineer should decide whether a sim-ple curve radius or a more complicated curve design is ap-propriate for the subject intersection. In general, a simple curveradius is appropriate for low-speed approaches to urban inter-sections; otherwise, a simple-curve-radius-with-taper or a3-centered curve design should be considered.

If a simple curve radius is deemed appropriate, Table 2-14is consulted once for each intersection approach. If the exist-ing right-turn radius on a given approach does not equal orexceed the value listed, then increasing the right-turn radiusshould be considered a viable alternative.

TABLE 2-14 Guideline for determining the need for a larger simple radius

If a simple-curve-radius-with-taper or 3-centered curve isappropriate, the guidance developed by Hasan and Stokes (22)should be consulted. Use of this guidance requires two typesof data:

• Major-road turn movement volume for the peak hour ofthe average day.

26

• Major-road 85th percentile speed (posted speed can besubstituted if data are unavailable).

In application, Figure 2-10 is consulted once for eachmajor-road approach. If the combination of major-roadapproach volume and right-turn volume intersects above or tothe right of the trend line corresponding to the major-roadoperating speed (estimated as the 85th percentile speed), thena simple-curve-radius-with-taper or a 3-centered curve shouldbe considered a viable alternative.

Conditions Affecting the Accuracy ofConclusions from the Signal Warrant Check

Overview

The MUTCD 2000 (1) signal warrants identify thresholdtraffic volumes below which an intersection is not likely tobe amenable to traffic signal control. When a check of thewarrants reveals that volumes exceed these thresholds, thenan engineering study should be conducted to verify whetherthe signal will improve intersection safety or operations. Thestudy is needed because the traffic or geometric conditions atthe subject intersection may differ enough from those as-sumed in the warrants that the results of the warrant checkwould be inaccurate.

Table 2-15 lists common problematic conditions. Theseconditions are described in this section. Included in thisdescription is information that can be used to determine if thecondition is likely to exist at the subject intersection. If it isdetermined that a problematic condition exists, then the effectof the condition on intersection operations should be fullyevaluated during the engineering study stage.

Right-Turn Volume on the Minor Road

Minor-road right-turn vehicles at an unsignalized inter-section incur less delay than the left-turn or through vehi-

TABLE 2-15 Problematic traffic, signalization, and geometric conditions

Figure 2-10. Guideline for determining the need for asimple-curve-radius-with-taper.

cles. Thus, an intersection where minor-road drivers are pri-marily turning right is less likely to derive benefit from sig-nalization than one where most drivers are crossing throughor turning left.

The effect of minor-road right-turns is also recognized inSection 4C.01 of the MUTCD 2000 (1). For shared-lane ap-proaches, the MUTCD 2000 indicates that a portion of theright-turning volume should be subtracted from the minor-road volume when evaluating the volume-based signal war-rants. For minor-road approaches with an exclusive right-turnbay and adequate right-turn capacity, the MUTCD 2000 indi-cates that all of the right-turn volume can be subtracted fromthe minor-road volume and that the right-turn bay should notbe included in the count of approach lanes. Without theseadjustments, it is likely that an intersection with moderate-to-high right-turn volume will yield inaccurate conclusionsfrom the warrant check.

The following method can be used to determine if right-turnvolume on the minor road could influence the warrant check.The method is based on signal warrant analysis guidelinesdeveloped by the Utah Department of Transportation (25).Using this method, the actual right-turn volume is reduced onthe basis of consideration of the major-road volume that con-flicts with the right-turn movement, the number of traffic lanesserving the conflicting volume, and the geometry of the subjectminor-road approach. The relationship between these factors isillustrated in Figure 2-11.

To determine if a heavy right-turn volume might misleadthe warrant check, a second warrant check can be performedusing an adjusted minor-road volume. This adjusted volumeis computed as follows:

with,

where

Vi = volume for movement i (movement numbers areshown in Figure 2-12), veh/h;

Vr9 = right-turn volume reduction for movement 9 (obtainedfrom Figure 2-11 using conflicting major-road volumeVc9), veh/h;

Vr12 = right-turn volume reduction for movement 12(obtained from Figure 2-11 using conflicting major-read volume Vc12), veh/h;

Vc9 = conflicting major-road volume for movement 9;veh/h;

Vc12 = conflicting major-road volume for movement 12,veh/h; and

Ni = number of approach lanes serving through move-ment i.

V VV

Nc12 6

5

5

0 5= +.

V VV

Nc9 3

2

2

0 5= +.

Adjusted MinorLarger of:Road Volume = + + −

+ + −[ ]V V V VV V V V

r

r

7 8 9 9

10 11 12 12

27

The volume reduction cannot exceed the corresponding right-turn volume (i.e., V −V9 r9 ≥ 0 and V12−Vr12 ≥ 0). Also, the“Right-turn bay provided” case in Figure 2-11 can be used forshared-lane approaches when the shared lane functions as ade facto right-turn lane.

If the second warrant check yields different conclusionsthan the first warrant check (i.e., the check based on un-adjusted volumes), then right-turn volumes may be enough toaffect the accuracy of the warrant check. In this situation, it isrecommended that the effect of right-turns be fully examinedduring the engineering study.

Heavy Vehicles on the Minor Road

Heavy vehicles on the minor-road approach to an un-signalized intersection should logically increase overall delay.This effect would result from the relatively large gaps that thedrivers of such vehicles need to cross or enter the major-roadtraffic stream safely. Thus, an intersection with a significantnumber of heavy vehicles is likely to derive more benefit fromsignalization than one where there are fewer heavy vehicles.

A study conducted by Henry and Calhoun (26, p. 31) re-vealed that the volume of heavy vehicles must be “unusuallyhigh” to have a significant impact on the warrant evaluationprocess. Their report indicates that the heavy-vehicle per-centage would have to exceed 5 percent during the peak traf-fic hour to constitute “unusual” conditions. Therefore, if thepercentage of heavy vehicles exceeds 5 percent or if the ana-lyst is uncertain of the effect of heavy vehicles (perhaps whencombined with other factors such as grade) on intersectionoperations, then the effect of heavy vehicles should be fullyevaluated during the engineering study.

Pedestrian Volume

Heavy pedestrian volume, as may be found in downtownareas, may significantly increase vehicular delays at unsignal-ized intersections. This increase stems from the additional con-

Figure 2-11. Minor-road right-turn volume reductionfor warrant check.

flict between vehicular and pedestrian flows at the intersection.Because of this interaction, an intersection with heavy pedes-trian volume is likely to derive more benefit from signalizationthan one where there are fewer pedestrians.

The Highway Capacity Manual 2000 (15) procedure fortwo-way stop-controlled intersection analysis accounts for pe-destrians by including them in the number of events (vehicleor pedestrian) that conflict with the non-priority movements.This approach can also be used to determine if pedestrianvolume might mislead the warrant check. Specifically, it issuggested that a second warrant check be performed with thefollowing adjusted volumes:

where

V = movement volume shown in Figure 2-12.i

Also, if the adjusted minor-road volume is based on a pedes-trian volume, then the approach should be considered to haveone approach lane.

If the second warrant check yields different conclusionsthan the first warrant check (i.e., the check based on unad-justed volumes), then pedestrian volumes may be enough toaffect the accuracy of the warrant check. In this situation, it isrecommended that the effect of pedestrians be fully examinedduring the engineering study.

Adjusted Major Road Volume

Adjusted Minor Larger of: Road Volume

= + + + + + + +

= + + + +

V V V V V V V V

V V V V V VV V

1 2 3 4 5 6 15 16

7 8 9 10 11 12

13 14

( , ,, )

28

Progressive Traffic Flow on the Major Road

Signal coordination promotes efficient traffic movementalong the roadway by providing for the uninterrupted progres-sion of through vehicle platoons. These platoons affect theoperation of an unsignalized intersection. If platoons arrivesimultaneously, there is a positive effect. Conversely, if pla-toons arrive in an alternating pattern, the effect is negative. Theextent of the effect typically increases as the distance betweenintersections decreases.

Chapter 17 of the Highway Capacity Manual 2000 (15) in-dicates that the effect of an upstream signal decreases as thedistance between it and the subject intersection increases; theeffect is negligible for distances in excess of 400 m (1,300 ft).Based on this guidance, platoon progression may affect theaccuracy of the conclusions from the warrant check when thesubject intersection is within 400 m (1,300 ft) of an upstreamsignalized intersection. If the analyst is uncertain of the effectof progression on the existing unsignalized intersection, thenthe effect of platoon progression should be fully evaluatedduring the engineering study.

Three-Leg Intersection

The performance of a two-way stop-controlled intersectioncan be greatly affected by the number of minor-road approachlegs. A four-leg intersection has higher delay per vehicle thanone with three legs because of the increased number of con-flicting movements. Therefore, the benefit derived by signal-

Figure 2-12. Movement numbers for right-turn volume adjustment.

izing a three-leg intersection may be substantially less thanthat derived by signalizing a four-leg intersection.

A simulation study conducted by Saka (27) indicated thatMUTCD 2000 Signal Warrant 1 (Condition A) has thresholdvolumes that are relatively low when applied to three-leg inter-sections. Saka recommended that a minimum major-road vol-ume of 1,000 veh/h (total of both directions) and a minimumminor-road volume of 200 veh/h should be sustained for the 8highest hours of the average day before the traffic signal alter-native is considered for a three-leg intersection. If this volumelevel is not sustained, but Signal Warrant 1 is satisfied, then itis possible that the conclusion from the warrant check is in-accurate. In this situation, the operation of the three-leg inter-section should be fully evaluated during the engineering study.

Added Through Lane on the Major Road

The outside through lane should be sufficiently long inadvance of and beyond the intersection that it is attractive tothrough drivers. A through lane will not be fully used whenit extends only a short distance before or after an intersection.Instead, only turning drivers may use it. The consequencesof this problem are significant if the analyst misjudges theutilization of the second lane of a two-lane approach.

Guidelines provided in Chapter 10 of the Highway Capac-ity Manual 2000 (15) can be used to determine how a lane mayfunction at a given intersection. These guidelines are based onthe length of the lane, as measured along the intersection’sapproach and departure legs. The guidance is illustrated inFigure 2-13.

Three regions are shown in Figure 2-13. The region labeled“Full lane” indicates combinations of approach and departurelane length that should allow the traffic lane to operate effec-tively as a through lane in the vicinity of the intersection. Theremaining two regions indicate lane lengths that would yieldpartial or no use of the lane by through drivers.

29

The number of through lanes used in the warrant checkshould reflect fully utilized lanes to avoid inaccurate conclu-sions. If the number of lanes used in the warrant check doesnot reflect “fully used” lanes (as suggested by Figure 2-13),then it is possible that the conclusions reached from thewarrant check will be inaccurate. Regardless, if the analystis uncertain as to the effective number of through lanes atan intersection, then the effect of lane use should be fullyevaluated during the engineering study.

Left-Turn Bay

The operation of the major- and minor-road traffic move-ments is affected by the approach lane assignments. Provisionof an exclusive left-turn bay (of adequate length) separates theleft-turn movement from the through movement and generallyimproves the operation of all movements, relative to a shared-lane arrangement. Therefore, an intersection that has sharedlanes is likely to derive more benefit from signalization thanan intersection that has one or more left-turn bays.

Guidance is provided in Section 4C.01 of the MUTCD2000 (1) as to how the effect of a left-turn bay can beaddressed in the warrant check. For left-turn bays that serveabout 50 percent of the approach traffic, it is suggested thatthe left-turn bay be included in the count of major-road (orminor-road) lanes. In contrast, for left-turn bays that serve aminor percentage of approach traffic, the MUTCD 2000 (1)suggests that the left-turn bay should not be included in thelane count. In either case, the left-turn volume is included inthe approach volume estimate.

Application of this guidance should improve the likeli-hood of accurate conclusions from the warrant check. How-ever, if the analyst is uncertain whether to include the left-turn bay in the lane count, then the effect of the left-turn bayshould be fully evaluated in an engineering study. Similarly,if an existing left-turn bay is believed to be too short, then itis possible that the conclusions reached from the warrantcheck will be inaccurate. In this situation, the benefit of thebay should be fully evaluated during an engineering study.

Right-Turn Bay on the Minor Road

The provision of an exclusive right-turn bay (or the avail-ability of sufficient approach width to facilitate operationof a de facto exclusive right-turn bay) can significantlyimprove the operation of the minor-road right-turn move-ment, relative to a shared-lane arrangement. Therefore, anintersection that has shared lanes is likely to derive morebenefit from signalization than an intersection with anexclusive right-turn bay on the minor-road approach. Amethod for determining whether the provision of a right-turn bay will affect the accuracy of the warrant check wasdescribed earlier in this chapter in the subsection, Right-Turn Volume on the Minor Road.

Figure 2-13. Effect of a lane’s length on its utilization bythrough drivers.

Wide Median on the Major Road

An unsignalized intersection with a major-road medianof sufficient width to accommodate (or store) one or morecrossing vehicles may operate as two separate intersections,each with a one-way major road. The frequency of such“two-stage” crossing maneuvers increases with increasingmedian width and with the willingness of drivers to cross intwo stages.

The procedures in Chapter 17 of the Highway CapacityManual 2000 (15) indicate that two-stage operation can dou-ble the capacity of the minor-road traffic movements (evenwhen the median is only 5 m in width). Therefore, in areaswhere drivers frequently cross in stages, an intersection witha wide median is less likely to benefit from signalization thanone where there is no median.

Intersections with median widths of 5 m (16 ft) or moremay be associated with inaccurate warrant check conclusionsif two-stage crossing behavior occurs with any frequency.

30

Therefore, when such conditions are believed to exist, thepotential for and the effect of two-stage crossings should befully evaluated during the engineering study.

Combination of Factors

The preceding discussion of problematic conditions isfocused on describing individual factors that may influ-ence the accuracy of the warrant check conclusions. Thediscussion associated with each condition was primarilybased on the assumption that only that condition would bepresent at the subject intersection. Additional care shouldbe taken when two or more problematic conditions arepresent at an intersection because interactions among theseconditions could amplify their effect on traffic operations.More important, the existence of two or more problematicconditions heightens the need for a thorough evaluationduring the engineering study.

31

CHAPTER 3

ENGINEERING STUDY

This chapter describes the engineering study stage of theengineering assessment process. During this stage, the vari-ous alternatives identified in the alternative identification andscreening stage are evaluated in terms of their performance.Those alternatives that offer some level of improvement, rel-ative to the existing intersection, are then advanced to thealternative selection stage. The alternative selection stage isthe subject of the next chapter.

A procedure for evaluating the operational efficiency ofintersection improvement alternatives is presented in this chap-ter. The presentation is divided into two sections. In the firstsection, the procedural steps involved in the operational assess-ment process are defined. The second section provides guide-lines for designing the signalized intersection alternative andfor using stochastic simulation models. The assessment of analternative’s other impacts (e.g., safety) is an important elementof a comprehensive engineering study; however, procedures toguide these assessments are outside the scope of this guide.

PROCESS

Overview

The procedure for conducting an operational evaluation ofan intersection (and its potential improvement alternatives)consists of three steps. These steps, described in this chapter;are as follows:

1. Determine study type,2. Select analysis tool, and3. Conduct evaluation.

In the first step, issues are addressed that define the natureof the study to be conducted. In the second step, the analystidentifies an analysis tool on the basis of consideration of traf-fic, geometry, and control conditions at the subject intersec-tion and its alternatives. Finally, in the third step, the analystdevelops the alternatives to a preliminary design level andthen uses the selected analysis tool to quantify the operationalperformance of the subject intersection and each alternative.

The objective of the engineering study stage is to determineif a proposed improvement alternative is justified. An alter-native is justified if an analysis indicates that (1) it providesa solution to the problem that precipitated the need for the

study and (2) all traffic movements will operate safely andefficiently. The objective of determining if a proposed im-provement alternative is justified or not is achieved througha process of designing and evaluating the viable alterna-tives identified in the alternative identification and screen-ing stage. The steps involved in this process are describedin the remainder of this section.

Step 1. Determine Study Type

The first step in the engineering study stage requires theconsideration of two issues that characterize the type of studyneeded for the subject intersection. A procedure for resolv-ing each issue is described in a separate task within this step.The two tasks are

a. Determine type of operation andb. Determine type of evaluation.

In the first task, the relationship between the subject inter-section and any upstream or downstream signalized intersec-tion is evaluated. Then, in the second task, the complexity ofeach alternative is assessed, as is the corresponding level ofanalytical detail needed for its evaluation. Together, thesedeterminations are used by the analyst to determine the bestbalance between the thoroughness of the study and the degreeof certainty needed in its findings.

1-a. Determine Type of Operation

Overview. The first step in the engineering study is tocharacterize the effect of an upstream signalized intersectionon the operation of the subject intersection. The existence ofnearby signalized intersections adds a level of complexity tothe analysis process because the platoons from these inter-sections can affect the operation of the subject intersection.Moreover, if a signal alternative is proposed for the subjectintersection, then signal timing strategies that promote pro-gression must also be included in the development of thisalternative.

Isolated or Non-Isolated. The nature of the evaluation tobe conducted is dependent on whether the subject intersection

operates in isolation or within a signal system. The evaluationof an intersection that lies within a signal system (i.e., a non-isolated intersection) requires consideration of the effects of theupstream signalized intersections. In contrast, the evaluationof an isolated intersection does not have this requirement.

Figure 3-1 can be used to characterize the type of opera-tion at the subject intersection. This figure was developed byBonneson and Fontaine (20) and is based on the “couplingindex” concept described by Orcutt (28) and by Hook andAlbers (29). They used this concept to define road segmentvolumes and lengths that are likely to benefit from signalcoordination. It is used in this document to define the degreeof isolation at the subject intersection.

Use of Figure 3-1 requires identification of the road segmentlength and volume. The road segment length is the distancebetween the subject intersection and the upstream signalizedintersection. The road segment volume is the peak-hour two-way volume flowing between these two intersections. All al-ternatives can be considered as “isolated” if the road segmentlength exceeds 1600 m (5,200 ft).

Three regions are identified in Figure 3-1. Each region char-acterizes traffic platoons in terms of their size and concentra-tion. The region labeled as “Sizeable platoons” reflects veryconcentrated and lengthy platoons on relatively short roadsegments. If the subject intersection is associated with volumeand length values that fall in this region, then it should be con-sidered a non-isolated intersection. All alternatives associatedwith this intersection should reflect consideration of platooneffects. If the traffic signal alternative is being considered, itshould be coordinated with the upstream signal.

The region labeled as “Moderate platoons” in Figure 3-1reflects the likelihood of sizeable platoons over a wide rangeof segment lengths. If the subject intersection is associatedwith volume and length values that fall in this region, then itshould be considered a non-isolated intersection. The effectof traffic platoons on unsignalized intersection operation ismore subtle in this region than that associated with the “Size-able platoon” region. Hence, alternatives associated with un-signalized control should not require consideration of platoon

32

effects. However, if the traffic signal alternative is being con-sidered, it should be coordinated with the upstream signal.

Finally, the region labeled as “Small platoons” in Figure 3-1reflects the likelihood of short and dispersed platoons onlow-volume roads. If the subject intersection is associatedwith volume and length values that fall in this region, thenit should be considered an isolated intersection. Alternativesassociated with the subject intersection do not need to reflectconsideration of platoon effects. If the traffic signal alternativeis being considered, it does not need to be coordinated with theupstream signal for the purpose of the engineering study.

In application, all approaches to the subject intersection thathave upstream signals would be checked using Figure 3-1. Ifany one approach is found to be “non-isolated,” then the inter-section should be evaluated as non-isolated. In addition, if theagency wants to include coordination with the signal alterna-tive (independent of the guidance from Figure 3-1), then thesubject intersection should be evaluated as non-isolated.

1-b. Determine Type of Evaluation

Overview. During this task, the engineer determines thetype of evaluation needed for the engineering study. Twotypes of evaluation are possible: (1) the formal engineeringstudy process as described in the remainder of this chapterand (2) the deployment and field evaluation of a proposedalternative. This latter type of evaluation is referred to as“informal” because it does not follow the formal engineeringstudy process. The conditions where an informal evaluationis appropriate are described in the next section.

Formal Versus Informal Evaluation. The informal (orfield) evaluation of an alternative represents a practical methodof assessing an alternative’s performance. This type of eval-uation is most appropriate for intersections that are effec-tively isolated from the effects of upstream signals and thathave typical traffic and geometric features. Several conditionsshould be met before an informal evaluation is considered.These conditions are

1. The intersection is isolated,2. Only one alternative is viable,

3. Geometric and traffic conditions are typical and con- sistent with the assumptions underlying the warrant or guideline used to verify the alternative’s viability ( see disc u s s i o n ooff pprroobblleemmaattiicc conditions in Chap-

ter 2 if signal control is the one alternative), and 4. The engineer’s experience and judgment indicate that

the alternative would resolve the operational problemthat precipitated the study without creating additionalproblems.

If the aforementioned conditions are satisfied, then theengineer may choose to forego the remaining steps in the for-

Figure 3-1. Volume and segment length combinations thatdefine isolated and non-isolated operation.

mal engineering study process. In this case, the engineeringstudy would consist of a field evaluation of the installed alter-native. The field evaluation would verify that the alternativesolved the operational problem that precipitated the study anddid not degrade the overall operation of the intersection. Ifone or more of the four conditions for an informal evaluationare not satisfied, then the formal engineering study should beconducted.

Step 2. Select Analysis Tool

During this step, the analyst will identify the capacityanalysis procedure or simulation model (hereafter referred toas “analysis tool”) that is most appropriate for the engineer-ing study. A procedure for making this identification is de-scribed in this section. The procedure consists of the followingtwo tasks:

a. Identify desired capabilities andb. Evaluate and select analysis tool

During the first task, the desired functions and output capa-bilities of the analysis tool are identified. Then, these functionsand capabilities are compared with those of available analysistools to determine which tool will be the most effective.

2-a. Identify Desired Capabilities

Overview. The analysis tool used for the engineeringstudy must be able to model the traffic control modes and thetraffic flow interactions found at the subject intersection andits alternatives. The tool must also be able to model a systemof signalized and unsignalized intersections when the subjectintersection is non-isolated. Finally, the analysis tool shouldalso be able to predict the desired measures of effectiveness.At the conclusion of this task, the analyst will have developeda list of desired capabilities that can be used in Task 2-b toselect the most appropriate analysis tool.

Any analysis tool used by the agency should be calibratedto replicate existing conditions and be verified for accuracy.However, formal calibration of an analysis tool for any oneengineering study may not be cost-effective. Instead, calibra-tion should be an annual (or bi-annual) undertaking based ondata obtained from a comprehensive study of traffic conditionsin the agency’s jurisdiction.

Traffic Control Modes. The traffic control modes that areconsidered by an analysis tool should include that of the exist-ing intersection and any proposed alternatives. Table 3-1 liststhe control modes that are often required in engineering stud-ies. The analyst should review this list and indicate by check-mark (✓ ) those modes that will be needed for the study ofinterest.

33

Analysis Factors. The analysis tool selected for the engi-neering study should be able to model the various analysisfactors relevant to the subject intersection. These factorsinclude (1) the ability to accept important traffic characteris-tics as inputs and reflect their influence on traffic operations,(2) the ability to model key traffic interactions, and (3) theability to accept important descriptors of intersection geom-etry and reflect their influence on traffic operations. The needto have any one of these factors in a particular model is de-pendent on conditions present at the subject intersection aswell as those conditions present in the alternatives under con-sideration. Table 3-1 lists analysis factors that are often re-quired for engineering studies. The analyst should reviewthis list and indicate by checkmark (✓ ) those analysis factorsneeded for the study of interest.

Measures of Effectiveness and Performance Indicators.The analysis tool selected for the engineering study should beable to predict the various measures of effectiveness and per-formance indicators relevant to the subject intersection. Theneed to have any one of these attributes in a particular modelis dependent on the types of operational problems being expe-rienced and the performance assessment procedures of theagency conducting the study. Table 3-1 lists the measures andindicators that are often required for engineering studies. Theanalyst should review this list and indicate by checkmark (✓ )those measures and indicators needed for the study of interest.

An important output measure of effectiveness is average-delay-per-vehicle. This measure is widely recognized by en-gineers and the motoring public as reflective of intersectionperformance. It is important that the analysis tool provide thisstatistic. For intersections with heavy transit or pedestrianflow, average-delay-per-person is another important statistic.

2-b. Evaluate and Select Analysis Tool

Overview. This section describes a framework for analy-sis tool selection. Initially, a case is made that only one analy-sis tool should be used for the evaluation. Following thisdiscussion, a step-by-step tool selection process is outlined.It is recognized that some agencies may have guidelines re-garding the analysis tool to be used in the engineering study.However, the analyst should still complete this step and con-firm that the tool used in the study is appropriate for theanalysis of the subject intersection and its alternatives.

One Analysis Tool. This section describes two require-ments that the analysis tool should satisfy to be used in theengineering study. First, the analysis tool should be able toexplicitly model the operational features of the intersectionand, if needed, the associated street system. This requirementrecognizes the need for accuracy in the evaluation. It is in-tended to discourage the creative extension of tools to situa-tions that they are not specifically developed to model. Thus,

if a coordinated semi-actuated control alternative is beingconsidered, the analysis tool should explicitly model platoonprogression, coordination, and actuated control.

Second, the delay statistics used to compare various alter-natives should be consistently defined and computed. Thisrequirement promotes the accurate evaluation of the relativeimpact of each alternative on intersection operations. Thus, if

34

traffic signal control is being considered as an alternative, thedelay computed for both the existing intersection and the sig-nalized intersection should be based on a common definition(e.g., control delay) and similarly derived equations.

Based on this discussion, it follows that one analysis toolthat can model all of the alternatives should be used for theengineering study. This approach would ensure equitable com-

TABLE 3-1 Analysis tool capabilities

parisons among alternatives by avoiding biases that mightbe present when two or more tools are used. Such bias couldresult from differences in modeling approach or delay def-inition. The use of one tool may also offer some analysisefficiency due to common input and output data formats.

Selection Process. The analysis tool selection process isbased on the principle that the analysis tool selected shouldprovide an equitable balance between analysis complexity,accuracy, and cost. It is also based on the assumption that oneanalysis tool will be used for all evaluations. The processconsists of the following four activities:

1. Identify applicable traffic control modes,2. Identify applicable traffic and geometric factors,3. Identify applicable output measures and performance

indicators, and4. Select analysis tool.

These activities are described in the following paragraphs.Tables 3-2, 3-3, and 3-4 are consulted during this task. Col-

lectively, these tables identify the capabilities of the variousanalysis tools used in the engineering study. One table col-umn should be allocated for each candidate analysis tool. Thetool-specific information entered into these columns will needto be provided by the analyst. The first column in each of thethree tables has been completed (for a hypothetical analysistool) to illustrate the manner in which relevant informationcan be recorded.

Some effort will be required initially to complete thesetables. However, once they are completed, the developmenteffort will be offset by a significant time savings during sub-sequent engineering studies. One benefit of this formal tabu-lation of tool capabilities is that it ensures that all relevantmodes, factors, and outputs have been considered (and accom-modated to the extent possible) before any effort is expended

35

in the evaluation process. These tables should be periodicallyupdated to maintain their effectiveness.

Activity 1: Identify Applicable Traffic Control Modes. Thecontrol modes identified in Task 2-a are cross-checked withthe tool capabilities identified in Table 3-2. Those analysistools that provide the desired capability should be identifiedand advanced to Activity 2.

To illustrate the use of Table 3-2, consider a situationwhere the existing, isolated intersection has two-way stop-control and where actuated signal control is considered to bea viable alternative. Based on the example information pro-vided in this table, it appears that ABC software is able tomodel both control conditions and can be advanced to Activ-ity 2. In contrast, if the intersection were non-isolated or iftwo-way yield control were an alternative, then this softwarewould not be appropriate because it cannot model either ofthese control modes.

Activity 2: Identify Applicable Traffic and Geometric Fac-tors. The second activity in the selection process is to cross-check the analysis factors identified in Task 2-a with the toolcapabilities identified in Table 3-3. Those analysis tools thatprovide the desired capability should be identified and ad-vanced to Activity 3. (Only those tools that have the desiredcontrol mode and analysis factor capabilities are advanced tothe next activity.)

