Incident Management, FAST-TRAC Phase IIB Deliverables€¦ · Van Aerde and Plum ( 1988). The effectiveness of each traffic control strategy depends on demand, the network. and control

IncidentManagement

FAST-TRACPhase IIB Deliverables

#20. The Model Analysis Report on the Benefitsof SCATS in Alleviating the Impacts of Incidents

EECS – ITS LAB – FT96 – 112

William C. TaylorSorawit Narupiti

Michigan State UniversityEast Lansing, 48824 – 1226

Telephone: (517) 355-5107Fax: (517) 432-1827

FAST-TRAC

ABSTRACT

Incident ManagementFAST-TRAC

Phase IIB Deliverables

#20. The Model Analysis Report on the Benefits of SCATS in Alleviating the Impacts of IncidentsEECS-ITSLAB-FT96- 112

BY

Dr. William C. Taylor

Sorawit Narupiti

Selecting the most appropriate traffic control strategy for incident congestion management can have amajor impact on the extent and duration of the resulting congestion. This research investigated the effective-nesses of several control strategies on various incident conditions. The selected control strategies representingpossible ITS technologies included traffic metering (ATMS), traffic diversion (ATIS), and traffic diversionwith signal timing modification (ATIS/ATMS). The analysis was conducted on a hypothetical dense grid sur-face street network. Mid-block incidents of various durations were tested. The results indicated that theATIS/ATMS based solution can reduce congestion duration up to 27 minutes, with a saving of 261 vehicle-hours of delay. A sensitivity analysis was performed to obtain the effectivenesses of various control strategiesunder different demand levels. Several alternate diversion plans and control variables for initiating signalmodification were also tested.

The Model Analysis Report on the Benefits of SCATS in Alleviating the Impacts of incidentsFAST-TRAC - Phase IIB #20

FAST-TRAC

TABLE OF CONTENTS1. INTRODUCTION ............................................................................................................................................... 1

1.1 Description of the problem.......................................................................................................................... 11.2 Statement of the problem ............................................................................................................................ 31.3 Objective and scope of the research .......................................................................................................... .31.4 Research approach ..................................................................................................................................... .4

2. LITERATURE REVIEW....................................................................................................................................52.1 Theoretical traffic signal control strategies ............................................................................................... .52.2 Signal control strategies and response to incidents in existing urban traffic control systems .............. .62.3 Route guidance and access control strategies.. ......................................................................................... .72.4 Effectiveness of control strategies for incidents ....................................................................................... .82.5 summary ................................................................................................................................................... 10

3. DESIGN OF STUDY ........................................................................................................................................ 123.1 NETSIM simulation model.. ..................................................................................................................... 123.2 Hypothetical network formulation ........................................................................................................... 123.3 Traffic demand and initial traffic signal setting.. .................................................................................... .133.4 Incidents.. .................................................................................................................................................... 153.5 Technological scenarios and control strategies ...................................................................................... .153.6 Measures of effectiveness (MOEs)) .......................................................................................................... 163.7 Study plan ................................................................................................................................................... 183.8 Summary .................................................................................................................................................... 18

4. SIMULATION ANALYSIS AND RESULTS.............................................................................................. .214.1 Effect of type and duration of incident.. .................................................................................................. .214.2 Effect of different control strategies ....................................................................................................... .294.3 Summary .................................................................................................................................................... 38

5. SENSITIVITY ANALYSIS ............................................................................................................................. 395.1 Sensitivity to trafficc demand change ....................................................................................................... .395.2 Sensitivity to diversion control alternatives.. .......................................................................................... -465.3 Sensitivity to signal control variable alternatives ................................................................................... .525.4 Sensitivity to signal control coverage.. .................................................................................................... .545.5 summary .................................................................................................................................................... 58

6. CONCLUSION AND RECOMMENDATION.. .......................................................................................... .596.1 Conclusion.. ................................................................................................................................................596.2 Recommendation ....................................................................................................................................... 61

BIBLIOGRAPHY.........................................................................................................................................63

- The Model Analysis Report on the Benefits of SCATS in Alleviating the Impacts of IncidentsFAST-TRAC - Phase IIB #20

1. INTRODUCTION1.1 Description of the problem

Selecting the most appropriate traffic controlstrategy for incident congestion management canhave a major impact on the amount and duration ofthe resulting congestion. With the implementationof the Intelligent Transportation System (ITS) tech-nology, several advanced traffic control options arenow available. However, this topic has not beenthoroughly studied. In this study, several ITS controlstrategies are tested to determine their effective-nesses under various incident and control conditions.