To illustrate the use of Table 3-3, consider a situationwhere the existing, isolated intersection has two-way stop-control and where the addition of an exclusive left-turn laneon the minor road is a viable alternative. Based on the exam-ple information provided in this table, it appears that ABCsoftware can model the effect of adding a left-turn lane andmay be appropriate for the analysis and should be advancedto Activity 3. However, this software is not sensitive to theoperational problems that would result from a left-turn bay

TABLE 3-2 Traffic control mode check sheet

overflow. Thus, if the left-turn bay were restricted to a rela-tively short length, this software would not be appropriatebecause it cannot model turn bay overflow.

Activity 3: Identify Applicable Output Measures and Per-formance Indicators. The third activity in the selection pro-cess is to cross-check the output measures and indicatorsidentified in Task 2-a with the tool capabilities identified inTable 3-4. Those analysis tools that provide the desired capa-bility should be identified and advanced to Activity 4. (Onlythose tools that have the desired control mode, analysis fac-tor, and output capabilities are advanced to the next activity.)

Activity 4: Select Analysis Tool. The fourth activity in theselection process is to review the list of candidate analysistools and select the one tool that has the lowest resourcecosts. These costs include data entry, software maintenance,training, and model calibration. A high degree of precision inthis estimate is not essential. To aid in this assessment, theanalyst may wish to seek the opinion of other engineers whohave experience using one or more of the analysis tools.

36

Step 3. Conduct Evaluation

The third step in the engineering study stage focuses onquantifying the operational performance of the existing inter-section and any proposed improvement alternatives. The tasksinvolved in this evaluation process are discussed in this sec-tion; they are

a. Gather information,b. Evaluate operational performance, andc. Determine alternative effectiveness.

In the first task, relevant traffic, signalization, and geometricdata are identified and collected. Then, these data are usedwith the analysis tool identified in Step 2 to evaluate the sub-ject intersection and any proposed alternatives.

3-a. Gather Information

Overview. There are two objectives of this task. The firstobjective is to define the evaluation time periods. The second

TABLE 3-3 Analysis factor check sheet

objective is to identify the data needed for Task 3-b, Evalu-ate Operational Performance. At the conclusion of this task,all of the data needed for the evaluation should be identifiedand collected.

Define Evaluation Periods. At the start of this task, theperiods for which traffic operations will be evaluated shouldbe defined. It is generally sufficient to evaluate operations forthree representative 1-hr periods. These hours should includethe morning and afternoon peak traffic demand hours as wellas one representative off-peak demand hour. The off-peakhour should represent an average of traffic conditions duringthe hours from 6:00 a.m. to 10:00 p.m., excluding the twopeak hours. Additional hours can be evaluated if traffic pat-terns are highly varied or if a more comprehensive assessmentof operations is required.

Identify Data. Much of the data needed for Task 3-b willhave been collected or derived during the alternative identi-fication and screening stage (see Chapter 2). However, ad-ditional data may be needed to calibrate the analysis toolselected in Task 2-b. Also, the data previously collected maybe incomplete in terms of the evaluation periods just defined.Data that may need to be collected include

• Turning movement counts,• Pedestrian and heavy vehicle counts,

37

• Approach grade,• Approach speed, and• Vehicle occupancy.

If the subject intersection is non-isolated, then additionaldata will be needed to characterize the operation of the signalsystem. Such data may include

• Turning movement counts at adjacent signalized inter-sections,

• Phase sequence and duration at adjacent signalizedintersections,

• Signal offset relationship between signalized inter-sections, and

• Geometry of adjacent signalized intersections.

Collect Data. The procedures for collecting the necessarydata will vary depending on whether the subject intersectionexists or is proposed for construction. For existing intersec-tions, appropriate data collection procedures are described inthe Manual of Transportation Engineering Studies (6). Forproposed intersections, techniques described in Appendix Ccan be used to estimate turn movement volumes from forecastaverage daily traffic demands. Regardless of the source, thedata should represent traffic conditions occurring on an “aver-age day” (i.e., a day representing traffic volumes normallyand repeatedly found at a location).

TABLE 3-4 Measures of effectiveness and performance indicators check sheet

3-b. Evaluate Operational Performance

Overview. The subject intersection and all viable alter-natives should be formally evaluated during this task. Thisevaluation should determine the average delay to each inter-section traffic movement, as well as the overall averageintersection delay. These delay measures will be used inTask 3-c to determine whether an alternative can effectivelyimprove the operation of the subject intersection.

Design Alternatives. Before the evaluation, the physicalelements of each alternative should be sized and the functionalelements identified (at a preliminary design level of detail) tofacilitate an accurate evaluation of the alternative’s operationaleffectiveness. The time expended in this design effort willincrease with the alternative’s complexity; however, the con-sequences of not devoting the time needed for this design canbe far more costly. For example, an overly simplified signal-ized intersection design may yield inaccurate performanceestimates and could result in the analyst recommending whatwill ultimately turn out to be an “unnecessary” signal.

With one exception, preliminary design guidelines for inter-section improvement alternatives are not provided in thisdocument. The analyst is directed to other documents, such asAASHTO’s A Policy on Geometric Design of Highways andStreets (Green Book) (5) or the agency’s roadway design man-ual, for this guidance. The one exception is the traffic signalalternative. Guidelines for the design of this alternative areprovided in the Guidelines section of this chapter.

Conduct Evaluation. At this point, the analyst shoulduse the analysis tool identified in Step 2 to evaluate the oper-ation of the subject intersection and the various alternativesidentified in the alternative identification and screening stage.

Analysis tools that are characterized as “stochastic” willrequire additional effort on the part of the analyst to ensurethat the output delay statistics are accurately interpreted. Thisrequirement results from the stochastic analysis tool’s explicitmodeling of traffic events and driver decisions as randomprocesses. This modeling approach has the advantage of pro-ducing a great deal of realism in the simulation of the trafficsystem; however, it has the disadvantage of requiring extraeffort in setting up and running the simulation model. Guide-lines for using stochastic analysis tools in the engineeringstudy are provided in the Guidelines section of this chapter.

3-c. Determine Alternative Effectiveness

Overview. The objective of this task is to determine if anyof the alternatives (including the “do nothing” alternative)evaluated in Task 3-b can serve traffic effectively. Criteria aredescribed in this task for making this determination. Initially,two descriptors of traffic service quality are defined. Then,these descriptors are used to describe the conditions that an

38

alternative must satisfy to be termed “effective.” At the con-clusion of this task, all effective alternatives are identified andadvanced to the alternative selection stage (see Chapter 4).

Definitions. Two descriptors are used to define thresholdconditions that, if exceeded, are associated with unacceptableoperations. These two descriptors are “acceptable level ofservice” and “acceptable operation.” Their definitions are

• Acceptable level of service—an overall intersectionoperation that can be described as level-of-service(LOS) D or better [as defined in the Highway CapacityManual 2000 (15)].

• Acceptable operation—an intersection traffic move-ment that can be described as operating at (1) LOS D orbetter or (2) a total delay less than 4.0 vehicle-hours (5.0 vehicle-hours for a multilane movement).

The “total delay” threshold (i.e., 4.0 or 5.0 vehicle-hours)is based on recommendations made by Henry and Calhoun(26) and by Sampson (30) as a result of their signal-warrant-related research. The “level-of-service” threshold (i.e., “D”)is based on the recommendation made in Table II-6 of theGreen Book (5). The analyst may substitute agency-preferredthresholds.

The effect of considering both total delay and level-of-service (defined in terms of average control delay) is illus-trated in Figure 3-2. The convex trend line shown in this fig-ure represents volume and delay combinations that equal 4.0(or 5.0) vehicle-hours of total delay. The horizontal trendline represents 35 seconds per vehicle (s/veh) of averagedelay. This level of delay represents an upper limit for LOSD at unsignalized intersections [based on definitions in theHCM 2000 (15)]. The thick line represents the combinationof total delay and average control delay that form an upperlimit on “acceptable operation.” Because the curved trendline is above that associated with LOS D for volumes less

Figure 3-2. Acceptable operating conditions atunsignalized intersections.

than about 420 vehicle-hours, a low-volume movement mayhave delays that yield LOS E or F but may still be consid-ered as having “acceptable operation” by this approach.

Conditions for Identifying Effective Alternatives. Con-ditions are described in this section that can be used to identifyalternatives that are operationally effective and, therefore,suitable for consideration in the alternative selection stage.A separate series is provided for the isolated and non-isolatedintersection types. Each condition must be satisfied beforean alternative can be considered effective. The conditions foreach intersection type follow.

For Isolated Intersections. An effective alternative should

1. Improve overall intersection operations by reducingaverage intersection delay (based on an assessment ofconditions existing throughout the typical day); and

2. Result in an “acceptable level of service” for the inter-section (based on an assessment of the peak trafficdemand hour); and

3. Result in the “acceptable operation” of all minor move-ments (based on an assessment of the peak trafficdemand hour).

For Non-Isolated Intersections. An effective alternativeshould

A. Improve overall system operations by reducing the aver-age system delay (based on an assessment of conditionsexisting throughout the typical day); and

B. Result in an “acceptable level of service” for eachsignalized intersection in the system (based on anassessment of the peak traffic demand hour); and

C. Result in the “acceptable operation” of all minormovements at each signalized intersection (based onan assessment of the peak traffic demand hour); and

D. Satisfy Conditions 1, 2, and 3 (as described immedi-ately above) for the subject intersection.

Four terms or phrases used in the preceding conditions aredefined as follows:

• The “average intersection delay” identified in Condi-tion 1 should be computed as a volume-weighted aver-age that reflects all approach traffic movements at theintersection.

• The “average system delay” identified in Condition Ashould be computed as a volume-weighted average thatreflects all approach movements at each intersection inthe system.

• The phrase “conditions existing throughout the day”used in Conditions 1 and A means conditions occurringduring the morning and afternoon peak hours as well asone off-peak hour; additional hours can be included inthe assessment to provide better representation.

39

• As a minimum, the “system” identified in ConditionsA and B should include the nearest upstream and down-stream signalized intersections. Additional signalizedintersections along the arterial street may also be included,if the analyst determines that they would be affected byimprovements to the subject intersection.

In application, these conditions would be assessed for eachalternative. If all conditions are satisfied for a given alternative,then the alternative is “effective” in terms of being able toimprove operations at the subject intersection. All “effective”alternatives should be advanced to the alternative selectionstage, as described in Chapter 4.

GUIDELINES

This section provides guidance related to select componentsof the engineering study stage. The guidance is presented inthe following two sections:

1. Guidelines for Designing the Signalized IntersectionAlternative and

2. Guidelines for Use of Stochastic Simulation Models.

The first section describes guidelines that can be used tomake preliminary design decisions for the traffic signal alter-native. The second section describes guidelines that can beused when the analysis tool simulates traffic flow as a randomprocess such that the predicted performance measures (e.g.,delay) vary with each simulation run.

Guidelines for Designing the SignalizedIntersection Alternative

Overview

This section describes guidelines for the preliminary designof a signalized intersection. The guidelines were obtainedprimarily from reference documents that are generally rec-ognized as authoritative guides on engineering design. Othersources for the guidelines include reports documenting signif-icant research efforts whose recommendations are intendedfor nationwide application. Where guidelines are not avail-able, they have been derived using traditional engineeringanalysis techniques. The guidelines address the followingdesign topics:

• Intersection geometry,• Controller operation,• Controller phase sequence,• Basic controller settings,• Controller settings for isolated operation, and• Controller settings for coordinated operation.

These guidelines should be used to design the signalizedintersection alternative if, during the alternative identifica-tion and screening stage, it is determined that the trafficsignal alternative is viable. These guidelines are intended toassist the engineer to develop a reasonable representation ofthe traffic signal alternative. In this regard, the traffic sig-nal alternative should be properly designed before beingevaluated because the level of service it provides is oftenvery sensitive to the selected phase sequence, signal timing,and geometry. If not properly designed, the traffic signalalternative’s performance could be grossly overestimatedor underestimated.

The guidelines provided in this section do not address allaspects of the preliminary design of a signalized intersec-tion. Although they are believed to be appropriate for themore typical signalized intersections, some judgment maybe needed in the design of intersections that have uniqueoperating or geometric conditions. These guidelines areintended to describe reasonable design choices solely forthe purpose of fairly evaluating the traffic signal alter-native; they are not intended to be used as a substitutefor applicable state or local design standards.

Several software products are available to automate ele-ments of the signal timing design process. Some products canalso be used to make geometric design decisions in a fairlyefficient manner. The analyst may substitute these products forthe guidelines offered in this section. The only stipulation isthat each of the “bulleted” elements in the preceding list shouldbe designed using sound, defensible engineering practices.

Intersection Geometry

This section describes guidelines for estimating the numberof lanes needed for each intersection traffic movement. Theseguidelines were derived from the information provided inAppendix A of Chapter 16 in the Highway Capacity Manual2000 (15). These guidelines represent a good starting point forthe intersection design. Once established, this idealized geom-etry can be modified to accommodate the lane configuration ofthe existing road system and the availability of right-of-way.

The guidelines are intended for separate application to eachintersection approach and movement and should be used toestimate the number of lanes needed during both the morningand afternoon peak demand hours. Of these two estimates, thelarger should be used to design the approach. If only one peakhour is evaluated, then the number of lanes determined for thethrough movement in the peak demand direction should alsobe used for the opposing through direction to ensure adequateservice during both peak hours.

A minimum of two lanes should be considered for theminor-road approaches to an arterial street or highway. Thesetwo lanes may serve any combination of through or turnmovements. This practice will minimize the effect of a newsignal on the major-road operation (except when pedestriancrossing times dictate a lengthy minor-road phase duration).

40

Provision of only one minor-road lane tends to result in aninequitable distribution of cycle time between major andminor movements. Such inequity is especially disruptive whenthe major road is coordinated.

Exclusive Left-Turn Lanes. The Highway CapacityManual 2000 (15) suggests that the following criteria can beused to estimate the number of lanes needed for the left-turnmovement:

• Provide one or more exclusive lanes, if a left-turn phaseis provided;

• Provide one exclusive lane, if 100 veh/h <Vlt <300 veh/h;and

• Provide two exclusive lanes, if Vlt > 300 veh/h.

Where the variable Vlt represents the left-turn movement vol-ume (in veh/h).

These volume thresholds are approximate because of thecomplex nature of shared-lane operations and are most ap-propriate when the opposing and adjacent through move-ments have flow rates of 450 veh/h or more. Exclusive turnlanes may not be needed for lower flow rates. The need foran exclusive left-turn phase can be determined using theguidance provided in the section, Left-Turn Phasing.

Through Lanes (or Shared Lanes). Highway CapacityManual 2000 (15) suggests that enough through lanes shouldbe provided on an approach to ensure that the through volume(as well as any right- or left-turn volume sharing the throughlanes) does not exceed 450 veh/h/ln.

Exclusive Right-Turn Lanes. Finally, the Highway Ca-pacity Manual 2000 (15) suggests that an exclusive right-turn lane should be considered when the right-turn volumeexceeds 300 veh/h and the adjacent through lane volumeexceeds 300 veh/h/ln.

The guidance provided in the previous two paragraphs isgeneralized in Figure 3-3. This figure indicates the number of

Figure 3-3. Relationship between approach volume andapproach lane assignments.

through and right-turn lanes needed for various volume com-binations. The Highway Capacity Manual 2000 guidance hasbeen extended for the development of this figure (e.g., it isassumed that right-turn volumes in excess of 600 veh/h mayrequire two exclusive turn lanes).

The x-axis of Figure 3-3 represents the through volumeon the subject approach as well as any left-turn volume thatshares the through lanes (i.e., when no left-turn lane is pro-vided). When consulting this figure, the analyst should findthe point where the right-turn and the through (plus sharedleft-turn) volumes intersect. The location of this point deter-mines the number of through lanes needed and, if appropriate,the number of exclusive right-turn lanes.

Example. To illustrate the use of Figure 3-3, consider anapproach with the following flow rates during the peakdemand hour:

• Left-turn volume - 50 veh/h,• Right-turn volume - 350 veh/h, and• Through volume - 800 veh/h.

The Highway Capacity Manual 2000 (15) guidelines notedpreviously indicate that a left-turn volume of 50 veh/h is in-sufficient to justify a left-turn lane. Figure 3-3 is then checkedusing a “through + shared left-turn volume” of 850 veh/h (= 800 + 50) and a right-turn volume of 350 veh/h (see dashedlines). The figure indicates that two through lanes are neededalong with one exclusive right-turn lane.

Reconciliation with Existing Roadways. After the inter-section geometry is defined, it must be reconciled with thatof the existing intersecting roadways. In general, the numberof through lanes at an intersection should be consistent withthe number of continuous through lanes on the roadway. Anadditional through lane at an intersection is generally not effi-

41

ciently used unless it extends 900 m (3,000 ft) or more down-stream; provision of such a length is often not practical. Simi-larly, it may not be practical to provide a dual left-turn lanewhen the receiving roadway has only one through lane.

Tradeoffs may be possible as the guidelines in this sectionare conservative. When the guidelines indicate that the inter-section needs 10 to 20 lanes in total, one or two lanes canprobably be excluded (if taken from a multilane movement)without causing major operational problems. Also, subtrac-tion of one lane from a given traffic movement may be off-set by adding a lane to a conflicting movement. However,significant changes in lane assignment or reductions in thetotal number of lanes may not be possible without causing asignificant increase in delay. Ultimately, it is the responsibil-ity of the engineer to define an intersection geometry thatprovides a reasonable balance between intersection opera-tions and the geometry that can reasonably be accommodatedby the existing roadways.

Controller Operation

This section provides guidelines for selecting the mostappropriate control mode. These modes include pretimed,full-actuated, and semi-actuated control. The most appropri-ate choice of control mode for a given intersection is primar-ily based on the location of the subject intersection relativeto nearby intersections and the type of signal operation usedas a result of this location.

Determining the type of control that is best suited to a par-ticular intersection is a fundamental and critically importantdesign decision. Table 3-5 describes the range of applicationof the principal types of control. The guidance provided inthis table was adapted from Gordon et al. (31, p. 7–9). Theguidance regarding “consistent” versus “fluctuating” trafficdemands that is provided in the table footnotes was obtainedfrom Kell and Fullerton (32, p. 31).

TABLE 3-5 Relationship between intersection operation and control mode

When local practice does not indicate the preferred controlmode, the information in Table 3-5 can be used to identify asuitable mode. Application of this table requires that theidentification of “intersection operation” and, in the case ofarterial street systems, the character of the crossroad trafficdemand. For example, the table indicates that an isolatedintersection probably should have full-actuated control. Onthe other hand, an intersection that is located along an arterialstreet and that has “light” crossroad traffic demands probablyshould have semi-actuated control. Finally, an intersectionlocated in a downtown street grid system probably should havepretimed control.

Controller Phase Sequence

This section describes several phase sequences often usedat signalized intersections. By definition, a minimum of twophases is needed at an intersection of two roads. The provi-sion of a third or a fourth phase sequence typically dependson whether the left-turn and opposing through movementsneed to be separated in time. If this separation is needed, it istypically accomplished by providing a left-turn phase. Guide-lines are provided in this section that describe conditionswhere a left-turn phase may be needed.

Phase Sequence Variations. At a preliminary designlevel, the analyst should determine the manner in which thethrough and the left-turn movements will be served by the con-troller phase sequence. At the simplest level, two phases canbe used to control the intersection. This two-phase sequence,shown as Sequence A in Figure 3-4, alternates the right-of-way between the two intersecting roads—one phase servingeach road. If a road has two-way traffic, then its phase servesopposing through movements in an exclusive or “protected”

42

manner. In contrast, the left-turning drivers are not protectedand must yield to the oncoming through vehicles. This type ofyielding left-turn service is referred to as “permitted” left-turnoperation.

If gaps in the oncoming stream are inadequate to servethe left-turn demand or if left-turn-related accidents areunusually frequent, then a protected left-turn phase is oftenadded to the phase sequence. This left-turn phase can occurbefore (i.e., lead), after (i.e., lag), or at the same time as theadjacent through-movement phase. Figure 3-4 illustratesseveral three- and four-phase sequences that include boththrough and left-turn phases.

Left-Turn Phasing. As a general rule, the number ofphases should be kept to a minimum because each additionalphase in the signal cycle reduces the time available to theother phases. Thus, two-phase operation with permitted left-turn movements should be considered as a “starting-point.”Left-turn phasing should only be provided if it will improveoperations or safety. The guidelines shown in Figure 3-5 canbe used to make this determination. They can also be used to determine whether the left-turn phase should operate asprotected-plus-permitted or protected-only. These guidelineswere derived from guidance provided by Kell and Fullerton(32) and by Orcutt (28).

Full application of Figure 3-5 requires knowledge of thecrash history, intersection geometry, sight-distance, left-turndelay, and volume conditions. The following assumptionscan be made if such information is unavailable:

1. If collision frequency is unavailable, it should be assumedto be less than the critical number.

2. If the road is built to satisfy AASHTO criteria, asdescribed in the Green Book (5), then it should beassumed that sight distance is available.

3. If the cycle length is unknown, it should be assumedto be 60 s in duration (this assumption can be checkedusing guidance provided in a subsequent section).

The flowchart in Figure 3-5 has two alternative paths fol-lowing the check of left-turn volume. One path requiresknowledge of left-turn delay; the other requires knowledgeof the left-turn and opposing through volumes. The left-turndelay referred to is that delay incurred when no left-turnphase is provided. If no information is available on the left-turn delay, the alternative path based on volume should beevaluated.

The application of Figure 3-5 requires the separate eval-uation of each left-turn movement. This evaluation shouldconsider both the morning and afternoon peak demand hoursof the average day. Left-turn phasing should only be pro-vided for the left-turn movement that satisfies the guidelinesin Figure 3-5. For the purposes of the engineering study, ifleft-turn phasing is needed for one peak hour, then it shouldbe provided for the entire day.Figure 3-4. Selected signal phase sequences.

Common Sequences. This section describes commonphase sequences used when some type of left-turn phaseprotection is provided. One of the more commonly usedsequences serves both left-turn movements during the firstphase (i.e., a “leading” left-turn phase) and both throughmovements during the second phase. This sequence tendsto pair movements having similar volume together. The advan-tage of such a pairing is that it provides some efficiency by con-currently serving movements that have similar volumes andservice time needs. Phase Sequence B in Figure 3-4 illustratesthis approach for the north-south road.

Modern, dual-ring controllers improve on the aforemen-tioned practice by allowing for the concurrent service ofseveral movements, including the adjacent left-turn andthrough movements. This type of service allows the left-turn and the through movement to be served in an overlap-ping manner. This capability improves intersection effi-ciency when opposing left-turn pairs or opposing throughmovement pairs have dissimilar volumes and is particularlyeffective when implemented with actuated control. Phase

43

Sequences C and D in Figure 3-4 illustrate this approach forthe north-south road.

Both leading and lagging left-turn phases have advantagesand disadvantages. Kell and Fullerton (32) list these advan-tages and disadvantages, many of which relate to driverexpectancy and behavior during the transition between theleft and through phases. Local practice will often dictate theconsistent use of a leading or lagging left-turn phase sequencethroughout the jurisdiction.

There is one noteworthy disadvantage associated with alagging, protected-plus-permitted left-turn phase. This dis-advantage relates to the potential conflict between left-turnvehicles clearing the intersection at the end of the throughphase and the opposing through vehicles. This problem issometimes referred to as the “left-turn trap” and is describedmore fully by Orcutt (28, p. 27 ). The trap can be avoidedby forcing the simultaneous termination of the leadingthrough phases or by eliminating the permitted portion of theprotected-plus-permitted left-turn operation. The problemcan also be avoided by using leading left-turn phasing.

Figure 3-5. Guidelines for determining the potential need for a separate left-turn phase.

If local practice does not dictate the order of presentationof the left-turn phases, it should be assumed that the left-turnphase (if needed) leads the adjacent through phase. The ade-quacy of this assumption can be checked during Step 3 of theengineering study and a lagging left-turn phase can be usedif it is found to yield better operation.

Special Sequences. As noted in the preceding section,typical practice is to serve opposing left-turns during onephase and opposing through movements during a differentphase. However, there is occasionally the need to provide aseparate phase for each approach. This need generally occurswhen the travel paths of opposing traffic movements physicallyoverlap within the conflict area. When a separate phase is pro-vided for each approach, the phase sequence is often referredto as “direction separation,” “split phasing,” or “sequentiallyphased roads.”

Direction separation phasing is less efficient than otherphase sequences when the left-turn and adjacent through vol-umes are not equal. As these volumes are rarely equal, direc-tion separation phasing is rarely efficient and should beavoided. When possible, intersection geometry should be mod-ified so that left-turn movements can be served simultaneouslyduring a common phase.

Basic Controller Settings

The development of an efficient timing plan for the sig-nal phase sequence is an essential task when designing the traffic signal alternative for the engineering study. Thecomponents of this plan vary depending on (1) whether apretimed, semi-actuated, or full-actuated control mode isused and (2) whether an isolated or coordinated operation isused. A pretimed timing plan includes specification of cyclelength and a green interval duration for each phase. A full-actuated timing plan includes specification of a minimumgreen, a maximum green, and a unit-extension setting foreach phase. A semi-actuated plan for a coordinated systemincludes a system cycle length, force-off and yield points,and a reference point offset. It also includes minimum green,maximum green, and unit-extension settings for the actuated(non-coordinated) phases.

Despite the aforementioned differences among controlmodes, there are some basic settings that all modes have incommon. Specifically, they all require yellow and all-redinterval durations for each phase. They also require sometype of minimum green interval. A procedure is described inthis section for determining these interval settings, as theyapply to both the pretimed, semi-actuated, and full-actuatedcontrol modes.

Two sections follow this section (i.e., Controller Settingsfor Isolated Operation and Controller Settings for Coordi-nated Operation). These sections describe procedures fordetermining controller settings specific to the method of sig-

44

nal operation. Within each of these sections, separate sub-sections describe procedures for determining settings for thepretimed and actuated control modes.

The procedures described in this section and the two thatfollow are intended to provide a balance between the effortneeded to develop a reasonable timing plan and the accuracyneeded for the engineering study. Other timing plans are pos-sible and can often be defined using other procedures or soft-ware analysis tools. Regardless of method used, the analyst’sgoal here is to develop a reasonably efficient timing plan thatfairly represents the traffic signal alternative in the engineer-ing study. This procedure is not intended for use in the finaldesign of a timing plan.

Number of Timing Plans. An efficient timing plan shouldbe one that results in minimal delay to all vehicles entering theintersection throughout the day. To achieve this result, the ana-lyst needs to establish a timing plan that accommodates vari-ability in traffic demands throughout the day. The nature of thisaccommodation will vary depending on whether the controlmode is pretimed or actuated.

For pretimed control, the analyst should develop no moretiming plans than would typically be implemented on a time-of-day basis in his (or her) jurisdiction. Typically, this wouldmean determining the green interval durations for one tothree timing plans, using either software-automated or man-ual techniques. Each plan would be based on volumes repre-sentative of the applicable period. For example, Plan 1 couldbe based on the peak hour of flow during the morning andwould be used for the 2-hr period from 7:00 to 9:00 a.m. Plan2 could be based on the afternoon peak and apply to the 2-hrperiod from 4:00 to 6:00 p.m. Finally, Plan 3 could be basedon a representative off-peak period (or weekend day) andapply to all non-peak hours.

Actuated control is better able to adapt to demand vari-ability than is pretimed control. For isolated operation, onewell-constructed timing plan should be able to serve trafficdemands for the entire day. Typically, this would mean deter-mining the maximum green settings suitable for both themorning and afternoon peak demand hours. For a given phase,the larger of the two maximum green settings determined inthis manner would be set in the controller and used for theentire day. For non-isolated operation, the cycle length,force-off, yield-point, and offset settings would need to bedetermined for both peak hours and for a representative off-peak period.

Change Interval Duration. Fundamental components ofthe timing plan are the yellow warning and all-red clearanceintervals. Together, these two intervals represent the changeinterval. Kell and Fullerton (32) indicate that there is con-siderable diversity in practice regarding the duration of thechange interval. Although they discuss several commonlyused strategies, they do not explicitly recommend any onestrategy. They do offer a table of change interval durations

that are based on theoretic relationships of driver response anddeceleration. The table offered by Kell and Fullerton is repro-duced as Table 3-6. Unless local practice dictates otherwise,it is recommended that the analyst use the change intervaldurations listed in this table for the engineering study.

Table 3-6 also provides minimum phase durations basedon pedestrian crossing times. Guidance for using this portionof the table is provided in the next section.

Minimum Green Time. The minimum green durationof a phase is dependent on driver expectancy and, in somesituations, pedestrian crossing time. Driver expectancy shouldbe accommodated in the minimum green for all phases (left-turn and through). In contrast, pedestrian crossing needsshould only be accommodated in through movement phasesthat have some pedestrian demand but no pedestrian call but-ton. In general, all phases should have a minimum duration of7.0 to 10.0 s to satisfy the expectancy of waiting drivers. Min-imum green durations based on pedestrian crossing times gen-erally exceed this amount and vary widely, depending on thewidth of the road being crossed.

45

For the purpose of the engineering study, it is recom-mended that a minimum green duration of 8.0 s be used forall phases, unless local practice dictates other values. The oneexception to this rule applies to through phases that cannotbe activated by a pedestrian call button. These phases shouldbe long enough for pedestrians (traveling in the same direc-tion as the vehicles served) to cross the intersecting road. Theminimum pedestrian phase duration Pp would equal the com-bined pedestrian reaction time and crossing time. The mini-mum green duration Gm would then equal the larger of 8.0 sor the minimum phase duration less the change interval Y(i.e., Gm = larger of: 8.0 or Pp − Y). Typical minimum pedes-trian phase durations are listed in the last row of Table 3-6.These durations are based on reaction times and walkingspeeds reported by Kell and Fullerton (32).

Controller Settings for Isolated Operation

This section describes a procedure for determining thecontroller settings for isolated intersections. The procedure

TABLE 3-6 Change interval duration and pedestrian-based phase duration

has been developed for application to intersections with bothpretimed and actuated control modes. Prior to discussing theprocedure, there is a brief discussion of cycle length, as itrelates to pretimed and actuated intersection operation. Then,a seven-step procedure is described for determining the set-tings for pretimed control. Finally, a procedure is describedfor determining the settings for actuated control. This latterprocedure is developed as a three-step extension of the pre-timed procedure. As such, it requires completion of the firstsix steps of the pretimed procedure.

Cycle Length. The efficient operation of a pretimed, sig-nalized intersection is highly dependent on cycle length.Webster (33) demonstrated that overly long and short cyclelengths increase delay. He derived a relationship between theminimum delay cycle length and critical lane volume. Thisrelationship is shown in Figure 3-6. The trends in this figureare based on an effective saturation flow rate of 1,500 veh/h/ln and 4.0 s lost time per phase. This saturation flow ratereflects the demand peaks, lane use, and traffic mix found ata typical intersection.

As the trends in Figure 3-6 indicate, low to moderate vol-umes can yield minimum-delay cycle lengths less than 40 s.In contrast, high volumes can yield cycle lengths in excess of180 s. When the cycle length is exceptionally short or long,the analyst may wish to consider modifying the number oflanes on one or more approaches (i.e., reducing the numberof lanes if the cycle length is below 40 s and increasing theirnumber if the cycle length exceeds 180 s).

For actuated operation, a well-designed detection/controlsystem will yield minimal delay operation by ensuring opti-mally utilized phase durations and cycle lengths. Therefore,for the engineering study of an actuated traffic signal alter-native, it can be assumed that the cycle length obtained fromFigure 3-6 is reasonably close to the average cycle lengthachieved by actuated operation for the analysis period. Asdescribed in a subsequent section, Actuated Phase Settingsand Detection Design, this average cycle length can be usedto determine reasonable maximum green settings when theproposed intersection has actuated control.

46

Pretimed Phase Interval Durations. Appendix B ofChapter 16 in the Highway Capacity Manual 2000 (15)describes a procedure for determining the green interval dura-tion of a pretimed phase. A simplified version of this proce-dure is described in Tables 3-7 and 3-8. These tables can beused to determine reasonably efficient green interval durationsthat should be adequate for the engineering study process.Many software analysis tools are also available that can auto-mate the process of determining pretimed phase duration.Such tools can be used instead of the manual process describedin this section.

The procedure described in this section is based on the fol-lowing assumptions: (1) the left-turn phases lead the throughphases, (2) one or more left-turn lanes are provided when aleft-turn phase is used, and (3) both left-turn movements on agiven road are protected when a left-turn phase is used. Otherprocedures should be considered when these assumptions arenot appropriate.

The procedure for determining the green interval durationis described in terms of an example application. Blank ver-sions of the worksheets are provided in Appendix C. The ex-ample intersection geometry and demand volumes (in veh/h)are illustrated in Figure 3-7. The east-west road is proposedto have a protected-plus-permitted left-turn phase. The north-south road will not have protected left-turn phasing. Thethrough phases will need to be long enough to serve pedes-trians. The speed on the north-south road approaches is 40 km/h (25 mph); that on the east-west road approaches is60 km/h (40 mph).

Step 1. Input Volume and Lane Geometry. The procedurefor defining a pretimed timing plan starts with Table 3-7. Thisworksheet is used to determine the total critical lane volume.This volume is then used in Table 3-8 to determine an appro-priate cycle length and phase durations. The worksheet is com-pleted from top to bottom. Initially, the movement volumesand lane counts are entered in the Volume and Lane GeometryInput section of the worksheet. For this analysis, the right-turnvolume is combined with the through movement volume. Sim-ilarly, any exclusive right-turn lanes would be included in thecount of through lanes for a given approach.

Step 2. Compute Adjusted Movement Volumes. The PhaseSequence Section of the table is completed next. This sec-tion is divided into three parts, depending on the type ofleft-turn phasing provided. Each road is considered sepa-rately. The first part of this section is completed when aroad does not have a protected left-turn phase. The secondpart is completed when a road has a protected-plus-permittedleft-turn phase. The third part is completed if protected-onlyleft-turn phasing is used. The first and second parts requirethe computation of adjusted volumes to account for theeffect of permitted left-turn activity. This computation isomitted in the third part, because protected-only left-turnphasing is used.Figure 3-6. Minimum delay cycle length.

For this example, the first part of the Phase Sequence Sec-tion is completed for the north-south road because it does nothave left-turn phasing. The opposing volumes for both thenorth and southbound left-turn movements are entered in thefirst row in the Phase Sequence Section. This information iscombined with that in the next three rows to convert the left-turn volume into an equivalent through volume. The left-turnequivalence factor EL is obtained from Figure 3-8. Values for

47

this factor are based on information provided in Appendix Cof Chapter 16 in the Highway Capacity Manual 2000 (15).

The northbound left-turn movement shares a lane with thenorthbound through movement and is opposed by 104 south-bound vehicles per hour. Figure 3-8 indicates that these con-ditions result in a left-turn equivalency factor of 1.5. Thisfactor is used, along with the number of vehicles that clear atthe end of the through phase (i.e., sneakers), to determine the

TABLE 3-7 Critical volume worksheet

“adjusted” left-turn volume. It is suggested that 90 sneakersper hour be used for this analysis. For the northbound left-turn movement, the adjusted volume is computed as 5 veh/h(= 1.5[93 − 90]). For through movements, the adjusted vol-ume is equal to the actual through movement volume (noadjustment is needed).

48

The adjusted volume for the east-west road is based onthe calculations described in the second part of the PhaseSequence Section. The adjusted left-turn volume is basedon a conservative estimate of the permitted left-turn capacity.It is suggested that 60 veh/h be used for this analysis. Forthe eastbound left-turn movement, the adjusted volume is

TABLE 3-8 Controller setting worksheet with calculations for pretimed control

computed as 45 veh/h (= 105 − 60). For through move-ments, the adjusted volume is equal to the actual throughmovement volume.

Step 3. Compute Lane Volumes. For this step, the ad-justed volume is used to compute the lane volume for eachmovement. For movements with exclusive lanes, the lanevolume is computed as the adjusted volume divided by thenumber of lanes. However, as indicated in Footnote 2 toTable 3-7, the lane volume for a shared-lane approach iscomputed as the total, adjusted approach volume divided bythe number of through lanes.

The north and southbound approaches both have sharedlanes; therefore, the provision in Footnote 2 applies to eachapproach. The lane volume for the northbound approach iscomputed as 206 veh/h/ln (= [5 + 408]/2).

The east and westbound approaches both have exclusivelanes; therefore, the lane volumes are computed as the adjusted

49

volume divided by the number of lanes. The lane volume forthe eastbound through movement is 251 veh/h/ln (= 502/2).

Step 4. Compute Critical Volumes. As a final step on theCritical Volume Worksheet, the critical lane volumes areidentified. A critical volume represents the largest lane vol-ume served by a given phase. The rules for making thisidentification are listed in the rows labeled Critical Volumes.

Based on the rules listed in the worksheet, the critical vol-ume for the north-south road represents the largest of the lanevolumes on the north and southbound approaches. Thus, thecritical lane volume is 206 veh/h/ln for the north-south road.

Two volumes are required for the east-west road becausethere are two phases being used to serve east-west traffic. Onevolume represents the larger lane volume of the two left-turnmovements; the second volume represents the larger volumeof the two through movements. Thus, the critical volume pairfor the east-west road is 141 and 403 veh/h/ln.

Step 5. Identify Change Interval and Minimum GreenDuration. To complete the development of the timing plan,the Controller Setting Worksheet (i.e., Table 3-8) must becompleted. The columns that correspond to through trafficmovements should always be completed. Those columns thatcorrespond to left-turn movements should only be completedwhen the movement is served by a left-turn phase.

As a first step, Table 3-6 is consulted to determine thechange interval duration and the pedestrian phase time for thethrough movements. The left-turn phase change interval canbe assumed to equal that associated with the adjacent throughmovement. It can also be assumed that there is no pedestrianservice during a left-turn phase (i.e., Pp =0.0). Finally, the min-imum green duration is computed as the larger of the timeneeded to satisfy driver expectancy (i.e., 8.0 s) and the timeneeded to serve pedestrians (if appropriate).

For the north-south through movement, Table 3-6 indicatesthat a change interval of 5 s is appropriate for an approachspeed of 40 km/h (25 mph) and a crossing distance of 18 m (60 ft) [based on five lanes at 3.6 m (12 ft) each]. Table 3-6also indicates that a pedestrian phase time of 21 s is needed forthe 18-m (59-ft) crossing distance. This phase time translatesinto a minimum green time of 16 s.

Step 6. Compute Critical Volume Sum and Cycle Length.The critical volume sum and cycle length are determined by completing the Critical Volume Summary Section ofTable 3-8. This section is completed by transferring the criti-cal volumes from the Critical Volume Worksheet, computingtheir sum, and then using this sum to determine the minimum-delay cycle length Co. The minimum-delay cycle length isobtained from Figure 3-6 for a given critical volume sum andnumber of phases.

For the example intersection, the particular cells com-pleted in the Critical Volume Worksheet are a reminder thattwo phases are used to serve east-west traffic and one phase

Figure 3-7. Example intersection used to illustrate timingplan development.

Figure 3-8. Through-vehicle equivalents for permittedleft-turn vehicles.

to serve north-south traffic. This phase combination is repre-sented by the row labeled “2 phases E-W, 1 phase N-S.” Thecorresponding critical volumes are transferred to this rowfrom the Critical Volume Worksheet. The sum of these threecritical volumes is found to be 750 veh/h/ln. Consultation ofFigure 3-6 indicates that the minimum-delay cycle length forthis volume (and three-phase operation) is 46 s.

Step 7. Compute Green Interval Duration. The green inter-val duration for each phase is determined by completing thePretimed Phases Section of the worksheet. In the first row ofthis section, the critical phase volume from the preceding sec-tion is transcribed to all cells corresponding to the movementsserved by the phase. Finally, these volumes are used to com-pute the green interval duration for each phase. The equationused for this purpose is:

where

Gi = green interval duration for movement i (i = 1, 2, 3,. . . , 8), s;

vc, i = critical volume for movement i, veh/h/ln;Σvc = sum of critical volumes, veh/h/ln;

C = cycle length, snp = number of phases, andYi = change interval for movement i, s.

Equation 1 distributes the available green time to the var-ious movements in proportion to the critical volume of thecorresponding phase. It assumes a phase lost time of 4.0 s.The green intervals are based on a cycle length that is ini-tially set to equal the minimum-delay cycle length Co. Ifone or more of the resulting green intervals do not satisfythe corresponding minimum green duration, then the cyclelength should be increased and Equation 1 reapplied. Thisprocess is repeated until all green intervals exceed the re-quired minimums.

For the north-south road, the critical volume of 206 veh/h/ln is recorded in both the northbound and southbound col-umns. Then, the north-south green interval duration is com-puted as 8 s (= 206/750 � [46 − 12] + 4 − 5). Unfortunately,this duration is less than the minimum green duration of 16 s. To resolve this deficiency, the cycle length is increasedand all green intervals are recalculated (using Equation 1).This process is repeated until the green interval durationsfor each movement just satisfy the minimum time needed.For this example, a cycle length of 75 s is found to yieldgreen intervals that satisfy the corresponding minimums. Itshould be noted that a cycle length that is different from theminimum-delay cycle length will likely produce longerdelays; however, this is an unavoidable consequence of satis-fying the minimum green requirement.

Gv

vC n Yi

c i

cp i= −( ) + −

∑, ( )4 4 1

50

Actuated Phase Settings and Detection Design. Chap-ter 11 of the Manual of Traffic Signal Design (MTSD) (32)describes procedures for determining the minimum green,maximum green, and unit extension settings for an actuatedmovement. The guidance provided in the MTSD for the min-imum green setting is consistent with that provided in a pre-vious section, Minimum Green Time. The guidance providedfor the maximum green setting is based on the use of theminimum-delay pretimed green interval duration. Specifically,the MTSD suggests that the maximum green setting for amovement can be estimated by computing the minimum-delaygreen interval and then multiplying it by a factor ranging from1.25 to 1.50.

Table 3-9 illustrates the recommended procedure for usingthe Controller Setting Worksheet to determine reasonable ac-tuated controller settings for the engineering study. This pro-cedure requires that Steps 1 through 6 from the previous sec-tion are completed first. Then, Steps 7 though 9 from in thissection would be completed. As noted previously in the sec-tion, Cycle Length, the cycle length obtained from Figure 3-6is assumed to equal the average cycle length achieved byactuated operation.

The example intersection used in the preceding section isalso used in this section to demonstrate the procedure steps.The pedestrian provisions noted previously for the “pre-timed” example intersection are modified for this demon-stration. Specifically, pedestrian buttons are provided forall approaches; therefore, pedestrian considerations will notaffect the through phase duration.

Step 7. Compute Minimum-Delay Green Intervals. As afirst step in determining the maximum green setting, the lanevolumes are transferred from the Critical Volume Worksheetto the first row in the Actuated Phases Section of the Con-troller Setting Worksheet. These volumes are then used tocompute the minimum-delay green interval for each actuatedmovement. This green interval is computed from Equation 1with the cycle length equal to the value obtained from Fig-ure 3-6 (i.e., C = Co).

For the example intersection, the lane volume of 403 veh/h/ln for the westbound through movement yields a minimum-delay green interval of 17 s (= 403/750 [46 − 12] + 4 − 5).

Step 8. Determine Maximum Green Setting. The maxi-mum green setting for each movement is based on consider-ation of the minimum-delay green interval and the minimumgreen setting. Two values are computed. One value is equalto the minimum-delay green interval multiplied by a factorof 1.3. The second value is equal to the minimum green set-ting plus 12 s. The maximum green setting is set to equal thelarger of these two values.

For the westbound through movement, the maximum greensetting is found to be 22 s which is the larger of 22 (= 1.3 � 17)and 20 (=8 +12). Following this approach, the maximum greensetting for each of the other movements is found to equal 20 s.

Step 9. Define the Detector Design and Unit Extension.The unit extension setting is based on the detector locationand operation. It is desirable that local practice regardingdetector location and operation be consulted at this pointbecause of the numerous combinations possible. If informa-tion on local practice is not available, the following guide-

51

lines can be used to define detector locations, operation, andunit extension settings that are consistent with the level ofdetail needed for the engineering study.

Two detector design options are described in the remainderof this section for the purpose of determining a reasonabledetector location, operation, and unit extension setting. One

TABLE 3-9 Controller setting worksheet with calculations for actuated control

design option applies to (1) low-speed [i.e., 70 km/h (45 mph)or less] through movements and (2) protected left-turn move-ments. The other option applies to high-speed through move-ments. Both options assume presence-mode detector opera-tion. It should also be noted that both options have a maximumallowable headway of 3.0 to 3.5 s. Headways in excess of thismaximum would need to occur before a phase could gap-out.This maximum headway is consistent with that of the “typi-cal” detection designs described in the MTSD (32).

For low-speed through movements and protected left-turnmovements, it is suggested that a continuous stop line detec-tion area (comprising one long loop or several 2-m (6-ft) loopdetectors) be used. This detection area would be 15 m (50 ft)in length. The unit extension would range in value from 0.5to 1.9 s, depending on approach speed. Unit extension valuesfor typical approach speeds are listed in Table 3-10. For theexample intersection, the north-south road’s 40 km/h (25 mph)approach speed would require a 1.2-s unit extension. Simi-larly, the east-west road’s 60 km/h (40 mph) approach speedwould require a 1.8-s unit extension.

For high-speed through phases, it is suggested that twoadvance loop detectors and a stop line detection area be usedfor the engineering study evaluation. Both advance detectionsare 2 m (6 ft) in length. One advance detector is located at apoint 4.5 s travel time from the stop line. A second advanceloop is located 3.0-s travel time from the stop line. The equiv-alent travel distances are listed in Table 3-10 for a range ofapproach speeds. The stop line detection area is 15 m (50 ft)in length and uses the call-delay feature to effectively disablethe detector’s operation during green, after the initial queueclears. The call delay setting for the stop line detection area islisted in Table 3-10. The unit extension for all high-speeddesigns is 1.5 s.

52

Controller Settings for Coordinated Operation

When a traffic signal is proposed to be added to an arte-rial or network street system, it should be coordinated withthe upstream and downstream signals. When possible, coor-dination should be provided for both travel directions; how-ever, it may not be possible to achieve two-way coordinationfor some combinations of signal spacing, cycle length, andtraffic speed.

The components of the coordinated timing plan vary de-pending on whether a pretimed or semi-actuated control modeis used; however, all modes include the yellow and all-redintervals for each phase. A pretimed timing plan includesdefining the cycle length, reference phase offset, and greeninterval duration for each phase. A semi-actuated plan in-cludes defining the system cycle length, force-off, and yieldpoints. It also includes minimum green, maximum green, andunit extension settings for the actuated (non-coordinated)phases.

This section describes a reasonable procedure for estab-lishing a coordinated timing plan at a level of detail suitablefor the engineering study. This procedure can be used if infor-mation on local practice regarding signal coordination is notavailable. Although this procedure represents a “manuallyapplied” process, it is anticipated that the analyst will preferto automate the coordinated timing plan development processto the extent possible by selecting a software analysis tool (inStep 2, Select Analysis Tool) that has this capability.

The procedure described in this section extends the ma-terial described in the previous two sections (Basic ControllerSettings and Controller Settings for Isolated Operation). Assuch, the worksheets completed in the previous section willneed to be completed for the subject intersection when it

TABLE 3-10 Unit extension settings for suggested detector designs1

is part of a coordinated system. This section describes onlycoordination-related extensions to the material described inthe previous sections.

Cycle Length and Offset. The cycle length used for thesubject intersection C should be set equal to the existing sig-nal system cycle length Cs. This cycle length should be com-pared with the minimum delay cycle length Co (identified inStep 6). If Cs is less than 0.75 Co, then the subject intersection’sgeometry may need modification (e.g., lanes added) to lowerCo. The goal is to reduce the subject intersection’s minimum-delay cycle length until the existing system cycle lengthexceeds 75 percent of Co.

The term “relative” offset is used in this section. It rep-resents the time between the start of through green at twoadjacent intersections (one of which would be the subjectintersection). The relative offset needed to provide two-way progression through the proposed signal can be esti-mated as one of two values. Either it will be about 0.5 C orit will be about 0.0 s. Both of these offsets should be eval-uated in terms of the platoon arrival times from the twoupstream intersections. The preferred offset will be the onethat has vehicles arriving on green in both directions. As analternative, one-way progression can be provided in thepeak travel direction by using a relative offset equal to thetravel time.

The remainder of this section describes a technique forselecting a system cycle length and a relative offset. Thistechnique can be used when there is some flexibility in set-ting the system cycle length. It is applicable to the followingtwo cases:

1. The major road currently has only one signalized inter-section in the vicinity of the proposed signal, or

2. The existing system cycle length can be changed to pro-mote good two-way progression through the proposedsignalized intersection.

Each of these cases is discussed separately in the paragraphsthat follow.

The cycle-selection technique is based on the use of eithera “single alternate” or “double alternate” progression scheme.Each scheme has a unique type of offset relationship be-tween adjacent intersections. The single alternate schemeuses a relative offset that equals the travel time betweenintersections. The double alternate scheme uses a relativeoffset equal to 0.0 s between one pair of intersections and arelative offset equal to the travel time between the next pairof intersections.

Both schemes are based on the assumption that the throughgreen interval and the cycle length are the same at each inter-section. However, small differences in green interval durationdo not preclude the use of these techniques. Another assump-tion inherent to these schemes is that the distance (or spacing)is constant between intersection pairs. Both schemes are

53

shown in Figure 3-9 in terms of a series of intersectionsspaced L units apart with each intersection operating at a cyclelength of C s.

As the trends in Figure 3-9 indicate, the single alternatescheme is more efficient than the double alternate schemebecause the single alternate scheme has a progression bandthat is equal to the full width of the green interval. The bandfor the double alternate scheme is less than the width of thegreen interval; the amount of reduction depends on the traveltime between the intersections. In general, the double alter-nate scheme provides reasonably good progression when thegreen-to-cycle-length ratio G/C for the major-road throughphase is 0.5 or more.

Case 1. Only One Signalized Intersection in the Vicinityof the Proposed Signal. As a first step, the distance betweenthe existing and proposed intersections must be identified.Then, this distance is converted into a spacing coefficient SC using the “single alternate” trend line in Figure 3-10a.Finally, this coefficient is used with Figure 3-10b to deter-mine the cycle length for a given average running speed onthe major road. The relative offset between the throughmovement phases at the two intersections would be equal tothe travel time.

Figure 3-9. Two offset relationships that provide goodtwo-way progression.

If the cycle length obtained by the aforementioned techniqueis unacceptably short, the process should be repeated using thedouble alternate trend line. In this situation, the relative offsetbetween the through phases at the two intersections would beequal to 0.0 s.

To illustrate the Case 1 technique, consider that a newsignalized intersection is proposed to be located 400 m(1,300 ft) from a nearby signalized intersection. No othersignals exist within 1.5 km (4,900 ft). The major-road run-ning speed is 70 km/h (45 mph). Figure 3-10a indicates thatthe spacing coefficient is 400 for the single alternate pro-gression scheme. Figure 3-10b indicates that a spacing co-efficient of 400 and a speed of 70 km/h (45 mph) require acycle length of 40 s. The offset would equal the travel timeof 20 s. For this example, 40 s is determined to be too shortfor a system cycle length. Thus, Figure 3-10a is revisited tofind that the double alternate scheme will yield a coefficientof 800. A check of Figure 3-10b indicates that this coeffi-cient, when combined with the 70 km/h speed, yields a

54

more reasonable cycle length of 80 s. The offset wouldequal 0.0 s. As noted previously, the double alternate schemeis most effective when the G/C ratio is 0.5 or more for allthrough phases.

Case 2. Existing System Cycle Length Can Be Changed toPromote Progression. The first step is to identify the dis-tance between the existing signalized intersections. If thisdistance is not constant among intersection pairs, then thedistance between the two signalized intersections that boundthe proposed signal should be used. This distance is thenconverted into a “spacing coefficient” using the “singlealternate” trend line in Figure 3-10a. Next, this coefficientis used with Figure 3-10b to determine the ideal cycle lengthfor a given average running speed on the major road. As forCase 1, if the cycle length is unacceptably short, then the dou-ble alternate scheme should be used to determine the cyclelength. Regardless of the scheme selected, the cycle lengthobtained from Figure 3-10b should be used when the pro-posed signal is installed, the relative offset to the proposedsignal would equal 0.0 s, and the G/C ratio for the throughphases should be 0.5 or more.

One exception to the technique described in the previousparagraphs exists when the distance between existing inter-sections is 800 m (2,600 ft) or more and the proposed signal isto be located at a point midway between existing intersections.In this special case, a single alternate scheme will generallyprovide very good two-way progression. To determine theproper cycle length, the spacing coefficient is defined to equalthe distance between the proposed signal and an existing sig-nal. This coefficient would then be used with Figure 3-10b todetermine the ideal cycle length for a given average runningspeed on the major road. The relative offset would equal thetravel time.

To illustrate the Case 2 technique, consider that a new sig-nalized intersection is proposed to be located midway be-tween two existing signalized intersections that are spaced800 m apart (2,600 ft). The major-road running speed is 70 km/h (45 mph). Figure 3-10a indicates that the spacingcoefficient is 800 for the single alternate progression scheme.Figure 3-10b indicates that a spacing coefficient of 800 anda speed of 70 km/h (45 mph) require a cycle length of 80 sThis cycle length is acceptable, so the proposed intersectionshould be designed to operate with a cycle length of 80 s. Inaddition, all through movement phases should have a G/Cratio of 0.5 or more to ensure reasonable progression band-width. Finally, the relative offset to the proposed intersectionshould be 0.0 s.

The example intersection in the preceding paragraph isalso a candidate for consideration of the aforementionedexception to Case 2. The distance between the proposedand existing signals would equal 400 m (1,300 ft). This dis-tance equates to a spacing coefficient of 400. Figure 3-10bindicates that this coefficient corresponds to a 40-s cycle

Figure 3-10. Relationship between spacing, progressionscheme, speed, and cycle length.

length for a speed of 70 km/h (45 mph). The offset wouldequal the travel time of 20 s. The use of a single alternatescheme is attractive because it provides very good pro-gression; however, it can only be used for this example ifthe intersections can operate effectively at a relatively short,40-s cycle length.

Pretimed Phase Interval Durations. If the intersectionsin the signal system are pretimed, the procedure described inthe section, Controller Settings for Isolated Operation, can beused to develop the timing plan. Specifically, the proceduredescribed for pretimed intersections can be used to determinethe appropriate phase interval durations. One exception to thisprocedure relates to the cycle length used in the Controller Set-ting Worksheet (i.e., Table 3-8). The cycle length used for allcalculations should equal either (1) the existing system cyclelength or (2) the cycle length obtained from the proceduredescribed in a previous section, Cycle Length and Offset. Thecycle length selected by either means would then be used tocompute the pretimed phase durations in the worksheet.

Actuated Phase Settings and Detection Design. If theproposed intersection will use semi-actuated control in a coor-dinated arterial system, the procedure described in a previoussection (Controller Settings for Isolated Operation) should beused to determine the appropriate phase settings and minormovement detection design. Specifically, the analyst shouldcomplete the Controller Setting Worksheet (i.e., Table 3-9)using the Actuated Phases section to compute the maximumgreen settings for the non-coordinated phases. The cyclelength used for all calculations should equal either (1) theexisting system cycle length or (2) the cycle length obtainedfrom the procedure described in a previous section, CycleLength and Offset.

Detectors should be used for all non-coordinated phases(i.e., all left-turn and minor-road through phases). If localpractice on detection design is not known, the “low-speed”design described previously should be used. Specifically, thisdesign would consist of a continuous stop line detection area(comprising one long loop or several 2-m loop detectors). Thedetection area would be 15 m (50 ft) in length. The unit exten-

55

sion would range in value from 0.5 to 1.9 s, depending onapproach speed. Unit extension values for typical approachspeeds are listed in Table 3-10.