1 .1 .1l Incidents and their effect on trafficTraffic incident is a term used for a random,

unplanned event which affects traffic operations. Itcan also be a temporary reduction in roadway capac-ity caused by accidents, roadway maintenance, orconstruction activities, or a temporary increase intraffic demand as commonly occurs immediatelybefore and after special events. Incidents, generallyreferred to as nonrecurrent events, can be dividedinto three basic categories based on their uncertaintyof occurrence (Holmes and Leonard, 1993):

a. Normal and generally accepted (although notnecessarily desirable)--such ason-streetparking. This type of incident is usually tol-erated by drivers as being part of normal traf-fic conditions on the network.

b. Expected and programmed--includes road-work and maintenance activities. Theiroccurrence is foreseen and planned, butunexpected to motorists.

c. Unexpected--such as vehicle breakdowns andaccidents. Neither drivers nor the trafficagency is prepared for this type of the inci-dent because the occurrence is unpredictable.

Incidents generally result in traffic congestion. Themagnitude and duration of the congestion is difficultto predict because, in most cases. neither the loca-tion, the time, nor the severity of an incident isknown beforehand. Limited information and knowl-

edge on incidents can result in non-optimal choices.such as an improper route choice and ill-timed signalsettings.

Each incident has a unique impact on traffic opera-tions in a network. Beaubien (1994) analyzed theeffect of incidents on delay for various incidenttypes. Similar work was also conducted in Texas(Dudek, 1976) and in London, England (Holmesand Leonard, 1993). Delay generally ranged from15-30 vehicle-minutes due to a vehicle disablementto several vehicle-hours due to major accidents withlane blockages.

Incident management has received considerableattention since incidents are one of the most pressingtraffic problems in urban areas. Lindley (1987,1989) reported that congestion due to incidentsaccounted for more than 61 percent of all urban con-gestion in 1984, and is expected to cause more than70 percent of the total congestion by the year 2005(National Cooperative Highway Research Program,1991).

1 .1 .2 Measures to alleviate congestiondue to incidents

Incidents commonly create two types of conges-tion: primary and secondary. Primary congestion iscaused by traffic queuing at a bottleneck. Secondarycongestion arises from the blockage of other inter-sections by primary congestion (Longley, 1968).The goal of incident management is to restore road-way capacity as quickly as possible, so as to limitprimary congestion, and at the same time avoid orreduce secondary congestion.

Incident management shares the same traffic treat-ment philosophy as recurrent congestion manage-ment, although different traffic control strategiesmay be required. Recurrent congestion occurs rou-tinely at specific locations and times of day. Anexample is the peak hour traffic jam. Both short-term and long-term solutions are considered asmethods to alleviate recurrent congestion. Consider-

The Model Analysis Report on the Benefits of SCATS in Alleviating the Impacts of IncidentsFAST-TRAC - Phase IIB #20

1

ing the techniques for reducing recurrent traffic con-gestion, which can be divided into three categories(Rathi and Leiberman. 1989):

a. Increasing the capacity of the road system--through construction of additional facilitiesor a physical improvement to provide addi-tional capacity.

b. Reducing traffic demand--through behavioralchanges or travel demand management ortraffic restraints.

c. Maximizing the use of available capacity--through traffic engineering practices aimedat minimizing the capacity reducing factors(e.g. through traffic regulation) or more effi-cient use of existing capacity (e.g. signalcontrol improvement).

The first method is not appropriate for incident man-agement. The unpredictable nature of incidentsmakes it infeasible to maintain excess capacity at allelements of the network. Behavioral changes arealso not applicable to incident management becauseof the time frame involved (long-term solution).However, traffic restrictions in congested areas, suchas metering and traffic diversion, may be applied toincident management. The efficient use of existingcapacity through traffic management is the onlyresponse applicable to nonrecurrent congestion.Therefore, measures to alleviate incident congestionfocus on maximizing the use of available capacity.