Force-Off and Yield-Point Settings. Semi-actuated coor-dinated systems require specification of force-off and yieldpoint settings to ensure smooth traffic progression. These set-tings represent fixed points during the cycle and are referencedto the start of the major-road through phase at the “master”intersection (assumed to be the existing upstream intersectionfor the purpose of the engineering study). To determine thesesettings, the analyst should complete the Controller SettingWorksheet (i.e., Table 3-8) using the Pretimed Phases sectionto compute equivalent pretimed durations for all signal phases.

The yield-point YP setting defines when the major-roadthrough phase can be terminated (so as to serve waitingminor-movement vehicles) without disrupting progression.The yield-point value can be set to equal the relative offsetplus the equivalent pretimed green duration of the major-road through movement phase (i.e., YP2,6 = Offset + G 2,6).This relationship is shown in Figure 3-11 for three-phaseoperation with the major-road through phase denoted bymovement numbers 2 and 6.

The force-off settings FOi define when each minor-movement phase should be terminated so that all phases canbe served and green can be returned to the major-road throughphase in time to serve an arriving platoon. One force-off set-ting is computed for each minor-movement phase. The force-off setting for the minor phase that follows the major-roadthrough phase is computed as the sum of the yield-point,major-road change interval, and minor-phase equivalent pre-timed green duration (i.e., FO1 = YP2,6 + Y2,6 + G4,8). If a sec-ond minor phase exists, then its force-off would be computedas the sum of the previous phase’s force-off setting, its changeinterval, and the second minor-phase equivalent pretimedgreen duration (i.e., FO2 = FO1 + Y4,8 + G1,5). This processwould repeat for all additional minor-movement phases.

Example. For example, consider the example intersectionpreviously considered in the development of Tables 3-7 and3-8. This intersection has a leading, protected left-turn phase

Figure 3-11. Relationship between offset, cycle length, yield-point,force-off settings.

on the east-west road. Assume that a 75-s cycle length witha 32-s relative offset was determined to provide good two-way progression. Based on this assumption, the equivalentgreen durations shown in Table 3-8 can be used to determinethe yield-point and force-off settings. The yield point for theeast-west through phase can be computed as 65 s (= 32 + 33).The next phase to occur in the phase sequence is the north-south phase. The force-off for this phase is computed as 86 s(= 65 + 5 + 16). As this value exceeds the cycle length, oneincrement of the 75-s cycle length is subtracted to obtain a“relative” force-off setting of 11 s. The last phase to occur isthe east-west left-turn phase. The force-off setting for thisphase is computed as 27 s (= 11 + 5 + 11).

Guidelines for Use of Stochastic Simulation Models

Overview

This section describes guidelines for the use of stochasticsimulation models. These models recreate the random eventsthat define an individual vehicle’s journey through the simu-lated street network. Although the simulation of randomevents is very realistic, it results in an element of uncertaintyin the output measures of effectiveness (MOEs). This uncer-tainty stems from the fact that the sequence of random eventschanges from one simulation run to the next (assuming adifferent random number seed is used each run) and pro-duces slightly different MOE values at the end of each run.This section provides guidance as to how the variability inMOEs can be minimized to the extent that reasonable con-fidence can be obtained in the conclusions reached from theengineering study.

Random Arrivals

By definition, the headway between arriving vehicles at anisolated intersection is a random variable that is exponen-tially distributed. An accurate assessment of isolated inter-section delay requires that the simulation model replicate thisbehavior. The analyst should verify that arrival headways onthe isolated intersection approach vary with some degree ofrandomness. Some clue as to the variability of arrivals can befound by observing the vehicle headways in the graphic ani-mation package that accompanies the simulation model. How-ever, a model’s ability to generate random headways should beverified by consultation of its user’s manual.

Concerns about random arrival headways are also relevantto the evaluation of systems of signals. Specifically, arrivalheadways to the external approaches of intersections locatedalong the arterial street (or around the network) should followthe Negative Exponential distribution. External approachesare those approaches that do not have an upstream signalizedintersection.

56

Random Number Seed

Most simulation models use one user-specified, randomnumber seed to determine the sequence of simulated ran-dom events. The analyst must change this seed with eachnew simulation run to ensure that the random processes inthe system are fully reflected in the output MOEs. If thesimulation model uses different seed numbers to controldifferent aspects of the simulated events, then all seeds shouldbe changed with each new simulation run.

Several techniques can be used with simulation modelsto minimize or isolate the variability due to some randomevents. For example, one simulation model allows the ana-lyst to specify that the same drivers, vehicles, and routes areused for each simulation run while still allowing driverdecisions to be randomly selected. This approach helps toreduce the simulation time needed to identify whether sig-nificant differences in delay exist among alternative controlstrategies. However, it may also bias the estimate of a move-ment’s true delay when several runs are made and the resultsaveraged because only a subset of the driver population willbe represented in the averaged values. In short, variance-reduction techniques can be useful if they are properlyapplied. Their proper application requires an understandingof the underlying statistical issues and the questions forwhich variance-reduction can be helpful in answering. Ifthere is any uncertainty about the use of variance-reduction,it is recommended that the guidance offered in the previousparagraph be followed.

Minimum Simulation Initialization Time

The initialization time used for the simulation relates to theperiod of time that elapses (relative to the start of the simu-lation run) before the model starts to collect MOE statistics.In general, the first few vehicles to enter the simulated net-work do not experience the level of delay and interaction thatis experienced by subsequent vehicles. Thus, these initialvehicles may bias the MOE statistics if they are included inthe MOE calculations.

To minimize the potential bias resulting from “start-up”effects, it is suggested that the analyst set the initializationtime to a value that exceeds the travel time between the twomost-distant points in the simulated network. This approachis intended to allow the simulated street network time to“fill” with vehicles before statistics are collected. Some sto-chastic simulation models automatically define (and apply)an initialization period based on the aforementioned rule.

Simulation Period and Run Duration

As mentioned previously, the variability in the predictedMOEs is a result of randomness in the simulated system. This

variability effectively reduces the precision of the predictedvalues. However, a desired precision can be achieved byincreasing the number of observations included in the aver-age MOEs reported by the simulation model. Several tech-niques are available for this purpose; however, the simplestand most direct technique is to increase the simulation perioduntil the desired precision is achieved.

In this guide, the “simulation period” is defined as the totaltime a given evaluation period is simulated, as measured bythe “clock” in the simulation software. The simulation periodmay exceed the period represented by the evaluation period(e.g., morning peak hour). The next three sections describe aprocedure for determining the simulation period. This proce-dure is based on three assumptions: (1) that delay is the basisfor evaluating the acceptability of an alternative; (2) that thedelays to individual traffic movements will be comparedwith a threshold value [e.g., a specified delay level-of-servicevalue obtained from the Highway Capacity Manual 2000(15)]; and (3) that the overall intersection delays will be com-pared among alternatives with the lower delay alternativebeing given preference.

Run Duration. The simulation period should consist ofone or more “runs,” with each run having a 1-hr duration.Thus, if the evaluation period is the morning peak hour andit is determined that a 2-hr simulation period is needed toachieve the desired precision, then the simulation periodshould consist of two runs of 1-hr duration each. The desiredMOEs would be computed as the average of the individualMOEs from each run.

The benefits of using a 1-hr run duration are that (1) it isconsistent with the evaluation period used in the engineeringstudy and (2) it provides a common time basis for comparisonwhen one or more movements are oversaturated. With regardto this second point, a fixed run duration facilitates the com-parison of delay among alternatives because it accounts for thetime dependency associated with oversaturated movementqueues (when they exist).

Other approaches to defining run duration are possible. Forexample, simulating volume patterns that occur during sev-eral, consecutive 1-hr periods would provide a better estimateof delay when an intermediate hour has one or more oversat-urated movements. However, this approach would requireconsiderably more analysis effort than needed for the engi-neering study. It is believed that the use of a common, 1-hrduration for each simulation run provides the best balancebetween the effort required for the engineering study and theprecision needed in the analysis results.

Minimum Simulation Period Based on IndividualMovement Delay. One condition used to determine if aproposed alternative is “effective” is that each traffic move-ment has an acceptable level of service. This verificationrequires the comparison of the predicted movement delaywith a specified threshold delay value. In this situation, it is

57

important to know if the predicted delay is truly larger thanthe threshold value and that any difference is not due onlyto random variation. The ability to make this determinationcan be enhanced by providing a sample size that is largeenough to minimize random influences.

Figure 3-12 can be used to determine the simulation periodneeded to limit the uncertainty in the delay estimate to 10 per-cent or less. The 10-percent trend line reflects a 90-percentconfidence level that the true delay is no more than 10 percentlarger than the model-predicted delay. The predicted delaycan be inflated by 10 percent and the result compared with thespecified threshold delay value. Statements can then be madeabout whether the threshold value has likely been exceeded.The development of Figure 3-12 is described elsewhere byBonneson and Fontaine (20).

A 10-percent error limit is recommended for the engi-neering study. Error limits other than 10 percent can beobtained by multiplying the time obtained from Figure 3-12by the factor: f = (10 / new error percentage)2. For example,a 20-percent error limit would require a simulation periodthat is one-fourth the duration of that needed for a 10-percenterror limit.

Minimum Simulation Period Based on Average Inter-section Delay. The comparison of delays between one ormore alternative intersection improvements is also an impor-tant consideration in the engineering study. Specifically, it isimportant to know that the predicted average intersectiondelay for one alternative is truly less (or more) than that pre-dicted for another alternative and that the difference is notdue to random variation. The ability to make this determina-tion can be enhanced by providing a sample size that is largeenough to minimize random influences. For this discussion,an existing intersection is also considered as an alternative(i.e., the “do-nothing” alternative).

Figure 3-13 can be used to define the minimum simulationperiod needed to determine when one alternative is trulyoperating with less overall delay than another alternative.

Figure 3-12. Minimum simulation period whencomparing delay to a threshold value.

The difference in average intersection delay among the twoalternatives is reflected in the “delay ratio” term, which isthe ratio of the smaller delay to the larger delay. The trendline indicates the simulation run time needed to have 90 per-cent confidence that the difference between a pair of pre-dicted delays is significant. The development of Figure 3-13is described elsewhere by Bonneson and Fontaine (20).

Procedure. The procedure for determining the time ele-ments of the simulation process includes consideration of theminimum simulation period needed for each traffic movementand for the overall intersection. The largest of these minimumswould then be used to define the total simulation time period.The minimum simulation period is then divided into a seriesof individual simulation runs of 1-hr duration.

As a first step, Figure 3-12 is consulted for each trafficmovement of interest for a given alternative. Only trafficmovements that have exclusive lanes can be examined inde-pendently. Movements that share a lane should be combinedand examined as a single movement. This process is repeatedfor each alternative.

As a second step, Figure 3-13 is consulted once for eachalternative using a flow rate that corresponds to the total enter-ing volume for the respective alternative. One simulationperiod is obtained from the figure for each alternative.

As a third step, the largest single time period obtained fromthe first and second steps is used to define the minimum sim-ulation period used. This period is then rounded to the near-est number of whole hours (e.g., 1.6 h is rounded to 2.0 h). Aseries of 1-hr simulation runs are then completed so that thetotal simulation time equals the simulation period. Each runshould be based on a different random number seed. Themovement and intersection delays obtained from each run arethen averaged to produce the desired MOEs.

Example. To illustrate the procedure, consider the inter-section shown in Figure 3-14 (volumes in veh/h). The inter-

58

section is currently unsignalized; however, a traffic signalalternative is being considered. A simulation analysis is to beconducted to determine the delay associated with each alter-native. The analyst wants to know if movement delays exceeda threshold delay level of 55 s/veh and if overall delay dif-ferences of 10 percent or more (i.e., a delay ratio of 0.9) arelikely to be significant from a statistical standpoint.

The first step is to determine the minimum simulationperiod required for threshold assessment. This time periodis dictated by the lowest movement volume because it willrequire the longest simulation time. The 50-veh/h left-turnmovements on the north and south approaches share a lanewith the through movements and thus should be combinedwith the through movements. This combination yields shared-lane volumes of 100 and 150 veh/h for the north and southlegs, respectively. After reviewing all volumes, the 75-veh/hleft-turn volumes on the east and west approaches are thelowest volume movement at the intersection and dictate theminimum simulation period based on movement consider-ations. Figure 3-12 indicates that a minimum simulationperiod of 2.2 hr would be needed based on individual move-ments.

The second step is to determine the minimum simulationperiod required for the comparison of alternatives. This pe-riod is dictated by the total entering volume at the inter-section. This volume is 1,000 veh/h for the example inter-section. Figure 3-13 indicates that a minimum simulationperiod of 0.5 hr will be necessary for both alternatives (asthey both have an entering volume of 1,000 veh/h).

The third step is to determine the minimum simulationperiod for the engineering study. For this example, the indi-

Figure 3-13. Minimum simulation period whencomparing delays between two alternatives.

Figure 3-14. Example intersection used to illustratesimulation run control procedure.

vidual movement delay evaluations require a simulationduration of 2.2 hr and the overall intersection delay com-parisons require a simulation duration of 0.5 hr. The largerduration of these two values, 2.2 hr, dictates the simulationduration. As a final step, this duration can be rounded to 2.0 hr.In summary, the simulation period should consist of two, 1-hrsimulation runs.

At the conclusion of the two, 1-hr runs for one alternative,predicted delays of 51 and 45 s/veh are obtained for the east-bound left-turn movement. These two delays are averaged toobtain an average delay of 48 s/veh. This delay, combinedwith a 10-percent error limit, would indicate that the true

59

mean delay is less than 53 s/veh. Because 53 s/veh is lessthan the threshold of 55 s/veh, the analyst can report (with90-percent confidence) that the movement delay does notexceed the threshold level. Similar statements could be madeabout the delays for the other traffic movements.

Average overall intersection delays for the unsignalizedand signalized alternatives are computed as 50 s/veh and 40 s/veh, respectively. The corresponding delay ratio is 0.8 (= 40/50). As 0.8 is less than 0.9, the analyst can con-clude (with 90-percent confidence) that the average 10-s/vehdelay difference is statistically significant and that the twoalternatives are associated with different delays.

60

CHAPTER 4

ALTERNATIVE SELECTION

This chapter describes the alternative selection stage of theengineering assessment process. This process is used to eval-uate problem intersections and identify effective improvementalternatives. During the alternative selection stage, the alter-natives identified at the end of the engineering study stage are evaluated to determine their effect on traffic operations,safety, and other factors. The alternative determined to be the “best overall alternative” based on this evaluation is thenrecommended for implementation at the subject intersection.

PROCESS

Overview

The alternative selection stage consists of three steps. Thesesteps, described in this chapter, are

1. Identify impacts,2. Select the best alternative, and3. Document the study

In the first step, each alternative that was deemed to be“effective” at the conclusion of the engineering study stage is evaluated to determine its effect on traffic operations andother factors. In the second step, these effects are aggregatedand compared in order to determine the best overall alterna-tive. Finally, in the third step, the analyst documents the studyprocess, the relevant findings, and the recommended courseof action in a study report.

The objectives of the alternative selection stage are (1) todefine a rational process for selecting the best overall alterna-tive and (2) to document the results of the assessment processin a study report. These objectives are achieved through thedevelopment and application of a uniform procedure for alter-native selection and through the agency’s formal documenta-tion of the study process. The steps involved in this stage aredescribed in the remainder of this chapter.

Step 1. Identify Impacts

This step is intended to assist the analyst in making a com-prehensive assessment of all effects that would result from theimplementation of each alternative. This assessment consistsof the following two tasks:

a. Identify Decision Factorsb. Assess Degree of Impact

In the first task, a list is developed that includes the deci-sion factors that will be considered in the selection process.Then, in the second task, the engineer systematically evalu-ates the effect of the alternative on each factor. This evalua-tion is based on a quantitative assessment of the degree ofeffect each alternative would have on a given factor.

1-a. Identify Decision Factors

The factors considered in the alternative selection processmay include traffic operations, traffic safety, constructioncost, aesthetics, environment, and right-of-way requirements.The specific factors considered for a given project will dependon the conditions present at the problem location, publicawareness of the problem, and size (or extent) of the proposedimprovements. For small projects with little visibility, theengineer may choose to focus on the traffic operations andsafety impacts. Cost factors may also be considered for mod-erately sized projects. For large projects or those with con-siderable visibility, all of the factors should be considered. Inthis latter situation, the analyst may need to enlist the supportof other professionals to evaluate impacts to the environmentor adjacent property.

Agency preferences may also dictate which factors are con-sidered. These preferences may reflect the philosophy of theagency administrators, the amount of quantitative informationneeded to assess a given factor, the degree of reliance placedon engineering judgment, and the desired precision of theengineering study. Table 4-1 lists the factors that may be con-sidered in the selection of alternative improvements at prob-lem intersections (the arrangement shown is arbitrary andimplies no order of importance).

The factors listed in Table 4-1 are consulted during this task.Those factors that have particular relevance to one or morealternatives (including the existing intersection, if applicable)are identified. Traffic operations (including capacity and levelof service) should always be one of the factors considered.Engineering judgment should be used to determine if impactsto the other, non-operations-related factors are enough to war-rant their consideration. The factors selected should reflect

61

conditions at the subject intersection, adjacent properties, andthe surrounding street network (if applicable).

1-b. Assess Degree of Impact

During this task, the factors identified in Task 1-a are re-viewed and evaluated for each alternative. The evaluationshould include some assessment of the effect of an alternativeon the associated factor.

The effect of each factor should be quantified in terms of arepresentative performance measure. For example, the oper-ational effect of each alternative could reflect the total delayexperienced during the average day. If safety is evaluated, thesafety impact of each alternative could reflect the annual num-ber of crashes expected at the site. If direct-cost is evaluated,the cost impact of each alternative could reflect its initial andannual costs. If the other factors are being considered, theireffects could also reflect a quantitative performance or impactmeasure that has some economic basis.

The precision of the impact estimate should reflect a balancebetween the importance of the study and the implications ofnot identifying the best overall alternative. The operationaleffects should be estimated during the engineering study stage.If other factors are being included, their effects can be esti-mated using a combination of experience and engineeringanalysis. In this regard, an engineer who has experience esti-mating the effects of similar projects may be able to providereasonable impact estimates for the alternatives of interest.Agency reports may also be a source of information about the

effects of similar projects. The analyst may also consult the lit-erature to determine the effects of some factors (e.g., safety).

An example application of the impact assessment processis provided in Table 4-2. Two alternatives are being com-pared. Alternative 1 is estimated to have 7 vehicle-hours ofdelay during the average day, 6 crashes per year, an initial costof $400,000, and a $2,000 annual maintenance cost. Theimpacts of Alternative 2 are also listed. The crash frequencyand direct-cost estimates are approximate and based on typi-cal values from similar projects.

For the example illustrated in Table 4-2, the agency deter-mined that it was appropriate to consider only traffic opera-tions, crash frequency, and initial costs for this project. The datain this table indicate that Alternative 1 has lower delay, highercrash frequency, and higher direct cost than Alternative 2.

Step 2. Select Best Alternative

During this step, the analyst will select the best alternativebased on consideration of the assessed impacts of the variousalternatives. Initially, weights are selected for the individualfactor impacts. Then, the best alternative is selected as theone having the lowest weighted total impact.

The weights used for each factor are estimated with the samelevel of precision used to estimate the factor impacts. To pro-vide equity among the various factors considered, the weightscan be based on the annualized worth of each factor; however,this is not a requirement. For example, if total motorist delay [invehicle-hours per day (veh-h/day)] is used to quantify the

TABLE 4-1 Factors that may be considered in the alternative selection stage

TABLE 4-2 Example procedure for assessing the degree of impact

TABLE 4-4 Typical study report outline

62

TABLE 4-3 Example procedure for alternative selection

impact on traffic operations, its weight would reflect theannual cost of time per vehicle. These weights should be esti-mated by an engineer experienced in estimating the impactof similar projects.

The annualized-worth basis for determining the weights isa suggestion; it is not a requirement of this procedure. Otherweighting schemes can be rationalized and should reflect the policies of the responsible agency, the time available for

alternative selection, and the experience of the analyst. How-ever, once the weights are estimated, they should be used forsubsequent projects to provide some degree of consistency inthe recommendations.

An example application of this procedure is provided inTable 4-3. The impacts identified in Table 4-2 are used for thisexample. The weight used for the delay impact is based on anestimate of the annual cost of time per vehicle (in thousands

63

of dollars). It assumes that there are 300 equivalent averagedays per year and that the cost of time is $15 per vehicle-hour(i.e., 4.5 = 300 � 15 / 1000). Thus, the product of this weightand the total delay yields an estimate of the annual cost ofdelay per year for a given alternative. Again, these weights areillustrative; the analyst should establish appropriate weightsfor his (or her) agency following these principles.

The weight used for crash frequency reflects an equivalentcost for the average crash (in thousands of dollars) and reflectsthe distribution of fatal, injury, and property damage crashesat intersections. The initial costs (in thousands of dollars)were given a weight of only 0.1 to reflect their annualizationover a 15-year period. These weights were used to computethe weighted total shown in the last row of the table. This totalrepresents the annual cost of each alternative (in thousands ofdollars). The lower total for Alternative 2 indicates that itshould be selected (over Alternative 1).

A formal economic analysis may also be used to determinethe best alternative. This analysis has the advantage of beinga defensible, quantitative means of assessing the relativeworth of various alternatives by using financial return (in dol-lars) as a common basis of comparison. Its disadvantage isthat it requires some effort to quantify the economic worth ofeach factor considered. A procedure for conducting an eco-nomic analysis of transportation projects is described in AManual on User Benefit Analysis of Highway and Bus-TransitImprovements (34).

Step 3. Document Study

During this step, the analyst documents the results of thestudy process in a study report. This report should describe

the study process, the relevant findings, and the recommendedcourse of action. The report should also include all support-ing discussion, information, and summary worksheets. Thecontent of the report should be readily understood by otherengineers and agency administrators.

The study report represents the agency document that sup-ports the action taken at the subject intersection. As such, thisreport must contain sufficient information to justify the rec-ommended action. The typical study report outline is listedin Table 4-4. The information in this table is partially drawnfrom the Montana Department of Transportation’s TrafficEngineering Manual (35).

As indicated by the information in Table 4-4, the studyreport should contain eight sections. The information gath-ered through the conduct of the procedures in Chapters 2 and3 would be described in Sections 1, 2, 3, 5, and 6. An analy-sis of the intersection’s crash history would be included inSection 4. A procedure for conducting the crash data analy-sis is described in the Institute of Transportation Engineer’sManual of Transportation Engineering Studies (6 ).

In general, the study report should be kept brief. As a rule-of-thumb, the report (excluding attachments) should be fourto six pages in length with about one paragraph of discussiondevoted to each of the bullet items in Table 4-4. A table maybe the most efficient means of summarizing the traffic vol-umes and the crash history. A short cover letter may also beadded to document the conveyance of the report to the appro-priate agency administrator.

64

BIBLIOGRAPHY

The documents listed in this section provide useful guid-ance on the conduct of safety studies at intersections. Collec-tively, they describe how to conduct and interpret crash his-tory analyses, conflict studies, and control device compliancestudies.

Traffic Engineering Handbook. 5th ed. Pline, J.L., ed. Chapter 7“Community Safety.” Institute of Transportation Engineers,Washington, D.C. (1999).

Ogden, K.W. Safer Roads: A Guide to Road Safety Engineering.Ashgate Publishing Company, Brookfield VT, England (1996).

Manual of Transportation Engineering Studies. Robertson, D.H.,ed. Prentice-Hall, Englewood Cliffs, New Jersey (1995).

Pietrucha, M., Opiela, K., Knoblauch, R., and Cringler, K.“Motorist Compliance with Standard Traffic Control Devices.”FHWA-RD-89-103. Federal Highway Administration, Washing-ton, D.C. (1989).

Parker, M.R. and Zegeer, C.V. “Traffic Conflict Techniques forSafety and Operations: Engineer’s Guide.” FHWA-IP-88-026.Federal Highway Administration, McLean, Virginia (1988).

Graham, J.L. and Glennon, J.C. Manual on Identification, Analysis,and Correction of High Accident Locations. Federal HighwayAdministration, Washington, D.C. (1975).

65

REFERENCES

1. Manual on Uniform Traffic Control Devices-Millennium Edi-tion. Federal Highway Administration, U.S. Department ofTransportation, Washington, D.C. (December 2000).

2. Bretherton, W.M., Jr. “Signal Warrants–Are They Doing theJob?” Compendium of Technical Papers for the 1991 ITEAnnual Meeting. Institute of Transportation Engineers, Wash-ington, D.C. (1992) pp. 163–167.

3. Neudorff, L.G. “Gap-Based Criteria for Signal Warrants.” ITEJournal, Vol. 55, No. 5. Institute of Transportation Engineers,Washington, D.C. (February 1985) pp. 15–18.

4. Bonneson, J.A. and Fontaine, M.D. Assessing Traffic ControlSignal Installations Using Capacity Analysis and Simulation:Interim Report. NCHRP Project 3-58. National CooperativeHighway Research Program, Transportation Research Board,National Research Council, Washington, D.C. (November 1999).

5. A Policy on Geometric Design of Highways and Streets. Amer-ican Association of State Highway and Transportation Offi-cials, Washington, D.C. (1994).

6. Manual of Transportation Engineering Studies. D.H. Robert-son, ed. Prentice-Hall, Englewood Cliffs, New Jersey (1995).

7. Hawkins, Jr., H.G. and Carlson, P.J. “Traffic Signal Warrants:Guidelines for Conducting a Traffic Signal Warrant Analysis.”Report No. 3991-2, Texas Transportation Institute, CollegeStation, Texas (1998).

8. Williams, J.C. and Ardekani, S.A. “Impacts of Signal Installa-tion at Marginally Warranted Intersections.” Research Report1350-1F, Center for Transportation Studies, University ofTexas at Arlington, Arlington, Texas (1996).

9. Kay, J.L., Neudorff, L.G., and Wagner, F.A. “Criteria forRemoving Traffic Signals–Technical Report.” Report No.FHWA/RD-80/104. Federal Highway Administration, Wash-ington, D.C. (1980).

10. Bissell, H.H. and Neudorff, L.G. “Criteria for Removing Traf-fic Signals.” Compendium of Technical Papers for the 1980ITE Annual Meeting. Institute of Transportation Engineers,Washington, D.C. (1980) pp. 56–66.

11. Pusey, R.S. and Butzer, G.L. “Chapter 13 - Traffic Control Sig-nals.” Traffic Engineering Handbook. 5th ed., J.L. Pline, ed.Institute of Transportation Engineers, Washington, D.C. (1999).

12. Kacir, K.C., Benz, R.J., and Hawkins, Jr., H.G. “Analysis ofFlashing Signal Operation.” In Transportation Research Record1421. Transportation Research Board, National Research Coun-cil, Washington, D.C. (1993).

13. Benioff, B., Carson, C., and Dock, F.C. “A Study of ClearanceIntervals, Flashing Operation, and Left-Turn Phasing at TrafficSignals.” Report No. FHWA-RD-78-48. Federal Highway Ad-ministration, Washington, D.C. (1980).

14. Box, P.C. “Warrants for Traffic Control Devices at Low-VolumeUrban Intersections.” ITE Journal. Institute of TransportationEngineers, Washington, D.C. (April 1995).

15. Highway Capacity Manual 2000. 4th ed. TransportationResearch Board, National Research Council, Washington, D.C.(2000).

16. Cleveland, D.E., Huber, M.J., and Rosenbaum, M.J. “Chapter 9 -On-Street Parking.” Synthesis of Safety Research Related to

Traffic Control and Roadway Elements. Vol. 1, Report No.FHWA-TS-82-232, Federal Highway Administration, Wash-ington, D.C. (1982).

17. Special Report 125: Parking Principles. Highway ResearchBoard, Washington, D.C. (1971).

18. Koepke, F.J. and Levinson, H.S. “Access Management Guide-lines for Activity Centers.” NCHRP Report 348. TransportationResearch Board, National Research Council, Washington, D.C.(1992) p. 70.

19. Robinson, B. et al. “Roundabouts: An Informational Guide.”Report No. FHWA-RD-00-067, Federal Highway Administra-tion, Washington, D.C. (2000).