Actions to reduce incident congestion can specifi-cally be categorized into three groups (Van Vurenand Leonard, 1994):

a. Incident control

b. Behavioral control or demand control

c. Network control

Incident control deals with the initial cause of con-gestion. Techniques include reduction in incidentduration (i.e., incident removal) and local trafficmanagement to reduce the impact of the incident atthe scene. Behavioral control includes the provisionof information to motorists so that they can adjust

their travel pattern to avoid the congested linksthrough route diversion, a departure time change. amode change, or even trip cancellation. The routechange is the only response to the incident that canbe applied to motorists currently on the streets. Net-work control is the efficient use of available networkcharacteristics, including throughput and storagecapacity. This method employs measures such assignal control alterations.

1 .1 .3 Strategies for incidentmanagement

Most traffic control measures attempt toimprove network efficiency, i.e., reduce delay for allvehicles using the system. With this goal in mind, acriterion by which success will be measured and atechnique to achieve the goal must be developed fora specific application. Such a technique is called astrategy.

In the course of incident management, traffic controlstrategies can be classified into two main categories:

a. Signal modification

b. Traffic diversion

Signal modification strategies are generally imple-mented at traffic control centers. Examples of signalmodification strategies are longer or shorter cycletime, phase changes to reflect current demand,changes in the green splits and offsets to maintainequal queues for conflicting movements, trafficmetering to avoid blockage, and reverse progression.The signal modification strategies require a signalcontrol system that is responsive to changing trafficdemand. Signal control strategies have been exten-sively reviewed in the literature (Wright and Hud-dart, 1989; Montgomery and Quinn, 1992; Quinn,1992; OECD, 1981; McShane and Pignataro, 1978).

The strategies for driver information are pre-tripinformation and route guidance. These strategiesattempt to provide knowledge of traffic conditions todrivers so they can make route choices which mini-mize the effect and extent of the incident. Commu-nication through in-vehicle devices or changeablemessage signs is required to give the drivers incident


2

FAST-TRAC

and routing information. These control policieswere surveyed in Van Aerde and Rakha (1989) andVan Aerde and Plum ( 1988).

The effectiveness of each traffic control strategydepends on demand, the network. and control char-acteristics. Signal alteration alone is applicable onlywhen the demand does not exceed the total networkreduced capacity after an incident occurs. Trafficmetering requires some links to be designated forqueue storage. The effectiveness of traffic routingdepends on the availability of alternate routes andtheir level of congestion. Therefore, there is no sin-gle most appropriate control strategy that can beapplied in all situations.

1 .1 .4 Intelligent Transportation Systems(ITS) and incident congestion

In the past decade, the ability to improve perfor-mance of the transportation systems has been madepossible through the advancement of computer,communication, and information technologies. Thiseffort has been pursued in many parts of the world.In the United States, Intelligent Transportation Sys-tems (ITS), formerly Intelligent Vehicle-HighwaySystems (IVHS), is a single phrase describing theuse of technologies to accomplish transportationgoals in many functional areas. The European com-munity calls the program Advanced TransportTelematics (ATT) and Road Transport Informatics(RTI) (Castling, 1994).

ITS offers potential incident congestion reductionthrough Advanced Traffic Management System(ATMS) and Advanced Traveler Information System(ATIS). The primary feature of ATMS is the provi-sion of real-time (dynamic) control to respond tochanging traffic conditions. ATIS provides travelerswith information required to ensure that their jour-ney is as efficient and safe as possible. The com-bined effect of ATMS and ATIS has been studied byRakha et. al. (1989) and Sarakki and Kerr (1994).There are several demonstration programs employ-ing these two technologies currently underway in theUnited States, Europe, and Japan. Thus, the scenario

of vehicles equipped with ATIS devices and circulat-ing in a network with an ATMS traffic control is inthe near future.

However, control strategies in the ITS environmentneed further investigation to determine the effective-ness of each strategy under various situations. Thebenefits of ITS deployment cannot be fully realizedif inappropriate control strategies are applied.

1.2 Statement of the problemTraffic control strategies aimed at reducing the

consequences of incidents have not been thoroughlydeveloped and tested. The research conducted forthis study is original in that different control strate-gies developed for a variety of incident situations inan ITS environment were tested. The effectivenessof controls under each individual ITS element wasobtained. The joint effect of ATMS and ATIS con-trol strategies was also examined as benefits gainedthrough their interaction may be lost if only one ofthe strategies is employed.

1.3 Objective and scope of theresearch

This research addresses the development andtesting of traffic control strategies designed to reducethe consequences of an incident. The scope of theresearch includes:

a.

b.

C.

d.

Identification of measures of effectiveness(MOEs) to indicate the impact of an incidenton an urban street network.

Investigation of the’impacts of incidents onthese MOEs under various conditions.

Development of routing and signal controlstrategies to cope with the incidents.

Determination of the limits of the effective-ness of these strategies by varying degrees ofdemand, incident severity, and incident dura-tion.

-

The Model Analysis Report on the Benefits of SCATS in Alleviating the Impacts of lncidentsFAST-TRAC - Phase IIB #20

3

1.4 Research approachThis research was based on traffic simulation

since this permits the analysis and comparison ofdifferent control strategies on the same road networkand incident. NETSIM (NETwork SIMulation) wasselected for this study because it could be modifiedto replicate control and drivers characteristics withinthe ITS environment. NETSIM is a microscopicinterval-based simulation model of urban traffic on asurface street network. The model was first devel-oped in the 1970s and has periodically beenenhanced. NETSIM version 5.0, which was used inthis study, includes many advanced features on traf-fic signal, driving behavior, and turning movementdescriptions and can provide data on the MOEs suit-able for the analysis (Federal Highway Administra-tion, 1995).

The research was based on a hypothetical dense gridnetwork with demand characteristics representativeof traffic conditions in the City of Troy, Michigan.

FAST-TRAC

When an incident was introduced into the network,the evolution and dissipation of congestion werestudied. Congestion resulting from an incident in anetwork without ITS was used as a base case, withtraffic performance analyzed for various types ofincidents. For the base case, the signal timing washeld constant and the impacts of the incidents onspecified MOEs were determined.

Several traffic signal control and route diversionstrategies were developed for each traffic situation.These strategies were then tested with the simulatednetwork to obtain performance measures in variousincident characteristics. Data for the MOEs werecollected for each control scheme and the resultswere compared and discussed.

The effectiveness of these control strategies underdifferent demand condition was then determined.Moreover, variations of control strategies were eval-uated.


4

2. LITERATURE REVIEWTraffic control strategies have been developed

and incorporated in urban traffic signal control sys-tems since the 1960s. Along with the advancementin traffic control systems and technologies, manytraffic control strategies have been developed to pro-vide efficiency, safety, and a reduction in fuel con-sumption. Most of the signal control policies in thepast were developed for recurrent traffic conditions,both for peak and off-peak traffic. As the nonrecur-rent traffic congestion problem on urban streetsincreases, recent interest has shifted to traffic controlstrategies under incident conditions. Traffic controlstrategies for incident conditions were not possibleuntil responsive traffic signal controls and communi-cation technologies existed, due to the requirementfor a prompt response to the incident.

Studies on control strategies for incidents started inthe 1970s with the evaluation of responsive signalcontrol systems and driver information systems onunexpected situations. Nonetheless, the first majorwork was conducted by Hunt and Holland in 1985,where an attempt was made to determine the effectof an incident in a network controlled by theSCOOT (Spilts, Cycle length, Offset OptimizationTechnique) traffic control system. Since thenincreased attention has been given to the ability ofeach modem traffic control systems to respond withspecial control strategies for incidents.

Most traffic control strategies for incidents arederived from those for recurrent congestion condi-tions (both undersaturated and saturated conditions).Because of this, the following section reviewed thecontrol strategies originally developed for recurrenttraffic conditions. The transferability to an incidentsituation was then discussed, along with the researchon effectiveness of numerous control strategies onincident-caused congestion.