20. Bonneson, J.A. and Fontaine, M.D. Assessing Traffic ControlSignal Installations Using Capacity Analysis and Simulation:Final Report. NCHRP Project 3-58. National Cooperative High-way Research Program, Transportation Research Board, Wash-ington, D.C. (2001).

21. Neuman, T.R. “Intersection Channelization Design Guide.”NCHRP Report 279. National Cooperative Highway ResearchProgram, Transportation Research Board, National ResearchCouncil, Washington, D.C. (1985).

22. Hasan, T. and Stokes, R.W. “Guidelines for Right-Turn Treat-ments at Unsignalized Intersections and Driveways on RuralHighways.” Transportation Research Record 1579. Transporta-tion Research Board, National Research Council, Washington,D.C. (1997) pp. 63–72.

23. Harmelink, M.D. “Volume Warrants for Left-Turn StorageLanes at Unsignalized At-Grade Intersections.” HighwayResearch Record 211. Highway Research Board, Washington,D.C. (1967) pp. 1–18.

24. Koepke, F.J. “Chapter 10 - Access Management.” Traffic Engi-neering Handbook. 5th ed., Institute of Transportation Engi-neers, Washington, D.C. (1999) p. 328.

25. Traffic and Safety Manual of Instructions. Division of Trafficand Safety. Utah Department of Transportation, Salt Lake City,Utah (November 1990).

26. Henry, R.D. and Calhoun, J.H.L. “Peak-Hour Traffic SignalWarrant.” NCHRP Report 249. National Cooperative HighwayResearch Program, Transportation Research Board, NationalResearch Council, Washington, D.C. (1982).

27. Saka, A.A. “Microscopic Simulation Modeling of MinimumThresholds Warranting Intersection Signalization.” Trans-portation Research Record 1421. Transportation ResearchBoard, National Research Council, Washington, D.C. (1993)pp. 30–35.

28. Orcutt, F.L. The Traffic Signal Book. Prentice-Hall, Inc.,Englewood Cliffs, New Jersey (1993).

29. Hook, D. and Albers, A. “Comparison of Alternative Method-ologies to Determine Breakpoints in Signal Progression.” Com-pendium of Technical Papers for the 69th Annual Meeting of theITE. CD-ROM. Institute of Transportation Engineers, Wash-ington, D.C. (1999).

30. Sampson, J. D. “Queue-Based Traffic Signal Warrants: The4Q/6Q Warrant.” ITE Journal, Vol. 61, No. 7. Institute ofTransportation Engineers, Washington, D.C. (1999) pp. 30–36.

66

31. Gordon, R.L. et al. “Traffic Control System Handbook.”Report No. FHWA-SA-95-032. Department of Transportation,Washington, D.C. (1996).

32. Kell, J.H. and Fullerton, I.J. Manual of Traffic Signal Design.Prentice-Hall, Englewood Cliffs, New Jersey (1982).

33. Webster, F.V. Traffic Signal Settings: Technical Paper 39.H.M. Stationary Office, London, England (1958).

34. A Manual on User Benefit Analysis of Highway and Bus-Transit Improvements. American Association of State High-way and Transportation Officials, Washington, D.C. (1977).

35. Traffic Engineering Manual. Chapter 12 - Traffic SignalDesign. Montana Department of Transportation, Helena, Mon-tana (July 1999).

A-1

APPENDIX A

CASE STUDY APPLICATIONS

References to software products in Appendix A are solelyfor the purpose of providing meaningful illustrations of theanalysis tool selection process. No endorsement of a productis implied by its inclusion in the guide.

INTRODUCTION

The objective of this appendix is to demonstrate the pro-cedures described in the Engineering Study Guide (ESG )through their application to several case studies. These casestudies are based on real-world intersections; however, someaspects of their traffic demand or geometry have been mod-ified to illustrate various elements of the procedures.

Four case studies are described in this appendix. Collec-tively, these case studies show the three stages of the engi-

neering assessment process (i.e., alternative identificationand screening, engineering study, and alternative selection).They also illustrate the ability of the process to deal with bothisolated intersections and intersections within signal systems(i.e., non-isolated). The attributes of the case studies are listedin Table A-1.

The case studies were selected to demonstrate the mannerin which the ESG could be used to identify and evaluate arange of alternatives. Collectively, the case studies considerboth roundabout and conventional intersection geometry; thelatter with two-way stop, multi-way stop, and traffic signalcontrol. By coincidence, the signal alternative is considered ineach case study. However, this result should not be construedto mean that the signal alternative is a required considerationin all engineering studies. It should also be noted that the sig-nal alternative was not always found to be the best alternative.

TABLE A-1 List of case study attributes

CASE STUDY 1

SYNOPSIS

This case study illustrates an application of the EngineeringStudy Guide (ESG) to an isolated, stop-controlled intersectionwhere minor-road drivers are experiencing excessive delay.The HCS software package was used to evaluate the operationof the intersection and its proposed alternatives. The additionof right-turn bays on the north and westbound approaches andthe addition of a left-turn bay on the southbound approach wasrecommended as the best method for alleviating motorist delay.

BACKGROUND

The local transportation agency has received many com-plaints about the intersection of County Routes 21 and 27.These complaints suggest that the delay to minor-road driv-ers may be excessive. Most of these drivers are traveling toand from a manufacturing plant located about 1⁄2 mile to theeast of the intersection on County Road 27. Because of thesecomplaints, an engineering assessment was undertaken todecide if improvements to the intersection were needed.

The intersection of County Routes 21 and 27 is in a ruralcommunity with approximately 11,000 residents. The inter-section has three approach legs; one leg is stop-controlled.County Route 27 (CR 27) is the minor road; it is oriented inan east-west direction and has stop-control at the intersec-

A-2

tion. County Route 21 (CR 21) is the major road; it is ori-ented in the north-south direction. The speed limit is postedat 45 mph on both roads. All approaches have one trafficlane. Figure A-1 illustrates the intersection geometry.

STAGE 1: ALTERNATIVE IDENTIFICATIONAND SCREENING

The first stage of the engineering assessment process in-volved diagnosing the problem at the intersection and deter-mining potential corrective measures. Candidate alternativeswere identified and given a preliminary screening to decidewhether they represented viable solutions to the problem.The alternative identification and screening stage consists ofthe following steps:

1. Define problem and cause,2. Select candidate alternatives, and3. Select viable alternatives.

Stage 1: Step 1. Define Problem and Cause

Gather Information

The first step was to gather information about the inter-section. This involved checking agency records and visiting

Figure A-1. Case Study 1: Existing intersection geometry (not to scale).

the site. The data gathered during this step are summarizedin the remainder of this section.

Historic Data. This intersection had not been the sub-ject of a prior engineering study, so archival traffic datawere not available. Crash data were obtained from thestate’s Department of Public Safety. These data indicatedthat a total of five collisions occurred at the intersection inthe last 12 months (two susceptible to correction by a traf-fic signal). Regional traffic counts recorded 1 year earlierindicated that the annual average daily traffic (AADT) isabout 13,500 vehicles per day on CR 21 and 3,000 vehiclesper day on CR 27.

A-3

Observational Study. The intersection was visited duringa typical weekday to determine whether the reported delayproblem was present at the intersection. The On Site Obser-vation Report completed during this study is provided as Fig-ure A-2. The observation study revealed that drivers on theminor road did seem to experience excessive delay during thepeak hour. Much of this delay appeared to be the result ofright- and left-turn vehicles having to share one traffic lane.The study also revealed that through drivers southbound on CR 21 occasionally experienced some delay and conflictbecause of vehicles waiting to turn left. Finally, it was notedthat some conflicts occurred between northbound through andright-turn vehicles (and associated delay) when the right-turn

Figure A-2. Case Study 1: On site observation report.

vehicle slowed to turn right. However, the delay associatedwith this latter conflict appeared to be fairly small.

Site Survey. A survey was also conducted during the sitevisit. Key geometric features of the intersection noted duringthis survey are summarized in Figure A-1. Of particular noteat this location are the railroad tracks, utility lines, and anearby residence. Specifically, railroad tracks run parallel toCR 21 and are about 60 ft west of CR 21. Electrical utilitypoles are about one-half the way between CR 21 and the rail-road tracks. A private residence is in the southeast quadrantof the intersection. A driveway to this residence is 500 ftsouth of the intersection on CR 21.

Define Problem and Cause

Assess Evidence. Based on the field observations, delayto minor-road drivers seemed excessive. There also appearedto be some delay and conflict associated with the major-roadleft- and right-turn movements. This latter problem stemmedfrom the interaction between the through and turning move-ments and the fact that these movements share a commonlane on both the north and south approaches. From this evi-dence, it was concluded that a problem existed and that fur-ther study was justified.

Define Problem and Identify Cause. The observationalstudy identified three potential problems at the intersection.These problems were

• Excessive delay on the minor road,• Some delay and conflict to through and left-turn move-

ments on CR 21 southbound, and

A-4

• Some delay and conflict to through and right-turn move-ments on CR 21 northbound.

Using the information provided in Table 2-1, possiblecauses for these problems were identified. This informationis summarized in Table A-2.

Define Influence Area. The intersection is several milesfrom nearby signalized intersections, so it operates as an iso-lated intersection. The subject intersection’s influence areawas defined to consist of the intersection conflict area and a300-ft length of roadway on each approach.

Stage 1: Step 2. Select Candidate Alternatives

Identify Potential Alternatives

Table 2-2 in the ESG was consulted to determine potentialalternatives that could address the observed problems at theintersection. The alternatives identified in Table 2-2 that areapplicable to the subject intersection are summarized inTable A-3.

Organize and Select Alternatives

The potential alternatives were subjected to a preliminaryscreening to eliminate any alternatives that would not be fea-sible at the site. Judgment was used to identify those alterna-tives that were clearly not feasible because of site-specificconstraints. Based on this screening, three alternatives wereeliminated. These alternatives are listed in Table A-4.

Based on this analysis, the following alternatives werefound to merit further study:

TABLE A-2 Case Study 1: Potential problems and causes

TABLE A-3 Case Study 1: Potential alternatives

• Convert to traffic signal control,• Add left-turn bay to major road,• Add right-turn bay to major road,• Add second lane on minor road, and• Convert to multi-way stop control.

Stage 1: Step 3. Select Viable Alternatives

The candidate alternatives selected in Step 2 were moreclosely examined in this step. This examination focused onevaluating each alternative to assess its “viability” (i.e., itsability to address the observed problems effectively).

Gather Information

Identify Data. Table 2-5 in the ESG was used to identifythe data needed to evaluate the candidate alternatives. Thesedata are summarized in Table A-5.

A-5

Based on the information in Table A-5, the minimumamount of data that must be collected to evaluate all of theguidelines was determined. The following data were identi-fied as necessary for guideline evaluation:

• Major- and minor-road turn movement volume (8 hours);• Pedestrian volume (4 hours);• Gap frequency (4 hours);• 85th percentile approach speed;• Progression quality;• Minor-road delay (1 hour);

✓ • Area population;✓ • Number of lanes; and✓ • Crash history by type.

Of these data, information about the last three items hadalready been gathered in a previous step. The minimum num-ber of hours for which data would be collected is indicated inthe list above.

TABLE A- 4 Case Study 1: Alternatives eliminated

TABLE A- 5 Case Study 1: Data needed to evaluate guidelines

Collect Data. Because it was unclear when the eight high-est traffic hours occurred, turn movement volumes wererecorded for a total of 13 hours. These counts were adjusted torepresent average-day volumes using the techniques describedin Appendix C. The adjusted volumes are summarized inTable A-6.

The morning peak demand occurred between 7 and 8 a.m.,the afternoon peak occurred between 4 and 5 p.m., and theoff-peak hour occurred between 9 and 10 a.m. No pedestrianswere observed during the study. Heavy vehicles comprisedabout 1.0 percent of the traffic stream on each approach. Aspot speed study was performed for CR 21; the 85th percentilespeed was found to be 50 mph.

A stopped-delay study was performed during the after-noon peak hour (i.e., 4 to 5 p.m.). This study focused on thedelay to minor-road vehicles. The average delay during thepeak hour was 40 s/veh and the total delay was 2.3 veh-h.

Assess and Select Alternatives

Convert to Traffic Signal Control. The traffic signalwarrants in the Manual on Uniform Traffic Control Devices-Millennium Edition (i.e., MUTCD 2000) (A-2) were evalu-ated to assess the viability of the traffic signal alternative. Acomplete warrant analysis for this intersection is documentedin Appendix B. The following warrants were satisfied:

• Warrant 1: Eight-hour vehicular volume (Conditions Aand B),

• Warrant 2: Four-hour vehicular volume, and• Warrant 3: Peak hour.

A-6

The presence of conditions that could misdirect the signalwarrant check was also assessed. A review of Table 2-15 inthe ESG revealed that two conditions exist that could affectthe warrant analysis conclusions. One condition relates to thefact that there are only three approach legs at the subjectintersection. The second condition relates to the significantnumber of right-turn vehicles on the minor-road approach.

The discussion associated with Table 2-15 indicated thatthe MUTCD’s volume-based warrants may have thresholdvalues that are too low for a three-leg intersection. Based onthe guidance in Chapter 2, the major and minor roads at athree-leg intersection may need to have volumes of 1,000 and200 veh/h, respectively, for the eight highest hours before atraffic signal is likely to be justified. A comparison of thesevalues with the intersection volumes indicated that the“1,000 veh/h criteria” was not met for any of the hours thatturn movement volumes were collected.

The impact of minor-road right-turn volume on the accu-racy of the warrant check was also investigated. Figure 2-11in the ESG was consulted to determine the extent to which theright-turn volume might be reduced in the warrant check.Based on this examination, it was found that the right-turnvolume probably should not be included in the warrant analy-sis. However, when Warrants 1 and 2 were re-evaluated with-out the right-turn volume, the warrants were still satisfied.

In summary, the warrant analysis indicated that a trafficsignal might improve traffic operations at the intersection.However, guidelines in Chapter 2 indicated that the three-leggeometry of the intersection may affect the accuracy of thewarrant check conclusions. From this analysis, it was deter-mined that the operation of the traffic signal alternative wouldbe carefully examined during the engineering study stage toconfirm the benefits of signalization.

TABLE A-6 Case Study 1: Turn movement volumes

Add Left-Turn Bay to Major Road. Figure 2-5 in theESG was examined to determine whether an exclusive left-turn bay on the major road was feasible. Based on this exam-ination, it was determined that a left-turn bay was justifiedduring the peak demand hour. This finding indicated that aleft-turn bay on the major road would be desirable.

Add Right-Turn Bay to Major Road. Figure 2-6 in theESG was examined to determine whether an exclusive right-turn bay on the major road was feasible. This figure is repro-duced as Figure A-3. The northbound volume combinationduring the afternoon peak hour is indicated by a solid circle.Speed measurements on the major road indicated that the 85thpercentile speed is 50 mph. The fact that this circle lies abovethe “50 mph” trend line indicates that a right-turn bay on thenorthbound approach is a viable improvement alternative.

Add Second Lane on Minor Road. Figure 2-4 in theESG was examined to determine whether adding a second

A-7

lane (i.e., an exclusive turn bay) on the minor road was fea-sible. This examination was based on a right-turn volumethat averaged 32 percent of the approach traffic stream dur-ing the afternoon peak hour. The results of this examinationare illustrated in Figure A-4.

The solid circle in Figure A-4 represents the volumeoccurring during the afternoon peak hour. By interpolation,the circle lies slightly above an equivalent “32%” trend line.Hence, it was concluded that an additional lane on the minorroad would be beneficial. Site conditions indicated that thislane could most easily be added as a right-turn bay.

Convert to Multi-Way Stop Control. The MUTCDprovides four criteria that can be used to determine whethermulti-way stop control might improve intersection opera-tions (these criteria are also cited in Chapter 2). One of thecriteria indicates that multi-way stop control may be justi-fied when (1) the total volume entering the intersection onthe major-road approaches averages at least 210 veh/h forany 8 hours of an average day, (2) the volume of vehicles onthe minor road averages at least 140 veh/h for the same 8 hours, and (3) the average minor-road delay is at least 30 s/veh during the highest hour. These volume thresholdsrepresent “70%” levels because the major-road speed exceeds40 mph.

Table A-7 summarizes the results of the multi-way stopcriteria evaluation. Based on this examination, it was deter-mined that the multi-way stop criteria are satisfied and thatmulti-way stop control is a viable alternative at this inter-section.

Summary of Viable Alternative

Based on the assessment of viable alternatives, all fivecandidate alternatives identified in Step 2 were selected for

Figure A-3. Case Study 1: Check of need for right-turnbay on major road.

Figure A-4. Case Study 1: Check of need for second lane on minor road.

A-8

TABLE A-8 Case Study 1: Viable alternatives

TABLE A-7 Case Study 1: Check of multi-way stop control criteria

more detailed analysis in the next stage of the assessmentprocess. These alternatives are summarized in Table A-8.

STAGE 2: ENGINEERING STUDY

The operational performance of the alternatives identifiedin the alternative identification and screening stage were eval-uated in the engineering study stage. This stage consists ofthree steps:

1. Determine study type,2. Select analysis tool, and3. Conduct evaluation.

Stage 2: Step 1. Determine Study Type

The first step of the engineering study stage required adetermination of the type of study needed. Initially, the rela-tionship between the subject intersection and any upstreamor downstream signalized intersection was evaluated. Then,the analysis detail required for the assessment of each alter-native was determined.

Determine Type of Operation

The interaction between the subject intersection and anyadjacent signalized intersection was evaluated to determinewhether the operation of the adjacent intersection should beconsidered in the analysis. An investigation of the locationof these intersections revealed that the nearest signalizedintersection is 6 miles to the north. Figure 3-1 in the ESG

indicates that distances exceeding 1,800 ft (when the two-way volume is 900 veh/h) are sufficient to isolate the sub-ject intersection from the adjacent signalized intersections.

Determine Type of Evaluation

Next, the intersection was examined to determine whethera formal evaluation of alternatives was required. The ESGsuggests that an informal evaluation (i.e., implementationand field study) is possible when only one viable alter-native is applicable at an isolated intersection. However,the fact that the subject intersection is associated with sev-eral viable alternatives required the conduct of a formalevaluation.

Stage 2: Step 2. Select Analysis Tool

During this step, an analysis tool was selected for evaluat-ing the operation of the subject intersection and its alterna-tives. This selection was based on an identification of thedesired analysis tool capabilities and a comparison of this listwith the actual capabilities of the available tools. The toolselected for use was the one that provided all of the desiredcapabilities.

Identify Desired Capabilities

Table 3-1 in the ESG was consulted to determine whichanalysis tool capabilities would be required for this evalua-tion. The minimum set of capabilities needed for the analy-sis tool are summarized in Table A-9.

A-9

TABLE A-9 Case Study 1: Desired analysis tool capabilities

Evaluate and Select Analysis Tool

Three analysis tools were available for the analysis. Thesetools are HCS (version 3.1c), CORSIM (version 4.3), andTRANSYT-7F (release 8). The capabilities of these toolswere compared with the list of desired capabilities identifiedin Table A-9. Only those tools that provided all of the desiredcapabilities were considered for use in the study. Table A-10illustrates the procedure used to assess the traffic control modecapabilities of the three analysis tools. Based on this assess-ment, TRANSYT-7F was eliminated from further consider-ation because it could not model the multi-way-stop-controlalternative.

This process was repeated for the desired analysis factorcapabilities. The analysis tool used for this study would needto be able to model heavy vehicle percentage, exclusive turnbays, and traffic lanes shared by through and turning vehicles.Based on an assessment of model capabilities, it was concludedthat both the HCS and CORSIM tools could model all of the

required analysis factors and provide the desired measure ofeffectiveness.

The effort required to use HCS and CORSIM was consid-ered in making the final tool selection. In this regard, theCORSIM tool was believed to have more capability (andassociated complexity) than was needed for this analysis.Also, considerable expertise had been developed by agencystaff in the use of HCS for isolated intersection evaluation.Based on these considerations, the HCS was selected as themost appropriate tool for the evaluation.

Stage 2: Step 3. Conduct Evaluation

The final step in the engineering study stage was to deter-mine the operational performance of the viable alternatives.As a first step, the need for additional information and fielddata was reviewed. Then, the operational performance ofeach alternative was quantified and its relative effectivenesswas determined.

TABLE A-10 Case Study 1: Traffic control mode check sheet

Gather Information

The need for additional data was assessed at this point inthe study. This assessment included consideration of theinput data requirements of the chosen analysis tool (i.e.,HCS) and the data needed for the benefit-versus-cost analy-sis. Based on this assessment, it was concluded that addi-tional data did not need to be gathered.

Evaluate Operational Performance

Design Alternatives. Some preliminary design decisionswere made regarding the operation and the geometry of thealternatives prior to their evaluation. Most of this effort wasdevoted to the development of a design and timing plan forthe traffic signal control alternative. These design decisionsare documented in this section.

Signal Alternative. The signal design process included adetermination of the intersection geometry and its signal tim-ing. Guidelines in Chapter 3 of the ESG indicate that anexclusive left-turn lane (or bay) may be needed when the left-turn volume exceeds 100 veh/h during the peak hour.Because the left-turn volume of 146 veh/h exceeds this thresh-old value, a left-turn bay on the southbound approach isincluded in the signal alternative.

Guidelines regarding the need for right-turn lanes are pro-vided in Figure 3-3 of the ESG. A check of these guidelinesindicated that a right-turn lane was not needed for the signalalternative and that one through lane on each approach wouldbe adequate.

Guidance regarding the operation of the signal controller isprovided in Table 3-5 of the ESG. This guidance indicated thatactuated operation is appropriate for isolated intersections.

Guidance regarding the need for left-turn phasing is pro-vided in Figure 3-5 of the ESG. A check of the informationin this figure indicated that left-turn phasing was not needed.Thus, the signal phase sequence should consist of two throughphases: one for the northbound and southbound approachesand one phase for the westbound approach.

Traffic volumes recorded for the afternoon peak periodwere used to determine reasonable controller settings. Theanalysis of the afternoon peak hour volumes is documentedin Tables A-11 and A-12. The settings determined from thisanalysis were reasoned to be sufficiently accurate for the engi-neering study evaluation; however, it was recognized thatthey might need to be refined if the signal alternative is ulti-mately selected.

The HCS does not provide for the direct input of minimumand maximum green times; instead, it requires entry of averagegreen interval duration when evaluating actuated phases. Thesedurations were computed using the procedures described inChapter 16, Appendix II, of the Highway Capacity Manual2000 (A-3). Once computed, they were used with the HCS todetermine the average delay during each analysis period.

A-10

TWSC-1 Alternative. According to Table 2-13 of theESG, the right-turn bay length on the minor road should belong enough to provide for right-turn vehicle storage. Fig-ure 2-8 of the ESG indicated that right-turn bay storagelength should be at least 25 ft. However, agency practice isto use 50 ft as a minimum bay length.

TWSC-2 Alternative. The length of the major-road left-turn bay at the unsignalized intersection was based on con-sideration of left-turn volume and approach speed. Accordingto Table 2-13 of the ESG, a left-turn bay should be longenough to provide for the deceleration and storage of left-turnvehicles. Figure 2-7 indicated that a storage length of about30 ft would be necessary. Also, Figure 2-9 indicated that anadditional 290 ft was needed to provide adequate decelerationdistance for an 85th percentile speed of 50 mph. Thus, themajor-road left-turn bay was designed to be 320 ft in length.

TWSC-3 Alternative. Table 2-13 of the ESG provided someinformation regarding the appropriate length of a right-turnbay on the major road. According to this table, a right-turn bayshould be long enough to provide for the deceleration of right-turn vehicles. As noted in the previous section, a bay length ofabout 290 ft is needed for turn vehicle deceleration. Thus, themajor-road right-turn lane was designed to be 290 ft in length.

Evaluate Alternatives. The HCS was used to evaluateeach viable alternative for the morning peak hour, afternoonpeak hour, and the off-peak hour. The individual movementdelays were highest during the morning peak hour; thesedelays are summarized in Table A-13.

The data in Table A-13 indicate that the more complexalternatives (i.e., Signal and TWSC-3) offer the least overalldelay during the morning peak hour. However, the low delayfor the Signal alternative is achieved at the “expense” ofincreased delay to the major-road through and right-turnmovements. It should be noted that the delays to throughvehicles due to right-turns from the major road were not esti-mated by the software; however, they were observed to bequite small and their omission from the delay summary wasdetermined to have negligible effect on the evaluation.

The variation in intersection delay over time is shown inFigure A-5. The trends in this figure indicate that the multi-way-stop-control alternative (MWSC) is consistently asso-ciated with the largest overall delay, reaching level of ser-vice (LOS) D during the peak hour. In contrast, the TWSC-3alternative (i.e., add a right-turn bay on the minor road andleft- and right-turn bays on the major road) would yield thelowest delay. The Signal alternative had the least delay vari-ation throughout the day; however, it causes more delay dur-ing the off-peak hour than most of the other alternatives.

Determine Alternative Effectiveness

The findings from the operations analysis were evaluatedto determine the effectiveness of each alternative. This deter-

A-11

TABLE A-11 Case Study 1: Critical volume worksheet

mination was made based on the definitions of “acceptablelevel of service” and “acceptable operation” as defined inChapter 3 of the ESG. Table A-14 summarizes the effective-ness of each alternative.

Only two alternatives were determined to be ineffective.They are (1) the current intersection alternative (TWSC-0)and (2) the multi-way-stop-control alternative (MWSC). The

signal alternative and all three of the turn-bay-related alter-natives were determined to be effective.

STAGE 3: ALTERNATIVE SELECTION

During the alternative selection stage, the alternativesadvanced from the engineering study stage are reviewed for

TABLE A-12 Case Study 1: Controller setting worksheet

A-12

their potential impact on traffic operations, safety, and theenvironment. The alternative selection stage consists of thefollowing three steps:

1. Identify impacts,2. Select best alternative, and3. Document study.

Stage 3: Step 1. Identify Impacts

Identify Decision Factors

Several factors were identified that might influence al-ternative selection. These decision factors include trafficoperations, traffic safety, and direct cost. Impacts to the envi-

ronment and area aesthetics are negligible for the proposedalternatives.

Assess Degree of Impact

Each alternative’s impact on traffic operations, safety, andcost was evaluated. The traffic operations impacts werequantified using motorist delay. The total delay during theaverage day was computed by assuming that each peak-hourdelay was incurred for 2 hours and that the off-peak delaywas incurred for 18 hours. For the Signal alternative, totaldelays of 3.1, 1.4, and 3.1 veh-h/h were predicted for themorning peak hour, off-peak hour, and afternoon peak hour,respectively. The average-day delay for the Signal alternativewas estimated as 38 veh-h/day (= 2 � 3.1 + 18 � 1.4 + 2 � 3.1).The total delay for the other alternatives was computed in asimilar manner.

The annual number of crashes was estimated using aver-age crash rates for intersections with characteristics similarto those of the subject intersection. This analysis suggestedthat the Signal alternative would have the most crashes at 9 per year. The two-way-stop-control alternatives were esti-mated to have lower crash frequencies at 4 or 5 crashes peryear, depending on whether turn bays are present on themajor road. The initial and annual costs of each alternative

A-13

were estimated using costs developed for similar intersec-tions during the past year. The estimated impacts of the alter-natives are summarized in Table A-15.

Stage 3: Step 2. Select Best Alternative

As a first activity of this step, the factor weights were iden-tified. These weights would be used to aggregate the impactsof the various factors identified in Step 1. The weightsselected are listed in Column 4 of Table A-15. The weight usedfor the delay impact is based on an estimate of the annual costof time per vehicle (in thousands of dollars). It assumes thatthere are 300 equivalent average days per year and that thecost of time is $15 per vehicle-hour (i.e., 4.5 = 300 � 15 /1000). Thus, the product of this weight and total delay yieldsan estimate of the annual cost of delay per year for a givenalternative.

The weight used for crash frequency reflects an equivalentcost for the average crash (in thousands of dollars) and reflectsthe distribution of fatal, injury, and property damage crashesat intersections. The initial costs (in thousands of dollars)were given a weight of only 0.1 to reflect their annualizationover a 15-year period. These weights were used to computethe weighted total shown in the last row of the table. Thistotal represents the annual cost of each alternative (in thou-sands of dollars).