2.1 Theoretical traffic signal controlstrategies

The first generation of signal control strategieswere based on off-line calculations for a fixed timesignal control system. At an individual intersection,Webster and Cobbe (1958) suggest that a signal splitstrategy should equalize the degree of saturation(DS) of all critical approaches to approximatelyyield the minimum intersection delay. This controlstrategy can handle varying demand in a day by hav-ing separate settings for different time periods. Thistechnique became general practice for most fixedtime signal controls, in both undersaturated and satu-rated traffic conditions.

For arterial and network considerations, Little(1966) introduced an off-line mathematical tech-nique to maximize the bandwidth. The principlewas to maximize the number of vehicles able to suc-cessfully encounter green signals when travelingalong a street. Over the years, several variations ofthis approach were developed. NCHRP Report 73(1969) evaluated several of these offset strategies.Off-line control techniques investigated in this reportwere Yardeni’s time-space design, Little’s maximalbandwidth, and delay/difference-of-offset. Threeresponsive control strategies, namely basic queuecontrol, cycle and offset selection, and mixed cyclemode, were also evaluated in this report. The resultsindicate that the cycle and offset selection methodand delay/difference-of-offsets techniques rank thebest for off-peak periods, whereas the mixed cyclemode and basic queue control were the best for peakperiods.

Perhaps the most comprehensive and most widely-applied control strategy for fixed time control settingis based on a computer optimization method.TRANSYT (Traffic Network Study Tool) (Robert-son, 1968) is an off-line program utilizing modifiedWebster’s method to calculate green splits, and ahill-climbing optimization technique to determineoffset and cycle length which minimizes a perfor-mance index. The logic for the offset calculation is

--


5 -

-

FAST-TRAC

similar to the delay/difference-of-offset method eval-uated in NCHRP Report 73. Similar programs toTRANSYT are SIGOP (Signal Optimization pro-gram) and PASSER II (Progression Analysis andSignal System Evaluation Routine), which haveslightly different calculation procedures.

The control strategies obtained from off-line calcula-tions are effective for average traffic conditions, butare not responsive to changing traffic patterns.When a traffic pattern changes, the solutions fromthese programs are no longer optimal. Thus, theTRANSYT method is suitable only for recurrenttraffic conditions. These methods are difficult toapply to incident conditions, where the incidentalters the capacity and demand pattern.

Several traffic-responsive signal control strategieswere developed for an individual intersection fur-nished with traffic detectors. Gazis and Potts (1964)developed a technique for time-dependent signal set-ting under varying demand. They used queue lengthas an input to minimize total aggregate intersectiondelay. The technique is also called “bang-bang”because the green time is set at a predeterminedmaximum value for the queued approach, and at aminimum value in other directions. When a queuein the first approach is cleared, the setting isreversed. This signal setting does not, in general,minimize the period during which one approach iscongested. d’ Ans and Gazis (1976) furthered thiscontrol method by means of linear programming.Church and Revelle (1978) formulated similar con-trol strategies with consideration of maximum wait-ing time and queue length. When the maximumqueue length was used as a control objective, theyfound that the solution tended to balance the queuelengths on the most saturated approaches of eachsignal phase, and the signal frequently switchedbetween phases. Michalopoulos and Stephanopo-lous (1977) reported that the queue constraint waseffective when the demand increases to the limitingvalue. The optimal control strategy at saturation issimply the balance of input-output to maintain con-stant queue length.

NCHRP Report 32 (1967) tested four control strate-gies, namely basic queue control, queue-lengtharrival rate control, modified space-presence control.and delay-equalization control. The results showedthat the modified space-presence control strategyyielded the lowest delay under low to medium inter-section demand (up to 2000 vehicles per hour for 4-lane, 4-leg junction). When the demand was greaterthan 2000 vehicles per hour, the basic queue controlstrategy was better than the others.

Many control strategies have been developed foroversaturated traffic conditions at an isolated inter-section. Gordon (1969) suggested that the controlobjective should be to maintain a constant ratioamong the respective storage spaces. Longley(1968) attempted to balance queue lengths on allapproaches. NCHRP Report 194 showed that,although the Longley control logic yielded lowerdelay than the off-line calculation, the queue-actu-ated control resulted in lower delay when the degreeof saturation was above 0.5. The report stated thatthe objective of signal control should be to avoidspillback and to provide equitable service. Thereport also gave some tactical control strategies toease queue blockage at an intersection.