The total impact values listed in the last row of Table A-15 indicate that the TWSC-3 alternative (i.e., add a right-turn bay on the minor road and left- and right-turn bays on the major road) has the lowest overall impact. In con-trast, the Signal alternative has the largest impact due inpart to its large initial cost, its large number of expectedcrashes, and its large total delay. Based on this analysis,the TWSC-3 alternative was selected as the best alterna-tive. A sketch of the recommended alternative is shown inFigure A-6.

Stage 3: Step 3. Document Study

A report documenting the results of the engineering studywas prepared and submitted.

TABLE A-13 Case Study 1: Delay summary for peak traffic hour

Figure A-5. Case Study 1: Average intersection delay.

A-14

TABLE A-14 Case Study 1: Alternative effectiveness

TABLE A-15 Case Study 1: Alternative impacts

Figure A-6. Case Study 1: Recommended alternative (not to scale).

CASE STUDY 2

SYNOPSIS

This case study illustrates the application of the Engi-neering Study Guide (ESG) to an isolated, two-way stop-controlled intersection where minor-road drivers are expe-riencing excessive delay. The SIDRA software packagewas used to evaluate the operation of the intersection andits proposed alternatives. Conversion to traffic signal con-trol was recommended as the best method of reducingmotorist delay.

BACKGROUND

Recently, several housing developments have been builton Barracks Road near its intersection with U.S. 29. Thesedevelopments have increased traffic volumes in general andtruck volumes in particular (many of which are constructionrelated). A local elected official requested the installation ofa traffic signal at the intersection to reduce delay and improvesafety. Based on this request, an engineering study wasundertaken to determine whether improvements to the inter-section were needed.

The intersection of Barracks Road and U.S. 29 is locatedin a rural area with a population of less than 9,000 people.The intersection has four approach legs; two opposing legsare stop-controlled. Barracks Road is the minor road; it isoriented in an east-west direction and has stop control at the

A-15

intersection. U.S. 29 is the major road; it is oriented in anorth-south direction. The speed limit is posted at 55 mph(90 km/h) on U.S. 29 and 30 mph (50 km/h) on BarracksRoad. The geometric layout of the intersection is shown inFigure A-7.

STAGE 1: ALTERNATIVE IDENTIFICATIONAND SCREENING

The first stage of the engineering assessment processinvolved diagnosing the problems at the intersection anddetermining potential corrective measures. The alternativeidentification and screening stage consists of the followingsteps:

1. Define problem and cause,2. Select candidate alternatives, and3. Select viable alternatives.

Stage 1: Step 1. Define Problem and Cause

Gather Information

The first step was to gather information about the inter-section. This involved checking agency records and visitingthe site. The data gathered during this step are summarizedin the remainder of this section.

Figure A-7. Case Study 2: Existing intersection geometry (not to scale).

Historic Data. This intersection had not been the subjectof a prior engineering study, so archival traffic data were notavailable. Crash data were obtained from the AccidentRecords Division in the state’s Department of Transporta-tion. A review of these data indicated that six crashes hadoccurred at the intersection in the last year (three susceptibleto correction by a signal). The annual average daily traffic(AADT) is 12,000 vehicles per day on U.S. 29 and 7,000vehicles per day on Barracks Road.

Observational Study. The intersection was visited dur-ing a typical weekday to determine whether the reporteddelay problems were present at the intersection. During thisvisit, vehicles on the minor road were observed to experienceexcessive delay during the peak hours. Also, some delay andconflict was observed between the northbound through andright-turn movements. An On Site Observational Report wascompleted to document the nature of the operational andsafety problems at the intersection.

Site Survey. A site survey was also conducted during thesite visit. Key geometric features of the intersection notedduring this survey are summarized in Figure A-7.

Define Problem and Cause

Define Problem and Identify Cause. Based on the re-sults of the observational study, it was found that sufficientevidence existed to justify continuing with the engineer-ing study. As a next step in this study, Table 2-1 of the ESGwas used to identify possible causes for the observeddelays. The findings from this assessment are summarizedin Table A-16.

A-16

Define Influence Area. The subject intersection’s influ-ence area was defined to consist of the intersection conflictarea and a 100-m length of each intersection approach.

Stage 1: Step 2. Select Candidate Alternatives

Identify Potential Alternatives

Table 2-2 in the ESG was consulted to identify alterna-tives that could address the observed problems at the inter-section. The alternatives identified in Table 2-2 that areapplicable to the subject intersection are summarized inTable A-17.

Organize and Select Alternatives

The potential alternatives were subjected to a preliminaryscreening to eliminate any alternatives that would not be fea-sible at the site. Based on this screening, several alternativeswere eliminated. The alternatives eliminated are listed inTable A-18.

Based on this analysis, the following alternatives were foundto merit further study:

• Convert to traffic signal control,• Add right-turn bay to major road,• Add second lane on minor road, and• Convert to roundabout.

Stage 1: Step 3. Select Viable Alternatives

The candidate alternatives selected in Step 2 were moreclosely examined in this step. This examination focused on

TABLE A-16 Case Study 2: Potential problems and causes

TABLE A-17 Case Study 2: Potential alternatives

evaluating each alternative to assess its “viability” (i.e., itsability to address the observed problems effectively).

Gather Information

Identify Data. Table 2-5 in the ESG was used to identifythe data needed to evaluate the candidate alternatives. Basedon the information in this table, it was determined that thefollowing data were needed for guideline evaluation:

• Major- and minor-road turn movement volume (8 hours),• Heavy vehicle volume (8 hours),• Pedestrian volume (4 hours),• Gap frequency (4 hours),• 85th percentile approach speed,• Intersection sight distance,

✓ • Area population,✓ • Number of lanes, and✓ • Crash history by type.

Of these data, information about the last three items hadalready been gathered in a previous step. The minimum num-ber of hours for which data would be collected is indicated inthe list above.

Collect Data. At the onset of this task, it was noted thatvehicular volumes and approach speed were common to mostof the alternatives. It was also noted that the satisfaction of onesignal warrant would confirm the viability of the signal alter-native. From this assessment, it was decided initially to collectonly volume and speed data. These data would allow a checkof the Manual on Uniform Traffic Control Devices-MillenniumEdition (i.e., MUTCD 2000) (A-2) Signal Warrants 1 and 2. Ifthese warrants were not satisfied, then additional data wouldbe collected to allow a check of Warrants 4 and 8.

Warrant 3 was not evaluated because the intersection wasnot believed to represent an “unusual case.” Specifically, theintersection was not near an office complex, manufacturingplant, or other facility that generated high volumes over ashort time period.

Because it was unclear when the 8 peak traffic hoursoccurred, turn movement volumes were recorded for a totalof 12 hours. The counts were obtained on a Wednesday inApril; they were then adjusted to eliminate bias due to weekly

A-17

and monthly variations. The estimated average-day volumesare summarized in Table A-19. A review of the hourly totalsindicated that the morning peak hour occurred from 7 to 8 a.m.,the representative off-peak period was reasoned to occurfrom 10 to 11 a.m., and the afternoon peak hour occurredfrom 4 to 5 p.m.

No pedestrians were observed during the study period.Heavy vehicles comprised about 30 percent of the trafficstream on the minor road and 2 percent on the major road. Aspot speed study was conducted for U.S. 29; the 85th per-centile speed was found to be 91 km/h.

Assess and Select Alternatives

Convert to Traffic Signal Control. The traffic signalwarrants in the MUTCD 2000 (A-2) were evaluated to assessthe viability of the traffic signal alternative. Based on thisevaluation, the following warrants were satisfied at this inter-section:

• Warrant 1: Eight-hour vehicular volume (Condition Aonly) and

• Warrant 2: Four-hour volume.

For the warrant evaluation, the “70%” values were usedbecause the major-road speed exceeded 70 km/hr. ConditionA of Warrant 1 was satisfied for all 12 hours for which datawere collected. Warrant 2 was satisfied for 11 of the 12 hours.It should be noted that the right-turn volume on the east-bound approach was not included in the approach totalbecause right-turn vehicles were (1) provided a turn bay and(2) observed to enter the major road without conflict.

The presence of conditions that could misdirect the signalwarrant check was also assessed. A review of Table 2-15 inthe ESG revealed that two conditions exist that could affectthe warrant analysis conclusions. These conditions are

• Heavy vehicles on the minor road and• Left-turn bays on the major road.

The existence of problematic conditions required addi-tional analysis to determine whether the conclusions from thewarrant check were valid. The exclusion of the right-turn

TABLE A-18 Case Study 2: Alternatives eliminated

volume on the eastbound approach was reasoned to be con-sistent with MUTCD 2000 guidance (A-2, p. 4-C2) and, thus,it was not believed to represent a problematic condition.

Frequent heavy vehicles on the minor road were a point ofconcern because their representation is much larger than typ-ically found at most intersections. In fact, guidance providedin Chapter 2 of the ESG indicated that intersections withmore than 5 percent heavy vehicles are atypical and may notbe fully reflected in the warrants. Because the subject inter-section has 30 percent heavy vehicles, it was decided that theoperational effects of heavy vehicles would be explicitlyconsidered in the evaluation of the signal alternative.

Chapter 2 of the ESG indicated that a left-turn bay on themajor road may affect the warrant check when the approachhas only one through traffic lane. In this instance, the approachhad two through lanes, so the effect of the left-turn bay wasdetermined to be insignificant, as it relates to the accuracy ofthe warrant check conclusion.

In summary, the warrant analysis indicated that a trafficsignal may improve traffic operations at the intersection.However, the guideline addressing the effect of “heavy vehi-cle percentage” indicated that the large number of trucks atthe intersection may have an impact on the accuracy of thewarrant check conclusion. Therefore, the operation of thetraffic signal alternative would be examined in greater detailin the engineering study stage. Moreover, the examinationwould explicitly include the effect of trucks.

Add Right-Turn Bay to Major Road. Figure 2-6 in theESG was examined to determine whether an exclusive right-turn bay would be beneficial on the northbound intersectionapproach. The examination was based on a major-road speed

A-18

of 90 km/h (55 mph) and the afternoon peak hour volumes.From this examination, it was concluded that the right-turnbay would be beneficial.

Add Second Lane on Minor Road. Figure 2-4 in theESG was examined to determine whether adding an exclu-sive turn bay to the westbound approach would improve traf-fic operations at the intersection. The examination was basedon the afternoon peak hour volumes and a westbound right-turn volume that constituted 42 percent of the approach vol-ume. Based on this evaluation, it was concluded that an addi-tional lane on the westbound approach would be beneficial.Site conditions indicated that this lane could most easily beadded as a right-turn bay.

Convert to Roundabout. Table 2-12 in the ESG wasevaluated to determine whether a roundabout was a viablealternative for this intersection. The results of this evaluationare shown in Table A-20.

The response to Question 2 in Table A-20 required an esti-mate of the daily entering volume and the roundabout’s max-imum service volume. Daily entering volume was estimated as19,000 veh/d based on AADTs of 12,000 and 7,000 veh/d forthe major and minor roads, respectively. Maximum servicevolume was estimated as 41,500 veh/d using the equation inthe footnote to Table A-20. This estimate was based on 2-laneroundabout approaches, 63 percent of the volume entering onthe major road, and 34 percent left-turns at the intersection.The ratio of major-road-to-minor-road volume was computedas 1.7 (= 12/7).

As indicated by the last column of Table A-20, the round-about was found to be a viable alternative for all of the fac-

TABLE A-19 Case Study 2: Turn movement volumes

tors considered except “heavy vehicle frequency” (i.e.,Question 6). Thus, there was some concern about the abilityof the roundabout design to serve frequent truck traffic.There was also some concern about the suitability of theroundabout to a rural, high-speed location. Despite these con-cerns, the roundabout alternative was advanced to the engi-neering study stage.

Summary of Viable Alternatives

Based on the assessment of viable alternatives, all fourcandidate alternatives identified in Step 2 were selected formore detailed analysis in the next stage of the assessmentprocess. These alternatives are summarized below:

• TWSC: Base case (existing intersection with no im-provements);

• Signal: Conversion to traffic signal control;• Bay: Two-way stop control with right-turn bays on west-

bound and northbound approaches; and• Roundabout: Conversion to a two-lane roundabout.

The major- and minor-road right-turn bay alternatives werecombined into one alternative. This alternative is denoted in the list above as “Bay.” The designation shown in boldtype is used in subsequent sections to refer to the variousalternatives.

STAGE 2: ENGINEERING STUDY

The operational performance of the alternatives identifiedin the alternative identification and screening stage wereevaluated in the engineering study stage. This stage consistsof three steps:

1. Determine study type,2. Select analysis tool, and3. Conduct evaluation.

A-19

Stage 2: Step 1. Determine Study Type

The first step of the engineering study stage required adetermination of the type of study needed. Initially, the rela-tionship between the subject intersection and any upstreamor downstream signalized intersection was evaluated. Then,the type of evaluation needed was determined.

Determine Type of Operation

The interaction between the subject intersection and anyadjacent signalized intersection was evaluated to determinewhether the operation of the adjacent intersection should beconsidered in the analysis. An investigation of the locationof these intersections revealed that there are no intersectionswithin several kilometers of the subject intersection. Fig-ure 3-1 of the ESG indicates that distances exceeding 730 m(for a major-road peak hour volume of 1,223 veh/h) are suf-ficient to isolate the subject intersection from the adjacentsignalized intersections.

Determine Type of Evaluation

Next, the intersection was examined to determine whethera formal evaluation of alternatives was required. The ESGsuggests that an informal evaluation (i.e., implementation andfield study) is possible when there is only one viable alterna-tive applicable at an isolated intersection. However, the factthat the subject intersection is associated with several viablealternatives required the conduct of a formal evaluation.

Stage 2: Step 2. Select Analysis Tool

During this step, an analysis tool was selected for evaluat-ing the operation of the subject intersection and its alternatives.This selection was based on an identification of the desiredanalysis tool capabilities and a comparison of this list with the

TABLE A-20 Case Study 2: Roundabout evaluation worksheet

actual capabilities of the available tools. The tool selected foruse was the one that provided all of the desired capabilities.

Identify Desired Capabilities

Table 3-1 in the ESG was examined to determine whichanalysis tool capabilities would be required for this evalua-tion. Based on this examination, it was determined that theanalysis tool selected must be able to evaluate the traffic con-trol modes used with the alternatives (i.e., two-way stop con-trol, actuated traffic signal control, and roundabout opera-tion). It was also determined that the analysis tool must beable to evaluate the percentage of heavy vehicles, left- andright-turn bays, and shared-lane approaches. Finally, it wasdecided that the analysis tool should provide a “level of ser-vice” indication.

Evaluate and Select Analysis Tool

Three analysis tools were available for the evaluation:HCS (version 3.1c), CORSIM (version 4.3), and SIDRA(version 5.02a). Each analysis tool was evaluated for its abil-ity to provide the desired capabilities. CORSIM was elimi-nated because it cannot be used to evaluate roundabout oper-ations. The HCS was also eliminated because it can only beused to estimate roundabout capacity. SIDRA was deter-mined to be the only analysis tool that could be used to eval-uate all of the alternatives.

Stage 2: Step 3. Conduct Evaluation

The final step in the engineering study stage was to deter-mine the operational performance of the viable alternatives.Before conducting this evaluation, the need for additionalinformation and field data was reviewed. Then, the opera-tional performance of each alternative was quantified and itsrelative effectiveness was determined.

Gather Information

The need for additional data was assessed at this point in thestudy. This assessment focused on the input data requirementsof the chosen analysis tool (i.e., SIDRA). The initial data col-lection effort was determined to be thorough enough that allnecessary data were available. Based on this finding, it wasconcluded that additional data did not need to be gathered.

Evaluate Operational Performance

Design Alternatives. Some preliminary design decisionswere made regarding the operation and the geometry of thealternatives before they were evaluated. Most of this effort

A-20

was devoted to the development of a design and timing planfor the traffic signal control alternative. However, someeffort was also expended on the design of the roundabout.

Signal Alternative. The signal design process included adetermination of the intersection geometry and its signal tim-ing. Guidelines in Chapter 3 of the ESG indicated that anexclusive left-turn lane (or bay) may be needed when the left-turn volume exceeds 100 veh/h. However, it also notes thatthis criterion may not be appropriate when the opposing andadjacent through volumes are less than 450 veh/h. The eastand west approaches of the intersection satisfy this criterionfor the afternoon and morning peak hours, respectively.However, the through volumes were exceptionally light dur-ing these hours. Therefore, it was decided that left-turn bayson the eastbound and westbound approaches would not beincluded in the intersection design for the signal alternative.

Guidelines regarding the need for right-turn lanes are pro-vided in Figure 3-3 of the ESG. A check of these guidelinesindicated that right-turn lanes were not needed for the signalalternative.

Guidance regarding the operation of the signal controller isprovided in Table 3-5 of the ESG. This guidance indicated thatactuated operation is appropriate for isolated intersections.

Guidance regarding the need for left-turn phasing is pro-vided in Figure 3-5 of the ESG. A check of the informationin this figure indicated that left-turn phasing was not needed.Thus, the signal phase sequence should consist of two throughphases: one for the north and southbound approaches and onephase for the east and westbound approaches.

Traffic volumes recorded for the afternoon peak periodwere used to determine the controller settings. This analysisfollowed the guidelines associated with Tables 3-7 and 3-8in the ESG. The settings determined from this analysis werereasoned to be sufficiently accurate for the engineering studyevaluation; however, it was recognized that they may need tobe refined if the signal alternative is selected.

Bay Alternative - Major-Road Approach. Table 2-13 ofthe ESG provided some information regarding the appropri-ate length of a right-turn bay on the major road. According tothis table, a right-turn bay should be long enough to providefor the deceleration of right-turn vehicles. Figure 2-9 indi-cated that about 115 m was needed to provide adequate decel-eration distance when the 85th percentile speed is 91 km/h.Thus, the northbound major-road right-turn bay was designedto be 115 m in length.

Bay Alternative - Minor-Road Approach. According toTable 2-13 of the ESG, the right-turn bay length on the west-bound approach should be long enough to provide for right-turnvehicle storage. Figure 2-8 of the ESG was consulted for eachof the three analysis hours. The morning peak hour yielded thegreatest storage length of 15 m. Thus, the westbound minorroad right-turn bay was designed to be 15 m in length.

Roundabout Alternative. It was determined that the round-about could be designed to accommodate the large number ofheavy vehicles. The design guidelines provided in Round-abouts: An Informational Guide (A-4) were used for this pur-pose. This document indicated that a two-lane roundaboutwith an inscribed circle diameter of 60 m (and central islanddiameter of 50 m) could accommodate the travel paths ofturning trucks.

Evaluate Alternatives. SIDRA was used to evaluate eachalternative for the morning peak, off-peak, and afternoonpeak hours. The individual movement delays predicted forthe afternoon peak traffic hour (i.e., 4:00 to 5:00 p.m.) aresummarized in Table A-21. It should be noted that similarlevels of delay were observed during the morning peak hour.

The data in Table A-21 indicate that the two more restrictivealternatives (i.e., Signal and Roundabout) offer the least over-all delay. This low delay would be achieved at the “expense”of increased delay to the major-road through and right-turnmovements. However, this increase was very small when com-pared with the decrease in delay to minor-road movements.

The variation in intersection delay over time is shown inFigure A-8. The trends in this figure indicate that the exist-ing intersection (TWSC) resulted in the largest delay. In con-trast, the roundabout resulted in the smallest delay.

A-21

Determine Alternative Effectiveness

The findings from the operations analysis were evaluatedto determine whether each alternative would serve trafficeffectively. This determination was made based on the defi-nitions of “acceptable level of service” and “acceptable oper-ation” as defined in Chapter 3 of the ESG. Table A-22 sum-marizes the results of this evaluation.

Only the Signal and Roundabout alternatives were found tosatisfy all three of the “effectiveness” criteria. The existingintersection and the turn bay alternative did not yield minor-road delays that were low enough to be considered acceptable.

STAGE 3: ALTERNATIVE SELECTION

During the alternative selection stage, the alternativesadvanced from the engineering study stage are reviewed fortheir potential impact on traffic operations, safety, and theenvironment. The alternative selection stage consists of thefollowing three steps:

1. Identify impacts,2. Select best alternative, and3. Document study.

Stage 3: Step 1. Identify Impacts

Identify Decision Factors

Several factors were identified that might influence alter-native selection. These decision factors included traffic oper-ations, traffic safety, and direct cost. Impacts to the environ-ment and area aesthetics were considered but were largelyunaffected by the proposed alternatives.

Assess Degree of Impact

Each alternative’s impact on traffic operations, safety, andcost was evaluated. The traffic operations impacts werequantified using total delay during the average day. An esti-mate of annual crash frequency was used to quantify safety

TABLE A-21 Case Study 2: Delay summary for peak traffic hour

Figure A-8. Case Study 2: Average intersection delay.

impacts. Estimates of the initial construction and annual main-tenance costs were used to estimate the direct-cost impacts.The total delay for the average day was estimated from theevaluation conducted in the engineering study stage. It wasassumed that the morning peak hour, off-peak hours, and after-noon peak hour existed for 2, 18, and 2 hours, respectively,during the average day.

The crash frequency for the roundabout was difficult toestimate as there was little data available for reference. Arecent report by Robinson et al. (A-4) indicated that multilaneroundabouts have more conflicts than single-lane round-abouts and cite data indicating that crash frequency reductionmay be small for this location. Given the high speed on themajor-road approach, it was conservatively assumed thatthere would be no change in annual crash frequency with theroundabout. On the other hand, data from similar signalizedintersections indicated that it was likely that crash frequencywould be reduced (to about four crashes per year) at thisintersection, if it were signalized. Finally, the cost of con-structing the roundabout was estimated to be $400,000 dueto the large amount of right-of-way, design complexity, andnew pavement area needed. Table A-23 illustrates the find-ings from the impact assessment.

A-22

Stage 3: Step 2. Select Best Alternative

The best overall alternative was selected by choosing impactweights that reflect an economic worth for delay and crashes.These weights were established by the agency for previousprojects and found to yield reasonable results. The weightsselected for this study are listed in Column 4 of Table A-23.

Weighted impact totals for each alternative are shown inthe last row of Table A-23. These totals indicate that the sig-nal alternative is slightly more desirable than the multilaneroundabout alternative. The roundabout is likely to performbetter operationally. However, there was some concern aboutthe ability of the roundabout to improve safety at this inter-section. The anticipated cost to construct the roundabout isalso expected to be much larger than the signal due to theroundabout’s larger size and geometric design complexities.

Based on this analysis, the Signal alternative was selectedas the best overall alternative for this intersection.

Stage 3: Step 3. Document Study

A report documenting the results of the engineering studywas prepared and submitted.

TABLE A-22 Case Study 2: Alternative effectiveness

TABLE A-23 Case Study 2: Alternative impacts

CASE STUDY 3

SYNOPSIS

This case study illustrates the application of the Engineer-ing Study Guide (ESG) to an isolated, stop-controlled inter-section where minor-street drivers are experiencing exces-sive delay. These delays are partially created by frequenton-street parking maneuvers occurring on the minor street.CORSIM was used to evaluate the intersection. The removalof some on-street parking on the minor street and its replace-ment with a right-turn bay was recommended as the bestmethod for alleviating motorist delay.

BACKGROUND

The intersection of Market Street and Maple Avenue hasbeen the focus of ongoing complaint about excessive delayduring peak hours and unsafe parking maneuvers. Recentincreases in traffic demand (due to growth in the downtownarea) have amplified these problems and the correspondingnumber of complaints. The number of complaints received inrecent weeks has prompted the need for an engineering eval-uation of traffic conditions at the intersection.

The intersection is on the edge of the downtown area in acity of 40,000 people. The intersection has three approachlegs with stop control on Maple Avenue. Market Street is themajor street; it is oriented in an east-west direction. MapleAvenue is the minor street; it is oriented in the north-southdirection. Both streets have a posted speed limit of 35 mph.The westbound approach has an exclusive left-turn bay; allapproaches have one through lane and no right-turn bays.The layout of the intersection is shown in Figure A-9.

A-23

STAGE 1: ALTERNATIVE IDENTIFICATIONAND SCREENING

The first stage of the engineering assessment processinvolved diagnosing the problem at the intersection anddetermining potential corrective measures. Candidate alter-natives were identified and given a preliminary screening todetermine whether they represented viable solutions to theproblem. The alternative identification and screening stageconsists of the following steps:

1. Define problem and cause,2. Select candidate alternatives, and3. Select viable alternatives.

Stage 1: Step 1. Define Problem and Cause

Gather Information

The first step was to gather information about the inter-section. This involved checking agency records and visitingthe site. The data gathered during this step are summarizedin the remainder of this section.

Historic Data. The intersection had not been the subjectof a prior engineering study, so archival traffic data were not available. Crash data were obtained from the AccidentRecords Division of the Department of Public Works. Thesedata indicated that a total of four collisions occurred at theintersection in the last year (two susceptible to correction bya signal). Regional traffic counts recorded 1 year earlier indi-cated an annual average daily traffic (AADT) of 12,600 vehi-

Figure A-9. Case Study 3: Existing intersection geometry (not to scale).

cles per day on Market Street and 11,500 vehicles per day onMaple Avenue.

Observational Study. The intersection was visited duringa typical weekday to determine whether the problems reportedwere present at the intersection. The observational studyrevealed that the vehicles on the minor street experiencedexcessive delay during the peak hours. Most of the delay wasdue to the difficulty that northbound left-turn drivers had find-ing an adequate gap in traffic on Market Street. This difficultytranslated into significant delay to both left-turn and right-turndrivers because they shared the same approach lane. Someconflict and delay was also observed to result from frequentparallel parking maneuvers on the minor-street approach.

Site Survey. A site survey was also conducted during thesite visit. Key geometric features of the intersection notedduring this survey are summarized in Figure A-9. Also notedduring this visit were the characteristics of the on-street park-ing on Maple Avenue. Parallel parking is permitted alongboth sides of Maple Avenue. Each parking space is about 25 ft long and has a posted 30-minute time limit. Observationsindicate that each parking space is used by about two vehiclesper hour. The parking lane extends along Maple Avenue towithin 5 ft of the intersection with Market Street. Buildingsalong Maple Avenue are close to the curb, having a setbackof 10 ft from the edge of pavement. This narrow setback dis-tance may make it unrealistic to consider widening MapleAvenue to provide additional traffic lanes at the intersection.

Define Problem and Cause

Define Problem and Identify Cause. Based on the re-sults of the observational study, it was found that sufficient

A-24

evidence existed to justify continuing with the engineeringstudy. Using the information provided in Table 2-1 of theESG, possible causes for the observed problems were identi-fied. This information is summarized in Table A-24.

Stage 1: Step 2. Select Candidate Alternatives

Identify Potential Alternatives

Table 2-2 in the ESG was consulted to determine potentialalternatives that could address the observed problems at theintersection. The alternatives identified in Table 2-2 are sum-marized in Table A-25.

Organize and Select Alternatives

The potential alternatives were subjected to a preliminaryscreening to eliminate any alternatives that would not be fea-sible at the site. Judgment was used to identify those alter-natives that were clearly not feasible due to site-specificconstraints. Based on this screening, “peak-hour parking pro-hibition” was eliminated in preference to “permanent parkingprohibition” due to observed delays during both peak and off-peak hours. Permanent prohibition of on-street parking wouldonly be applicable to the intersection influence area to mini-mize the adverse effect of parking loss on local merchants. Anincrease in right-turn radius was eliminated in preference tothe “add second lane” alternative due to the extent of observedtraffic queues. The roundabout alternative was eliminated dueto right-of-way limitations. Finally, left-turn prohibition waseliminated because the city council had previously deter-mined that this alternative was not acceptable in the down-town area.

TABLE A-24 Case Study 3: Potential problems and causes

TABLE A-25 Case Study 3: Potential alternatives

Based on this analysis, the following alternatives werefound to merit further study:

• Add second lane on minor street (by removing on-streetparking near the intersection) and

• Convert to traffic signal control.

Stage 1: Step 3. Select Viable Alternatives

The candidate alternatives selected in Step 2 were moreclosely examined in this step. This examination focused onevaluating each alternative to assess its “viability” (i.e., itsability to address the observed problems effectively).

Gather Information

Identify Data. Table 2-5 in the ESG was used to identifythe data needed to evaluate the candidate alternatives. Basedon the information in this table, it was determined that thefollowing data were needed for guideline evaluation:

• Major- and minor-road turn movement volume (8 hours),• Heavy vehicle volume (8 hours),• Pedestrian volume (4 hours),• Gap frequency (4 hours),• 85th percentile approach speed,

✓ • Area population,✓ • Number of lanes, and✓ • Crash history by type.

Of these data, information about the last three items hadalready been gathered in a previous step. The minimum num-ber of hours for which data would be collected is indicated inthe list above.

Collect Data. At the onset of this task, it was noted thatvehicular volumes and approach speed were common tomost of the alternatives. It was also noted that the satisfac-tion of one signal warrant would confirm the viability of thesignal alternative. From this assessment, it was decided ini-tially to collect only volume and speed data. These datawould allow a check of the Manual on Uniform Traffic Con-trol Devices-Millennium Edition (i.e., MUTCD 2000) (A-2)

A-25

Signal Warrants 1 and 2. If these warrants were not satisfied,then Warrants 4 and 8 would be checked.

Warrant 3 was not evaluated because the intersection wasnot believed to represent an “unusual case.” Specifically, theintersection was not near an office complex, manufacturingplant, or other facility that generated high volumes over ashort period.

Traffic volumes were measured during 8 hours on a Mon-day in October. These counts were adjusted for weekly andmonthly variations to yield an estimate of the average-dayvolume. The morning peak hour, off-peak hour, and after-noon peak hour were found to occur from 7:00 to 8:00 a.m.,1:00 to 2:00 p.m., and 4:30 to 5:30 p.m., respectively. Theturn movement volumes for these hours are provided inTable A-26.

During the study, heavy vehicles were found to compriseabout 2 percent of all traffic entering the intersection. Pedes-trian volumes were constant throughout the day with about10 ped/hr crossing Maple Avenue and 15 ped/hr crossingMarket Street (total of all crossings on both corners). A spotspeed study was conducted for Market Street; the 85th per-centile speed was found to be 34 mph.

Assess and Select Alternatives

Convert to Traffic Signal Control. The traffic signalwarrants in the MUTCD 2000 (A-2) were evaluated toassess the viability of the traffic signal alternative. Basedon this evaluation, it was determined that Warrant 2: Four-Hour Vehicular Volume was satisfied. It should be notedthat the left-turn bay on the westbound approach wasincluded in the count of major-street lanes (i.e., 2 lanes) forthe following reasons: (1) the approach had only one lane,and (2) the left-turn volume was about one-half of the totalapproach volume.

To complete the warrant check, the presence of conditionsthat could misdirect the signal warrant check was assessed.A review of Table 2-15 in the ESG revealed that two condi-tions exist that could affect the warrant analysis conclusions.One condition relates to the fact that the intersection has onlythree approach legs. The second condition relates to the sig-nificant number of right-turn vehicles on the minor-streetapproach.

TABLE A-26 Case Study 3: Turn movement volumes

The discussion associated with Table 2-15 indicated thatthe MUTCD’s volume-based warrants might have thresholdvalues that are too low for a three-leg intersection. Based onthe guidance in Chapter 2 of the ESG, the major and minorroads at a three-leg intersection may need to have volumes of1,000 and 200 veh/h, respectively, for the 8 highest hoursbefore a traffic signal is likely to be justified. A check of thevolumes at the subject intersection indicated that this“1,000/200 veh/h” criterion was not met.

The impact of minor-street right-turn volume on the accu-racy of the warrant check was also investigated. Figure 2-11in the ESG was consulted to determine the extent to whichthe right-turn volume might be reduced in the warrant check.Based on this examination, it was found that the minor-streetright-turn volume probably should not be included in thewarrant analysis (except during the afternoon peak hour). Infact, when Warrant 2 was re-evaluated without this right-turnvolume, it was not satisfied.

In summary, the warrant analysis indicated that a trafficsignal might improve traffic operations at the intersection.However, guidelines in Chapter 2 of the ESG indicated thatthe three-leg geometry and ample right-turn capacity of theintersection might have an impact on the accuracy of the war-rant check conclusions. It was determined that the operationof the traffic signal alternative would be carefully examinedduring the engineering study stage to confirm the benefits ofsignalization relative to the other alternative.

Add Second Lane on Minor Street. Figure 2-4 in theESG was examined to determine whether adding an exclu-sive right-turn bay to Maple Avenue would improve trafficoperations at the intersection. This bay would be created byremoving some on-street parking on the east side of MapleAvenue. The evaluation was based on a right-turn volumethat represented 73 percent of the approach traffic streamduring the afternoon peak hour. Based on this examination,it was found that a right-turn bay could improve traffic oper-ations during the peak traffic demand hours.

Summary of Viable Alternatives

Based on the assessment of viable alternatives, the candi-date alternatives identified in Step 2 were selected for moredetailed analysis in the next stage of the assessment process.These alternatives are summarized below:

• TWSC: Base case (existing intersection with no im-provements);

• Signal: Conversion to traffic signal control; and• Bay: Stop control with right-turn bay on the minor street

(i.e., Maple Avenue).

The designation shown in bold type is used in subsequentsections to refer to the various alternatives.

A-26

STAGE 2: ENGINEERING STUDY

The operational performance of the alternatives identifiedin the alternative identification and screening stage wereevaluated in the engineering study stage. This stage consistsof three steps:

1. Determine study type,2. Select analysis tool, and3. Conduct evaluation.

Stage 2: Step 1. Determine Study Type

The first step of the engineering study stage required adetermination of the type of study needed. Initially, the rela-tionship between the subject intersection and any upstreamor downstream signalized intersection was evaluated. Then,the analysis detail required for the assessment of each alter-native was determined.

Determine Type of Operation

Figure 3-1 of the ESG was consulted to determine whetherthe subject intersection was isolated from the operation of theupstream intersections. This figure can be used to determineif an intersection is isolated based on consideration of the dis-tance to the nearest signalized intersection and the two-waytraffic volume. The nearest signalized intersection in eachdirection of travel is 1⁄2 mile from the subject intersection.The afternoon peak-hour two-way volumes on MapleAvenue, Market Street (west leg), and Market Street (eastleg) are 951, 844, and 1,149 veh/h, respectively. Based onthis information and the guidelines in Figure 3-1, it was con-cluded that the subject intersection was effectively isolatedfrom nearby signalized intersection operations.

Determine Type of Evaluation

Next, the intersection was examined to determine whethera formal evaluation of alternatives was required. The ESGsuggests that an informal evaluation (i.e., implementationand field study) is possible when an intersection has only oneviable alternative. However, the fact that the subject inter-section is associated with several viable alternatives requiredthe conduct of a formal evaluation.

Stage 2: Step 2. Select Analysis Tool

During this step, an analysis tool was selected for evaluat-ing the operation of the subject intersection and its alterna-tives. This selection was based on an identification of thedesired analysis tool capabilities and a comparison of this listwith the actual capabilities of the available tools. The tool

selected for use was the one that provided all of the desiredcapabilities.

Identify Desired Capabilities

Table 3-1 in the ESG was examined to determine whichanalysis tool capabilities would be required for this evalua-tion. Based on this examination, it was determined that theanalysis tool selected must be able to evaluate all of the traf-fic control modes used with the alternatives (i.e., two-waystop control and actuated traffic signal control). It was alsodetermined that the analysis tool must be able to evaluateleft- and right-turn bays, shared-lane approaches, and on-street parking activity. Finally, it was decided that the analy-sis tool should provide a level-of-service indication.

Evaluate and Select Analysis Tool

Three analysis tools were available for the evaluation:HCS (version 3.1c), CORSIM (version 4.3), and SIDRA(version 5.02a). Each tool was evaluated for its ability toprovide the desired capabilities. Based on this evaluation,CORSIM was determined to be the only analysis tool thatcould explicitly model parking maneuver frequency andduration on a stall-by-stall basis.

It was noted that the version of CORSIM used did not pro-vide an estimate of control delay (as would be needed to deter-mine level of service). However, previous efforts by city engi-neers to validate this tool had revealed that CORSIM’s “queuedelay” estimate was within 5 percent of the control delay(based on field measurements). Therefore, the “queue delay”obtained from CORSIM was used for the evaluation with therecognition that any level-of-service estimate would beapproximate.

Stage 2: Step 3. Conduct Evaluation

The final step in the engineering study stage was to deter-mine the operational performance of the viable alternatives.Before conducting this evaluation, the need for additionalinformation and field data was reviewed. Then, the opera-tional performance of each alternative was quantified and itsrelative effectiveness was determined.

Gather Information

The need for additional data was assessed at this point in thestudy. This assessment focused on the input data requirementsof the chosen analysis tool (i.e., CORSIM). The initial datacollection effort was determined to be thorough enough that allnecessary data were available. Based on this finding, it wasconcluded that additional data did not need to be gathered.

A-27

Evaluate Operational Performance

Design Alternatives. Some preliminary design decisionswere made regarding the operation and geometry of the alter-natives before they were evaluated. Most of this effort wasdevoted to the development of a design and timing plan forthe traffic signal control alternative. However, some effortwas also expended on the design of the right-turn bay.

Signal Alternative. The signal design process included adetermination of the intersection geometry and its signal tim-ing. Guidelines in Chapter 3 of the ESG were used for thispurpose. In particular, guidelines regarding the need forright-turn lanes are provided in Figure 3-3. A check of theseguidelines indicated that a northbound right-turn lane wasnot needed for the signal alternative.

Guidance regarding the operation of the signal controller isprovided in Table 3-5 of the ESG. This guidance indicated thatactuated operation is appropriate for isolated intersections.

Guidance regarding the need for left-turn phasing is pro-vided in Figure 3-5 of the ESG. A check of the informationin this figure indicated that a protected-permitted left-turnphase was needed for the westbound left-turn movement dur-ing the afternoon peak hour.

Traffic volumes recorded for the afternoon peak hour wereused to determine the controller settings. This analysis fol-lowed the guidelines associated with Tables 3-7 and 3-8 inthe ESG. The settings determined from this analysis werereasoned to be sufficiently accurate for the engineering studyevaluation; however, it was recognized that they may need tobe refined if the signal alternative is selected.

Bay Alternative. Table 2-13 of the ESG provided someinformation regarding the appropriate length of a right-turnbay on the minor street. According to this table, a right-turnbay should be long enough to provide for right-turn vehiclestorage. Table A-26 indicated that the conflicting volumeduring the afternoon peak hour is 407 veh/h (=308 + 197/2).Extrapolation of Figure 2-8 of the ESG for a right-turn vol-ume of 340 veh/h indicated that the right-turn bay on theminor street should be at least 260 ft long during the peakhour. However, due to the desires of local merchants to main-tain as much on-street parking as possible and the fact that a50-ft bay would be adequate for all other hours, it was deter-mined that a reasonable maximum bay length was 100 ft. Theuse of this length required the permanent removal of fourparking spaces.

Evaluate Alternatives. Because a stochastic simulationmodel (i.e., CORSIM) was being used, the minimum simu-lation period had to be determined so that statistically reli-able results could be obtained. This period was establishedby considering the minimum simulation period based on the individual movement delays and the average intersec-tion delay. Guidelines in Chapter 3 of the ESG were used to

determine the minimum simulation period based on the vol-ume of the individual movements and the total volume enter-ing the intersection.

The minimum simulation period based on individualmovement delay was determined first. Traffic volumesrecorded for the morning peak, afternoon peak, and off-peakhours were used for this determination. Movements thatshared a traffic lane were combined, as suggested in the ESG.Figure 3-12 of the ESG was used to determine the minimumsimulation period needed for each movement volume. Thesmallest movement volume for a given hour dictated theminimum simulation period based on individual movementdelay. The results of this evaluation are shown in Column 7of Table A-27.

Next, the minimum simulation period based on averageintersection delay was determined. The total entering volumesare 969 vehicles for the morning peak, 1,075 vehicles for theoff-peak, and 1,472 vehicles for the afternoon peak hour. Fig-ure 3-13 of the ESG was used to determine the minimum sim-ulation period needed for the overall intersection. The resultsof this evaluation are shown in Column 9 of Table A-27.

The last step in this process was to compare the minimumsimulation period based on individual movements with thatbased on the overall intersection. The larger of these two val-ues was selected as the minimum simulation period used forthe evaluation. It was recognized that a simulation of thisduration would provide the statistical precision needed in alldelay estimates. The simulation period used for the evalua-tion is shown in Column 10 of Table A-27. For each timeperiod, the minimum simulation period was found to be 0.8

A-28

or 0.9 h. For simplicity, the minimum value was rounded upto an even 1 hour for all time periods.

CORSIM was used to evaluate each viable alternative forthe morning peak, afternoon peak, and off-peak hours. Theindividual movement delays predicted for the afternoonpeak traffic hour are summarized in Table A-28. The Bayalternative (i.e., two-way stop control with right-turn bay onMaple Avenue) resulted in the smallest intersection delay.The existing alternative (TWSC) produced unacceptabledelay (i.e., it exceeded the delay threshold associated withLOS D and had more than 4.0 veh-hr of delay) on the minorstreet. In contrast, the Signal alternative resulted in thelargest intersection delay.

The impact of each alternative on average intersectiondelay was also examined. Figure A-10 illustrates the averageintersection delay for each alternative. The trends are con-sistent with those found in Table A-28 and indicate that theBay alternative results in the smallest delay for all threeanalysis periods.

The delay associated with each alternative was tested todetermine whether it was significantly different from thatof the other alternatives. The comparison technique wasbased on an examination of the “delay ratio” (i.e., smallerdelay divided by larger delay), as described in Chapter 3 ofthe ESG. Ratios less than 0.9 indicate that the two delays(and corresponding alternatives) are significantly different.Table A-29 summarizes the comparison of each pair ofalternatives.

The delay ratio for each pair of alternatives was found tobe less than 0.9. This finding indicated that the delays asso-

TABLE A-27 Case Study 3: Minimum simulation period evaluation

TABLE A-28 Case Study 3: Delay summary for peak traffic hour

ciated with each pair are significantly different (with 90 per-cent confidence) and that the trends shown in Figure A-10 arevery likely to be accurate. Most important, the analysis con-firms that the Bay alternative is the best overall alternative interms of traffic operations.

Determine Alternative Effectiveness

The findings from the operations analysis were evaluatedto determine whether each alternative served traffic effec-tively. This determination was made based on the definitionsof “acceptable level of service” and “acceptable operation”defined in Chapter 3 of the ESG. Table A-30 summarizes theeffectiveness of each alternative.

As suggested by the response provided in the last columnof Table A-30, only one alternative was found to satisfy the“effectiveness” criteria. Specifically, the Bay alternative (i.e.,the addition of a right-turn bay on the minor street by remov-ing 100 ft of parking) was the only alternative that satisfiedall three criteria. It reduced the average intersection delay to

A-29

acceptable levels and provided minor-street delays of 4 veh-hor less during the peak hour.

STAGE 3: ALTERNATIVE SELECTION

During the alternative selection stage, the alternativesadvanced from the engineering study stage are reviewed fortheir potential impact on traffic operations, safety, and theenvironment. The alternative selection stage consists of thefollowing three steps:

1. Identify impacts,2. Select best alternative, and3. Document study.

Stage 3: Step 1. Identify Impacts

In general, factors are typically identified that might influ-ence alternative selection. These decision factors includetraffic operations, traffic safety, direct cost, environment, andaesthetics. In this case, an alternative’s impact on local busi-ness was also relevant.

An evaluation of the impact of the Bay alternative (i.e., aright-turn bay on the minor-street approach) indicated thattraffic operations and safety at the intersection would begreatly improved for a nominal direct cost (i.e., the cost ofdelineating a right-turn bay and removing parking signs).However, the indirect cost to local businesses through theremoval of four parking slots was difficult to quantify. It wasnoted that convenient customer parking would still be avail-able within 150 ft of the affected businesses. It was alsothought that the improved operation of the intersection maybring some customers back to area businesses. In summary,the cost of parking removal was reasoned to be acceptable,given the benefits received directly by motorists and indi-rectly by the area businesses in terms of increased businessactivity.

Figure A-10. Case Study 3: Average intersection delay.

TABLE A-29 Case Study 3: Analysis of delay reduction associated with selected alternatives

Stage 3: Step 2. Select Best Alternative

Based on this analysis, the Bay alternative (i.e., a right-turn bay on the minor-street approach) was selected as thebest overall alternative.

A-30

Stage 3: Step 3. Document Study

A report documenting the results of the engineering studywas prepared and submitted.

TABLE A-30 Case Study 3: Alternative effectiveness

CASE STUDY 4

SYNOPSIS

This case study illustrates the application of the Engi-neering Study Guide (ESG) to an intersection being plannedfor construction on a busy arterial street. The location of theproposed intersection is very near to a signalized intersec-tion. The potential for queue spillback from the signalizedintersection into the subject intersection is high and wouldcreate excessive delay if it occurred. The CORSIM simula-tion model was used to analyze the operation of the pro-posed intersection and its two nearby signalized intersec-tions. The objective was to resolve queuing problems andminimize the impact of the new intersection on arterialoperations.

BACKGROUND

Main Street is a busy arterial street that provides for a sig-nificant amount of east-to-west intracity travel. A developeris proposing to build an office park on the south side of MainStreet between Rio Road and Sunset Boulevard. The devel-oper is also proposing to fund the extension of a city street(i.e., Elm Street) through the property to provide a point ofaccess to the arterial street system. The proposed intersectionwould have three legs with Elm Street representing the minorstreet. It would be located about 400 ft west of Sunset Boule-vard and 800 ft east of Rio Road.

The City Traffic Engineer is concerned that projectedElm Street traffic demands (at the intersection with MainStreet) would exceed the capacity of a two-way stop-controlledintersection. There is also a concern that the projected de-mands may justify a traffic signal and that this signal maymake it difficult to maintain good traffic progression alongMain Street. Finally, there is a concern that the intersection’sproposed location is too close to the signalized intersectionat Sunset Boulevard. Queues from this intersection may spillback into and block traffic at the new intersection. Becauseof these concerns, an engineering study was undertaken toevaluate the feasibility of adding a new intersection to MainStreet.

Main Street is major arterial in a city of 100,000 people.The intersections of Main Street with Rio Road and SunsetBoulevard are signalized and are coordinated with one an-other. The timing plan for these two intersections was devel-oped about 3 years ago. Traffic progression could be signif-icantly improved if the timing plan were updated. All streetshave a posted speed limit of 35 mph. Figure A-11 illustratesthe proposed location of the Elm Street intersection. It alsoshows the two signalized intersections that are adjacent toElm Street.

A-31

STAGE 1: ALTERNATIVE IDENTIFICATIONAND SCREENING

The first stage of the engineering assessment processinvolved diagnosing the problem at the intersection and deter-mining potential corrective measures. Candidate alternativeswere identified and given a preliminary screening to deter-mine whether they represented viable solutions to the prob-lem. The alternative identification and screening stage con-sists of the following steps:

1. Define problem and cause,2. Select candidate alternatives, and3. Select viable alternatives.

Stage 1: Step 1. Define Problem and Cause

Gather Information

The first step was to gather information about the arterial.Information about the proposed intersection was also col-lected; however, much of the information available was basedon traffic projections and the proposed site layout. Informa-tion was also gathered from agency records and a visit to thesite. The data gathered during this step are summarized in theremainder of this section.

Historic Data. The last engineering study of the arterial wasa corridor study conducted about 10 years ago. This study eval-uated the arterial’s ability to serve intracity travel and made rec-ommendations about improvements needed to improve arterialcapacity. Regional traffic counts recorded within the past yearindicated that the annual average daily traffic (AADT) on MainStreet is 29,400 vehicles per day.

Observational Study. The arterial segment was visitedduring a typical weekday to determine the character of traf-fic flow, the extent of the traffic queues, and the geometry ofthe intersections. The observational study revealed that traf-fic queues from the intersection of Main Street and SunsetBoulevard extend 450 ft back (toward Rio Road) during theafternoon peak traffic hour. Observations indicate that thisqueue length could be reduced by improved signal timing atSunset Boulevard.

Site Survey. A site survey was also conducted during thesite visit. Key geometric features of the arterial noted duringthis survey are summarized in Figure A-11.

Traffic Projections. The developer was asked to define thecharacteristics of the proposed business park (e.g., number

A-32

Figure A-11. Case Study 4: Existing arterial geometry (not to scale).

of employees, types of business, and square footage). Thesedata were then used to develop an estimate of the weekdaytraffic demand on Elm Street. Based on this analysis, theAADT for Elm Street was estimated as 9,200 vehicles per day.

Define Problem and Cause

Assess Evidence. The City Traffic Engineer’s concernsappear to have merit based on the findings of the observa-tional study. Possible problems that might occur from theconstruction of the proposed intersection would include (1) excessive delay on the Elm Street approach and (2) dis-ruption to traffic progression if the Elm Street intersectionwere signalized. Also, two operational problems were foundon Main Street. One problem was excessive delay andqueuing on the eastbound approach to the intersection ofMain Street and Sunset Boulevard A second problem waspoor traffic progression along Main Street. Based on theresults of the observational study, it was found that suffi-cient evidence existed to justify continuing with the engi-neering study.

Define Problem and Identify Cause. Using the infor-mation provided in Table 2-1 of the ESG, possible causes forthe aforementioned problems were identified. This informa-tion is summarized in Table A-31.

A-33

Stage 1: Step 2. Select Candidate Alternatives

Identify Potential Alternatives

Tables 2-2 and 2-3 in the ESG were consulted to determinepotential alternatives that could address the observed problemsat the intersection. The alternatives identified in these tablesare summarized in Table A-32. Two-way stop control wasadded to the list of potential alternatives to represent the “base”control mode because it represented the least-restrictive modethat could be used at an arterial-collector intersection.

Organize and Select Alternatives

The potential alternatives were subjected to a preliminaryscreening to eliminate any alternatives that would not be fea-sible at the site. Judgment was used to identify those alterna-tives that were clearly not feasible due to site-specific con-straints. Table A-33 lists the alternatives eliminated andprovides the reason for this action.

Based on this analysis, the following alternatives werefound to merit further study:

• Use two-way stop control (at Elm Street);• Convert to traffic signal controlled coordinated system

(at Elm Street);• Adjust signal timing at downstream signal;

TABLE A-31 Case Study 4: Potential problems and causes

TABLE A-32 Case Study 4: Potential alternatives

• Modify signal coordination (for all signalized intersec-tions on Main Street); and

• Add traffic lanes at downstream signal (on Main Streetat Sunset Boulevard).

Stage 1: Step 3. Select Viable Alternatives

The candidate alternatives selected in Step 2 were moreclosely examined in this step. This examination focused onevaluating each alternative to assess its “viability” (i.e., itsability to address the observed problems effectively).

Guidelines in Chapter 2 of the ESG were used to assessthe viability of traffic signal control at Elm Street. Similarguidelines were not available for assessing the viability ofsignal timing adjustment, signal coordination, and addingtraffic lanes to downstream signalized intersections. Thus,these alternatives were advanced to the engineering studystage for a formal evaluation.

Gather Information

Identify Data. Table 2-5 in the ESG was used to identifythe data needed to evaluate the candidate alternatives. Basedon the information in this table, it was determined that thefollowing data were needed for guideline evaluation:

• Major- and minor-road approach volume for proposedintersection (8 hours),

• Heavy vehicle volume for proposed intersection (8 hours),• 85th percentile approach speed (on Main Street),• Progression quality (on Main Street),

✓ • Area population, and✓ • Number of lanes for proposed intersection.

These data are sufficient to evaluate the Manual on UniformTraffic Control Devices-Millennium Edition (i.e., MUTCD2000) (A-2) Signal Warrants 1, 2, 3, and 6. Information aboutthe last two items had already been gathered in a previousstep. The minimum number of hours for which the data werecollected is indicated in the list above.

A-34

Collect Data. At the onset of this task, it was noted thatthe satisfaction of one signal warrant would confirm the via-bility of the corresponding alternative. From this assessment,it was decided to base the warrant check on volume andspeed data. These data would allow a check of Warrants 1, 2,and 3. If none of these warrants were satisfied, then Warrant6 would be evaluated.

To facilitate the warrant check for the proposed inter-section, an 8-hour count was synthesized using a combina-tion of forecast and existing traffic data. The objective of thisapproach was to derive a reasonably accurate estimate of thecombined traffic volumes for each of the 8 highest hours ofthe average day. The forecast AADT for Elm Street (includ-ing development-related and external through traffic) wasconverted to hourly traffic volume estimates using the proce-dures described in Appendix C. Main Street traffic volumeswere obtained from an 8-hour manual count.

A spot speed study was conducted for Main Street; the 85th percentile speed was found to be 37 mph. Heavy vehicleswere noted to comprise about 2 percent of the traffic stream.

Assess and Select Alternatives

The traffic signal warrants in the MUTCD 2000 (A-2) wereevaluated to assess the viability of the traffic signal alternative.The results of this evaluation are summarized in Table A-34.Based on this evaluation, it was determined that Warrants 1, 2,and 3 were satisfied. It should be noted that the Elm Streetapproach was assumed to have a minimum of two traffic lanesbased on the magnitude of the approach volume.

To complete the warrant check, there was an assessment ofpossible conditions that could misdirect the signal warrantcheck. A review of Table 2-15 in the ESG revealed that pro-gressive traffic flow on the major street could affect the warrantanalysis conclusions. Discussion associated with Table 2-15indicated that dense traffic platoons can affect operations at theproposed intersection to the extent that conclusions reachedfrom the warrant analysis may be inaccurate. Based on thisanalysis, it was concluded that the operation of the trafficsignal alternative would be carefully examined during theengineering study stage to confirm the benefits of signalization.

TABLE A-33 Case Study 4: Alternatives eliminated

Summary of Viable Alternatives

The candidate alternatives selected in Step 2 were deter-mined to be viable alternatives and suitable for formal eval-uation in the engineering study stage. However, these alter-natives were reconfigured to form alternatives that morecomprehensively addressed the observed operational prob-lems. The revised alternatives follow:

1. Base: Two-way stop control at Elm Street intersectionwith existing signal timings;

2. Bay: Two-way stop control at Elm Street intersection,adjust timing and coordinate existing signals at RioRoad and at Sunset Boulevard, and add a right-turn bayon eastbound Main Street at its intersection with Sun-set Boulevard; and

3. Signal: Convert to traffic signal control at Elm Streetintersection, adjust timing and coordinate all signalizedintersections, and add a right-turn bay on eastboundMain Street at its intersection with Sunset Boulevard

The designation shown in bold type is used in subsequentsections to refer to the various alternatives.

STAGE 2: ENGINEERING STUDY

The operational performance of the alternatives identifiedin the alternative identification and screening stage wereevaluated in the engineering study stage. This stage consistsof three steps:

1. Determine study type,2. Select analysis tool, and3. Conduct evaluation.

Stage 2: Step 1. Determine Study Type

The first step of the engineering study stage required adetermination of the type of study needed. Initially, the rela-

A-35

tionship between the subject intersection and any upstreamor downstream signalized intersection was evaluated. Then,the analysis detail required for the assessment of each alter-native was determined.

Determine Type of Operation

Figure 3-1 in the ESG was consulted to determine whetherthe proposed intersection would be isolated from the opera-tion of upstream intersections. This figure can be used todetermine if an intersection is isolated based on considera-tion of the distance to the nearest signalized intersection andthe two-way traffic volume. The nearest intersection to theeast (i.e., Sunset Blvd.) is 400 ft from the proposed inter-section. The nearest intersection to the west (i.e., Rio Road)is 800 ft from the proposed intersection. The peak hour two-way volume on both street segments is about 2,700 veh/h.Based on this information and the guidelines in Figure 3-1, itwas concluded that the intersection and all its alternativeswould be treated as “non-isolated.”

Determine Type of Evaluation

Next, the intersection was examined to determine whethera formal evaluation of alternatives was required. The ESGsuggests that an informal evaluation (i.e., implementationand field study) is possible when an intersection is isolatedand has only one viable alternative. However, a formal eval-uation was required in this case, because the subject inter-section was classified as “non-isolated.”