2.2 Signal control strategies andresponse to incidents in existingurban traffic control systems

Under a fixed time control system, the TRAN-SYT method of delay minimization is a popularmethod to determine signal timings (Woods, 1993).However, such timings cannot respond to unex-pected incidents. Because the signal setting is fixedfor a time-of-day period, there is no special controlstrategy for incidents.

Similar control strategies were used in the UTCS(Urban Traffic Control System) first generation con-trol system as in the fixed time control systems, butsignal plans could be changed every 15 minutes. Asignal timing plan suitable for current traffic condi-tions is selected from a set of pre-calculated plan.The UTCS first generation system includes splitadjustment for the Critical Intersection Control

The Model Analysts Report on the Benefits of SCATS In Alleviating the Impacts of Incidents 6FAST-TRAC - Phase IIB #20

FAST-TRAC

(CIC). The signal split is altered if oversaturation isdetected, and queue control comes into place at thiscritical intersection.

The UTCS second generation system computes anew signal plan instead of “looking up” a selectionin the plan library. Signal split and offset are thenadjusted at the critical intersection. The UTCS thirdgeneration system controls intersections indepen-dently, with fully adaptive split, offset, and cyclelength determination. The signals can be changedevery 3-6 minutes. Because the second and thirdgeneration systems are on-line and more responsiveto changing traffic patterns, they can respond to inci-dent-related traffic but the effectiveness of these sys-tems to an incident has not been evaluated (Kay,Allen, and Bruggeman, 1975).

In the Japanese UTMS (Urban Traffic ManagementSystem), five control strategies are selected based onthe level of traffic demand. Three strategies corre-spond to levels before network saturation.- Signalsare optimized similar to TRANSYT in undersatu-rated conditions. In oversaturated conditions, theobjective is switched to prevent blockages and togive priority to main roads by restraining accessfrom side streets. In the case of incidents, the systemkeeps priority routes clear during particularly severecongestion (Woods, 1993).

The SCATS (Sydney Co-ordinated Adaptive TrafficSystem) of Australia also has different control strate-gies for various demand levels. Signal splits and off-sets are selected from embedded plans calculated byan off-line program such as TRANSYT, while thecycle length is calculated every cycle. However, thesystem also makes use of some tactical controls ateach intersection (Lowrie, 1982). Response to inci-dents is primarily activated by traffic operators.When the detectors are covered by traffic for certainperiods of time, an alarm is signaled to the trafficoperator who sets the traffic control. SCATS tacticallogic itself can also respond to incident-related con-gestion. The logic is the same as the normal recur-rent traffic operation. At each intersection. tacticalcontrol strategies include:

a. Signal split selection from a library according

b.

C.

d.

to degree of saturation:

Green time gap-out:

Green time early cut-off due to inefficient use -of green time; and

-Phase skip if no demand is placed in the pre-vious cycle length.

At a “strategic” level of control, offset and cycle timeare selected in response to the current traffic situationbased on the plan selection process. However, theplans are not typically developed for incident situa-tions. In principle, when an incident occurs on alink, there is a reduction in traffic at the downstreamintersection. The reduction of the green time for thatdirection will be given to other phases by means ofany of the four strategies. At an intersectionupstream from the incident, if blockage exists andreduces the flow, the green split is reduced by thesplit plan change and the early cut-off. SCATS doesnot have logic to prevent intersection blockage.

The British SCOOT system is simply an on-line ver-sion of TRANSYT. The control strategies are tominimize delays and stops at all intersection in thenetwork in all ranges of demand. Signal splitschange incrementally based on current demandobtained from detectors every four seconds. Offsetsand cycle times are adjusted every few minutes. Theresponse to incidents relies on the adaptive logic ofthe system. In the case of an incident where the traf-fic demand change is so rapid that SCOOT cannotadapt to it, two methods can be imposed. The firstmethod is that SCOOT is suspended and falls backto manual operation. The other method is to invokea special plan run. SCOOT also has a gating featurethat limits flow into a particular sensitive area.