Stage 2: Step 2. Select Analysis Tool

During this step, an analysis tool was selected for evaluat-ing the operation of the subject intersection and its alternatives.This selection was based on an identification of the desiredanalysis tool capabilities and a comparison of this list with theactual capabilities of the available tools. The tool selected foruse was the one that provided all of the desired capabilities.

TABLE A-34 Case Study 4: Signal warrant check summary

Identify Desired Capabilities

Table 3-1 in the ESG was examined to determine whichanalysis tool capabilities would be required for this evaluation.Based on this examination, it was determined that the analysistool selected must be able to evaluate all of the traffic controlmodes used with the alternatives (i.e., two-way stop controland signal coordination). It was also determined that the analy-sis tool must be able to evaluate left- and right-turn bays, right-turn-on-red at signalized intersections, and spillback from adownstream signal. Finally, it was decided that the analysistool should provide a level-of-service (LOS) indication.

Evaluate and Select Analysis Tool

Three analysis tools were available for the evaluation: HCS(version 3.1c), CORSIM (version 4.3), and TRANSYT-7F(release 8). Each tool was evaluated for its ability to provide thedesired capabilities. Based on this evaluation, CORSIM wasdetermined to be the only analysis tool that could fully evalu-ate the effect of queue spillback on intersection operations.

It was noted that the version of CORSIM used did not pro-vide an estimate of control delay (as would be needed todetermine LOS). However, previous efforts by city engineersto validate this tool had revealed that CORSIM’s “queuedelay” estimate was within 5 percent of the control delay(based on field measurements). Therefore, the “queue delay”obtained from CORSIM was used for the evaluation with therecognition that any LOS estimate would be approximate.

Stage 2: Step 3. Conduct Evaluation

The final step in the engineering study stage was to deter-mine the operational performance of the viable alternatives.Before conducting this evaluation, the need for additional

A-36

information and field data was reviewed. Then, the opera-tional performance of each alternative was quantified and itsrelative effectiveness was determined.

Gather Information

The need for additional data was assessed at this point inthe study. This assessment focused on the input data require-ments of the chosen analysis tool (i.e., CORSIM). The fol-lowing data were identified as necessary for the evaluation:

• Major- and minor-street turn movement volumes for allintersections;

• Signal phase sequence and interval duration for bothsignalized intersections; and

• Relative offset between signals.

It was determined that the evaluation would include themorning and afternoon peak traffic hours as well as one mid-day off-peak hour. Turn movement counts were recorded ona Tuesday in July at the two signalized intersections. Thesecounts were adjusted for weekly and monthly variations(using the procedure described in Appendix C) to obtain anestimate of the average-day volume. Turn movement vol-umes for the proposed intersection were derived using typi-cal turn movement percentages for similar arterial/collectorintersections in the vicinity. The turn movement volumesused in the evaluation are summarized in Table A-35. Signalphase sequence, interval duration, and relative offset wereobtained from agency signal timing records.

Evaluate Operational Performance

Design Alternatives. Some preliminary design decisionswere made regarding the operation and the geometry of the

TABLE A-35 Case Study 4: Turn movement volumes

alternatives before they were evaluated. Most of this effortwas devoted to the development of an intersection design andtiming plan for the traffic signal control alternative. How-ever, some effort was also expended in the design of theright-turn bay at Sunset Boulevard.

Signal Alternative. The signal design process included adetermination of the intersection geometry and its signal tim-ing. Three through lanes were assumed for the major-streetapproaches to be consistent with the cross section of MainStreet.

Guidelines in Chapter 3 of the ESG indicated that an exclu-sive left-turn lane (or bay) may be needed when the left-turnvolume exceeds 100 veh/h. The left-turn volume on the west-bound approach satisfies this criterion for the peak and off-peak hours. Thus, a left-turn bay was included in the inter-section design for the Signal alternative.

Guidelines regarding the need for right-turn lanes are pro-vided in Figure 3-3 of the ESG. These guidelines indicatedthat a right-turn lane was not needed on the eastboundapproach. They also indicated that a right-turn lane was notneeded on the northbound approach. However, it wasdecided that two lanes would be provided on the north-bound approach to minimize the proposed intersection’simpact on arterial progression. These two lanes would includea bay for the left-turn movement and a lane for the right-turnmovement.

Guidance regarding the operation of the signal controlleris provided in Table 3-5 of the ESG. This guidance indicatedthat pretimed operation is appropriate for intersections hav-ing “heavy and consistent” traffic demands. As these char-acteristics were consistent with traffic demands at the pro-posed intersection, pretimed operation was assumed for thesignal design.

Guidance regarding the need for left-turn phasing is pro-vided in Figure 3-5 of the ESG. A check of the information inthis figure indicated that left-turn phasing was needed on themajor street. Thus, the signal phase sequence consisted of threephases: one for the westbound left-turn and through move-ments, one for the east and westbound through movements,and one for the northbound left- and right-turn movements.

Traffic volumes recorded for the morning peak, afternoonpeak, and mid-day off-peak periods were used to determinethe pretimed controller settings. This analysis followed theguidelines associated with Tables 3-7 and 3-8 in the ESG.The settings determined from this analysis were reasoned tobe sufficiently accurate for the engineering study evaluation;however, it was recognized that they might need to be refinedif the signal alternative is selected.

Finally, the cycle length and offset of the signal systemwere determined using guidelines provided in Chapter 3 ofthe ESG. Guidelines associated with Figure 3-10 indicatedthat a 50-s cycle length was needed to provide good two-wayprogression for the existing two signalized intersections(based on a 1,200-ft signal spacing and 35-mph speed). The

A-37

relative offset between the two existing signals would beequal to the travel time of 25 s. The proposed intersectionwould have an offset of 0.0 s relative to one of the adjacentsignals. It was decided to reference this 0.0-s offset to thenearer signal of the two (i.e., Sunset Boulevard).

Bay Alternative. Figure 3-3 in the ESG was consulted todetermine whether an exclusive right-turn bay was appropri-ate on the Main Street approach at the intersection with Sun-set Boulevard. Field observations indicated that a bay wouldreduce the extensive queues found on the eastbound approachby adding capacity and increasing queue storage. Based on theinformation in Figure 3-3, it was determined that two throughlanes and one right-turn lane would be adequate; however, thepeak hour volumes were almost large enough to justify threethrough lanes and one right-turn lane. Therefore, to maintaincontinuity along the route, the existing three through laneswere maintained and a right-turn bay was added.

Base Alternative. Figure 2-4 in the ESG indicates that aminimum of two lanes are needed for the stop-controlledapproach to a two-way stop-controlled intersection. Based onthis information, the Base alternative was assumed to havetwo lanes on the Elm Street approach.

Evaluate Alternatives. Because a stochastic simulationmodel (i.e., CORSIM) was being used, the minimum simu-lation period had to be determined so that statistically reli-able results could be obtained. This period is established byconsidering the minimum simulation period based on theindividual movement delays and the average intersectiondelay. Guidelines in Chapter 3 of the ESG were used to deter-mine the minimum simulation period based on the volume ofthe individual movements and the total volume entering theintersection.

The minimum simulation period based on individual move-ment delay was determined first. Traffic volumes for the morn-ing peak, afternoon peak, and mid-day off-peak hours wereused for this determination. Movements that shared a trafficlane were combined, as suggested in the ESG. Figure 3-12 ofthe ESG was used to determine the minimum simulation periodneeded for each movement volume. The smallest movementvolume for a given hour dictated the minimum simulationperiod based on individual movement delay. The results of thisevaluation are shown in Column 6 of Table A-36.

Next, the minimum simulation period based on averageintersection delay was determined. The total entering volumeat the proposed intersection was consistently the smallestvolume of the three intersections. This volume is listed inColumn 7 of Table A-36. Figure 3-13 of the ESG was usedto determine the minimum simulation period needed for theoverall intersection. The results of this evaluation are shownin Column 8 of Table A-36.

The last step in this process was to compare the minimumsimulation period based on individual movements with that

based on the overall intersection. The larger of these two val-ues was selected as the minimum simulation period used forthe evaluation. It was recognized that a simulation of thisduration would provide the statistical precision needed in alldelay estimates. The simulation period used for the evaluationis shown in Column 9 of Table A-36. To control for time-dependencies in the delay statistic, the 2-hour simulationperiod was divided into two, 1-hour simulation runs. Thedelay statistics from both runs were then averaged for sub-sequent analyses.

Table A-37 summarizes the average delay experiencedduring the afternoon peak hour at each intersection. The Bayalternative (i.e., two-way stop control at Elm Street, add aright-turn bay, and re-timing the existing signals) offeredthe lowest overall delay at two of the three intersections.The Base alternative (i.e., two-way stop control at Elm Street)was the only alternative where some traffic movementsoperated at LOS D or worse (based on a delay threshold of 35 s/veh for unsignalized movements and 55 s/veh for sig-nalized movements).

The impact of each alternative on average intersectiondelay was examined. Table A-38 lists the average delay at

A-38

each intersection for each alternative and study period. Thedata in this table indicate that the Bay alternative tends to offerthe lowest delay at the Elm Street and the Sunset Boulevardintersections. The Signal alternative offers the lowest delay atthe Rio Road intersection.

Figure A-12 illustrates the total delay experienced withinthe three-intersection arterial for each alternative during eachanalysis period. The Bay alternative consistently producesthe lowest total delay. The Base alternative consistently pro-duces the largest delays.

Determine Alternative Effectiveness

The findings from the operations analysis were evaluatedto determine whether each alternative served traffic effec-tively. This determination was made based on the definitionsof “acceptable level of service” and “acceptable operation”as defined in Chapter 3 of the ESG. Table A-39 summarizesthe results of this evaluation.

For a movement to have “acceptable operation,” its level-of-service (LOS) should be no worse than “D” or its total

TABLE A-36 Case Study 4: Minimum simulation period evaluation

TABLE A-37 Case Study 4: Delay summary for peak traffic hour

A-39

TABLE A-38 Case Study 4: Average intersection delay

Figure A-12. Case Study 4: Total delay in signal system.

TABLE A-39 Case Study 4: Alternative effectiveness

It improved overall system operation and provided accept-able LOS and acceptable operation for both signalized inter-sections. However, as noted in Table A-38, it increased delayat the subject intersection. As a result, the Signal alternativewas not found to be effective.

The Bay alternative was the only alternative found to beeffective. This alternative involves re-timing the existing sig-nals at Rio Road and Sunset Boulevard to reflect current traf-fic volumes and adding a right-turn bay on the eastboundapproach at Sunset Boulevard. The new intersection at ElmStreet would have two-way stop control.

STAGE 3: ALTERNATIVE SELECTION

During the alternative selection stage, the alternatives ad-vanced from the engineering study stage are reviewed fortheir potential impact on traffic operations, safety, and theenvironment. The alternative selection stage consists of thefollowing three steps:

1. Identify impacts,2. Select best alternative, and3. Document study.

delay should be less than 4.0 vehicle-hours. Total delay wascomputed using the movement volumes and delays reportedin Tables A-35 and A-37, respectively. For the Base alterna-tive, the movement delays at the signalized intersection ofMain Street and Sunset Boulevard were worse than LOS Dand in excess of 4.0 vehicle-hours. Hence, this intersection’soperation was considered “unacceptable.”

The Signal alternative was found to satisfy most of the“effectiveness” criteria, as defined in Chapter 3 of the ESG.

A-40

Stage 3: Step 1. Identify Impacts

In general, factors are typically identified that might influ-ence alternative selection. These decision factors include:traffic operations, traffic safety, direct cost, environment, andaesthetics. The factors believed to be most relevant includedtraffic operations, safety, and cost. Impacts to the environ-

ment and aesthetics were considered, but reasoned to be un-affected by the proposed alternative.

Stage 3: Step 2. Select Best Alternative

An evaluation of the impacts associated with the Bay alter-native indicated that it had relatively low overall impact.

Figure A-13. Case Study 4: Recommended alternative (not to scale).

Based on this analysis, the Bay alternative was selected forimplementation. This alternative included construction of atwo-way stop-controlled intersection on Main Street, additionof a right-turn bay on the eastbound approach of the intersec-tion at Sunset Boulevard, and adjustment of the arterial signaltiming plan to improve traffic progression. A sketch of therecommended alternative is shown in Figure A-13.

Stage 3: Step 3. Document Study

A report documenting the results of the engineering studywas prepared and submitted.

A-41

REFERENCES

A-1. Manual of Transportation Engineering Studies. D.H. Robert-son, ed. Prentice-Hall, Englewood Cliffs, New Jersey (1995).

A-2. Manual on Uniform Traffic Control Devices-Millennium Edi-tion. Federal Highway Administration, U.S. Department ofTransportation, Washington, D.C. (December 2000).

A-3. Highway Capacity Manual 2000. 4th ed. TransportationResearch Board, National Research Council, Washington,D.C. (2000).

A-4. Robinson, B. et al. “Roundabouts: An Informational Guide.”Report No. FHWA-RD-00-067. Federal Highway Adminis-tration, Washington, D.C. (2000).

B-1

APPENDIX B

MUTCD TRAFFIC CONTROL SIGNAL WARRANT WORKSHEETS

INTRODUCTION

This appendix describes a series of worksheets that can beused to document the findings of the MUTCD 2000 traffic con-trol signal warrant check. The following pages illustrate theuse of the worksheets for Case Study 1 described in AppendixA. Blank worksheets are provided following this exampleapplication.

EXAMPLE APPLICATION

The intersection of County Routes 21 and 27 is in a ruralcommunity with approximately 11,000 residents. The inter-section has three approach legs; one leg is stop-controlled.County Route 27 (CR 27) is the minor road; it is oriented in aneast-west direction and has stop control at the intersection.County Route 21 (CR 21) is the major road; it is oriented in thenorth-south direction. The speed limit is posted at 45 mph onboth roads. All approaches have one traffic lane. Figure B-1illustrates the intersection geometry.

Crash data were obtained from the state’s Department ofPublic Safety. These data indicated that a total of five colli-sions occurred at the intersection in the last 12 months (twosusceptible to correction by a traffic signal). Regional trafficcounts recorded 1 year earlier indicated that the annual aver-age daily traffic demand (AADT) is about 13,500 vehiclesper day on CR 21 and 3,000 vehicles per day on CR 27.

Because it was unclear when the 8 highest traffic hoursoccurred, turn movement volumes were recorded for a totalof 13 hours. These counts were adjusted to represent average-day volumes using the techniques described in Appendix C.The adjusted volumes are summarized in Table B-1.

The morning peak demand occurred between 7 and 8 a.m.,the afternoon peak occurred between 4 and 5 p.m., and theoff-peak hour occurred between 9 and 10 a.m. No pedestrianswere observed during the study. Heavy vehicles constitutedabout 1.0 percent of the traffic stream on each approach. Aspot speed study was performed for CR 21; the 85th percentilespeed was found to be 50 mph.

A stopped-delay study was performed during the highest-volume hour (i.e., 4 to 5 p.m.). This study focused on thedelay to minor-road vehicles. The average delay during thepeak hour was 40 s/veh and the total delay was 2.3 veh-h.

A complete warrant analysis for this intersection is summa-rized in the worksheets on the following pages. The analysisrevealed that Warrants 1, 2, and 3 are satisfied and that a traf-fic control signal may be a viable alternative for the subjectintersection.

WORKSHEETS

Blank worksheets for documenting the signal warrantcheck follow the pages illustrating the use of the worksheetsfor Case Study 1 described in Appendix A.

B-2

TABLE B-1 Turn movement volumes

Figure B-1. Existing intersection geometry (not to scale).

B-3

B-4

B-5

B-6

B-7

B-8

B-9

B-10

B-11

B-12

C-1

APPENDIX C

DATA FOR GUIDELINE EVALUATION

INTRODUCTION

This appendix describes the data collection requirementsfor the engineering assessment process. These requirementsare categorized into the following sections:

1. Procedures for Collecting Site Data and2. Procedures for Estimating Traffic Data.

The first section describes procedures for collecting dataneeded for the alternative identification and screening stageof the assessment process. The second section describes aprocedure for estimating the data needed for the engineeringstudy stage. Blank worksheets for recording the data areincluded in the third section of this appendix.

PROCEDURES FOR COLLECTING SITE DATA

This section describes two data collection activities for thealternative identification and screening stage of the assess-ment process. These data are needed to define the operationalproblems at the subject intersection and identify their causes.One data collection activity includes a site visit where a con-dition diagram and an on site observation report are com-pleted. A second data collection activity requires the acquisi-tion of crash records for the site and a presentation of the crashhistory in a collision diagram. The activities associated withthe development of these summary documents are describedin the next three sections.

Condition Diagram

The condition diagram represents a complete physicalrecord of relevant site characteristics. The diagram consistsof a plan-view, scale drawing of the features in the vicinityof the subject intersection. Some of the items that might beshown on the diagram include

• Names of major and minor road;• Functional class (arterial, collector, local);• Intersection angle;• Lane configuration and lane use markings;• Approach grade;• Lane widths;• Length of turn bays;• Location and legend of traffic signs;• Location of on-street parking;

• Location of medians and islands;• Location of transit stops;• Location of driveways;• Location of any roadway lights;• Surrounding land uses;• Location of sidewalks and crosswalks;• Location of drainage structures;• Distance to nearest traffic signals;• Nearby railroad crossings; and• Location of potential sight obstructions (may include

vegetation, utility poles, fences, buildings, mailboxes,and controller cabinets).

The items shown on the diagram can vary, depending onthe analyst’s assessment of the relevance of various items (asthey relate to the reported problems) and the anticipated ex-tent of potential improvements. For example, an anticipatedsigning change may not require as much detail as when traf-fic lanes are proposed to be added to an approach. A samplecondition diagram is shown in Figure C-1.

An accurate and complete diagram will help the engineerselect alternatives that are appropriate for the site. Locationmeasurements should be performed with a distance measur-ing wheel (or odometer). This device offers a good balancebetween precision and ease of measurement.

Photographs of the site may also be taken to supplementthe information obtained from the sketch. Photos can providea visual record of the conditions present at the intersection.

A blank condition diagram is included in the Worksheetssection of this appendix.

On Site Observation Report

An important component of the problem-cause identifica-tion process is the firsthand observation of traffic operations.As such, it is essential that the engineer responsible for theassessment be present during the site visit. This visit shouldbe scheduled to coincide with the occurrence of the reportedproblems (e.g., p.m. peak traffic demand hour).

The observational study should include several compo-nents. First, the engineer should drive through the subjectintersection and attempt to experience the problem, espe-cially if it relates to sight distance limitations. Then, theengineer should observe traffic operations at the inter-section. An On Site Observation Report, as shown in Fig-ure C-2, should be completed during the site visit. A blankreport form is included in the Worksheets section of thisappendix.

During the observational study, the engineer should attemptto identify the circumstances that contribute to the reported (orobserved) problems. The questions included in the report areintended to direct the engineer’s attention to situations thatoften cause problems at intersections. However, the engineershould also look for other site-specific factors that may con-tribute to the problem.

The questions on the report form are answered by check-ing “No,” “Not Sure,” or “Yes.” A response of “Not Sure” or“Yes” should be considered as an indication that a problemmay exist at the intersection. The engineer should describethe problem and its likely cause(s) in the Comments sectionof the report.

C-2

Collision Diagram

The collision diagram documents the spatial orientation ofthe crashes that have occurred at the intersection. The dia-gram shows all crashes that have occurred during the past 1to 3 years. The collision diagram includes a sketch (withoutscale) of the intersection curb lines and identifies roads bytheir name and direction. A symbol drawing is then added tothe diagram for each crash that occurred. The basic charac-teristics of each accident, such as date, day of week, time ofcrash, weather, pavement condition, and number of injuries,are recorded next to each symbol. A sample collision dia-gram is shown in Figure C-3.

Figure C-1. Sample condition diagram.

The information provided on the diagram can provideimportant clues to the cause of the reported problem. Forexample, frequent right-angle collisions may indicate thatsight lines are obstructed; frequent nighttime collisions mayindicate a need for roadway lighting. More information onthe preparation of and interpretation of collision diagrams isprovided by Hummer (C-1). A blank collision diagram work-sheet is provided in the Worksheets section of this appendix.

PROCEDURES FOR ESTIMATING TRAFFIC DATA

This section describes a procedure for estimating the ap-proach and turning movement volumes at a proposed inter-

C-3

section. The objective of the procedure is to estimate thesevolumes for the average day. The “average day” is defined asa day representing traffic volumes normally and repeatedlyfound at a location. The average day is typically a weekdaywhen volumes are influenced by employment; however, itcould be a weekend day when volumes are influenced byentertainment or recreation.

The procedure is based on the assumption that the intersec-tion does not exist at the time of the study and that only an esti-mate of the average daily traffic (ADT) is available for eachof the intersecting roadways. If one of the roads exists and atraffic count is taken on it, steps are described for adjusting thecount to offset seasonal variations in traffic demand. If hourly

Figure C-2. Sample onsite observation report.

traffic volumes on each intersection approach are availablefrom existing counts or from a traffic assignment process as-sociated with the proposed intersection, then Steps 1, 2, and 3can be skipped.

Step 1. Estimate Volume Distribution Factors and Turn Percentages

As a first step, the volume distribution factors and turnmovement percentages are estimated. The volume distribu-tion factor D represents the percentage of the two-way vol-ume on a given intersection leg at the intersection. Typicalpercentages range from 40 to 60 percent. If local field mea-

C-4

surements are not available, then D can be estimated usingthe factors in Table C-1.

The turn movement percentages must also be establishedin this step. The percentages in Table C-1 can be used if esti-mates based on local traffic patterns are unavailable. Thesepercentages were reported by Hauer et al. (C-2) based on ananalysis of turn volumes on 283 intersection approaches.

Step 2. Obtain Volume Adjustment Factors

The procedure requires adjustment factors that reflecthourly, weekly, and monthly volume variations for roadwaysin the vicinity of the proposed intersection. These factors are

Figure C-3. Sample collision diagram.

typically obtained from the permanent network of controlcount locations established by the state’s Department ofTransportation. Table C-2 illustrates these factors for a typi-cal urban location. These factors are for illustrative purposesonly; the analyst must obtain similar factors from the localstate agency.

Step 3. Compute Approach Volumes

The approach volumes for a given hour are estimatedusing the hourly adjustment factors from Table C-2 and thevolume distribution factors identified in Step 1. The equationused for this computation is:

C-5

where,

Va,i = approach volume during hour i (i = 1, 2, 3, . . . , 24),veh/h;

ADT = average daily traffic, veh/d;fh,i = hourly volume adjustment factor for hour i (percent

of AADT); andDi = volume distribution factor for hour i.

The ADT variable can represent the annual average dailytraffic (AADT) or an estimate of it based on the planningprocess. It can also represent a 24-hour count made on a

V ADTf

Da ih i

i,,= × ×

100(C-1)

TABLE C-1 Typical volume distribution factors and turn movement percentages

TABLE C-2 Illustrative hourly, weekly, and monthly volume adjustment factors

specific day, in which case, the day-of-week and month-of-year adjustments described in Step 5 should be used to removeany bias due to weekly and monthly volume variations.

To illustrate the application of Equation C-1, consider aproposed intersection on an existing road. The existing roadis a major arterial having an AADT of 10,000 veh/d. Themajor road is oriented in a north-south direction. The minorroad is planned as a collector with an estimated ADT of5,000 veh/d. The intersection is located on the northeast sideof a major city. During the afternoon peak hour, traffic flow ispredominantly in the northbound and eastbound directions.The volume distribution is estimated as 60 percent in each of these travel directions (and 40 percent in the opposite direc-tions) during the afternoon peak. This trend is reversed duringthe morning peak.

The intersection approach volumes calculated using Equa-tion C-1 are shown in Table C-3 for the eight highest hoursof the day. Volumes for these hours were computed to facil-itate the warrant check. The turn movement volumes shownare the subject of discussion in the next calculation step.

Step 4. Compute Turn Movement Volumes

The turn movement volumes are computed using the turnpercentages identified in Step 1. For each approach, the turnmovement volume is computed as the product of the turn per-centage and the approach volume. Table C-3 illustrates thiscomputation for the eastbound approach. The turn percent-ages were obtained from Table C-1 for the “Collector to arte-rial” facility type.

Occasionally, the turn movement volumes obtained fromthis procedure are not in reasonable proportion to the origi-nal ADT volumes associated with each intersection leg. Thisdisparity may occur when one leg ADT is unusually large (or

C-6

small) relative to the other leg ADTs. If this situation occurs,the procedure described by Hauer et al. (C-2) or that providedin Chapter 10 of the Highway Capacity Manual 2000 (C-3)can be used to refine the estimate of turn movement volumes.

Step 5. Compute Average-Day Volumes

Two adjustments to the volume data are considered duringthis step. First, an adjustment is needed if the volumes usedare based on counts taken on a specific day of the week andmonth of the year. This adjustment converts the count data toaverage-day-of-year volumes. It can be applied to the countdata in whatever form it is collected (e.g., 24-hour two-waytotal or 6-hour turn movement count); however, this procedureillustrates the adjustment being applied to the estimated turnmovement counts. The equation for making this adjustment is:

where

Vt,doy = turn movement volume reflecting the average-day-of-year, veh/h;

Vt = turn movement volume count, veh/d;fw = weekly volume adjustment factor (percent of

AADT); andfm = monthly volume adjustment factor (percent of

AADT).

If the turn movement volumes are derived from AADT val-ues or otherwise represent average-day-of-year volumes,then the adjustment using Equation C-2 is not needed.

The second adjustment converts the average-day-of-yearvolume to the average-day volume required for the engineer-

V Vf f

t doy tw m

, = × ×100 100(C-2)

TABLE C-3 Example computation of approach volumes

ing study. The average-day volume is computed by adjustingthe average-day-of-year volume to reflect the average week-day. The data in Table C-2 suggest that the average weekdayvolume is 105 percent of the average-day-of-year volume.This percentage is based on the average of the five weekdayaverages listed in Table C-2. A similar trend may be foundin the adjustment factors obtained in Step 2 for the subjectlocation.

The average-day volume is computed using the followingequation:

where

Vt*= average-day turn movement volume, veh/h; andfad = average weekday adjustment factor.

The use of Equations C-2 and C-3 are illustrated inTable C-4. This example assumes that the ADTs used to es-timate the turn movement volumes were obtained from a 24-hour count taken on a Wednesday in April. The factorslisted in Table C-2 indicate that Wednesday is about equal tothe average day-of-week; however, a count in April mayactually underestimate the true annual average daily volumeby 5 percent. Therefore, Equation C-2 is needed to convertthe turn movement volumes to average-day-of-year volumes.Then, Equation C-3 is needed to convert these volumes toaverage-day volumes. The total adjustment from both equa-tions combined represents an 11.0-percent (= 1.05/0.95/1.00)increase in the measured volumes. The resulting, average-dayvolumes are listed in the last three columns of Table C-4.

V Vf

t t doyad* ( ),= × −

1003C

C-7

This step of the procedure was illustrated using the moretypical case where the weekdays are a better representationof the “average day” than the weekend days. If the weekenddays are a better reflection of the “average day,” then thevariable fad would be estimated as the average of the Satur-day and Sunday day-of-week volume percentages. The othercomputations would be the same as previously stated.

WORKSHEETS

Blank worksheets for the following studies and analysesfollow the Reference section:

1. Condition Diagram,2. On Site Observation Report,3. Collision Diagram,4. Critical Volume Worksheet, and5. Controller Setting Worksheet.

REFERENCES

C-1. Hummer, J.E. “Chapter 11 - Traffic Accident Studies.” Man-ual of Transportation Engineering Studies. D.H. Robertson,ed. Prentice-Hall, Englewood Cliffs, New Jersey (1995).

C-2. Hauer, E., Pagitsas, E., and Shin, B.T. “Estimation of TurningFlows from Automatic Counts.” Transportation ResearchRecord 795. Transportation Research Board, National Re-search Council, Washington, D.C. (1981) pp. 1–7.

C-3. Highway Capacity Manual 2000. 4th ed. TransportationResearch Board, National Research Council, Washington,D.C. (2000).

TABLE C-4 Example computation of average-day volumes

C-8

C-9

C-10

C-11

C-12