2.3 Route guidance and accesscontrol strategies

Van Aerde and Rakha (1989) studied the poten-tial of two route guidance strategies, namely user-optimum and system optimum assignment. Theuser optimum strategy follows Wardrop’s first prin-ciple that a driver is assumed to choose a routewhich will minimize their journey time through the

-

The Model Analysis Report on the Benefits of SCATS In Alleviating the Impacts of lncidentsFAST-TRAC - Phase IIB #20

7

FAST-TRAC

network, and all other drivers equipped with theroute guidance instrument make their choice by thesame criterion. The system optimum strategy is thateach driver with the route guidance equipment isdirected to the path which minimizes the overalltravel time to all drivers. They concluded that sys-tem optimized routings are complex and impracticalfor any but the most trivial networks.

The difficulty of integrating route guidance/signalcontrol was initially reported by Allsop and Charles-worth (1977). They suggested that the route guid-ance optimal strategy was dependent on signalcontrol. In normal traffic operation, an optimal rout-ing can be found for each signal setting. When thesignal timings are changed, the optimal routingchanges. This effect is also reported by Charles-worth (1978) and Maher and Akcelik (1975).

Although optimal route guidance control is difficultto obtain, modem systems utilize advanced commu-nication systems to obtain real-time traffic data fromindividual vehicles. The real-time travel time dataare then used to determine the optimal path. Thesystems are based on the user optimal strategy. Eachequipped vehicle is provided with information on theshortest time to its destination, based on currenttravel time on each road section. The systemsinclude EURO-SCOUT, CACS, SOCRETES (Cas-tling, 1994).

Various techniques can also be used to disseminatetraffic information to motorists. These include radiobroadcasting, police control, variable message signs,and pre-trip planning.

The concept of access control can be applied to inci-dents if a route guidance system is available. In anincident situation, access to the area is controlled bymeans of traffic diversion. Traffic is diverted to otherroutes to avoid the obstruction. For a wider area,where there are numerous alternative routes, accesscontrol can be done through the use of route guid-ance to m-route traffic to non-congested routes.

Traffic metering (gating) is another technique to con-trol access. The traffic is screened at an entry pointto limit the number of vehicles allowed into the con-

gested area. Many countries have applied this tech-nique for peak-hour congestion (May and Westland.1979). Rathi (1991) conducted a comprehensivestudy on the traffic metering on a grid network ofManhattan, New York, which the overall travel timereduction in the order of 20 percent was obtained.

2.4 Effectiveness of control strat-egies for incidents

Hunt and Holland (1985) studied the effect of aSCOOT control strategy and traffic diversion on thereduction in congestion due to an incident on ahypothetical network. A key assumption in thestudy was that the flows were chosen so that duringthe road closure there was no oversaturation underfixed time control. This implies that any increase indelay is due to lack of responsiveness, rather thanlack of capacity. In the “before” case, demandwhich created a degree of saturation of 42 percent atany intersection on the major arterial was selected toconform to the assumption. The volume level wasselected to resemble off-peak traffic in Coventry,England. Each junction had a simple two-phase sig-nal. The control strategies in this study were modifi-cations of green split and offsets as a result of theSCOOT embedded adaptive control logic and trafficdiversion. All traffic was obliged to turn away fromthe incident at an upstream intersection onto one par-allel arterial then return to the original route at anintersection downstream of the incident. The resultsfrom the simulation are shown in Table 2.1.

The results indicate a large incident delay reductionfor adaptive signal control over the fixed time sys-tem. However, the assumptions on the treatment ofright turns on the diverted route (vehicles are oper-ated on the left side of the road in England) are ques-tionable. The assumption was made that there wasno opposing traffic for the right turners and thusthere was no waiting time for this traffic. The paral-lel arterials, which received the diverted traffic, didnot connect to other links and thus the need to retainprogression on the adjacent arterials with the outsidenetwork was not considered.


8

FAST-TRAC

Table 2.1 Delay And Journey Time For Routes Affected By The Incident(Hunt and Holland, 1985)

Delay Journey time

Scenario Route/Control Seconds per veh Increase Seconds per veh Increaseover normal over nor-before case mal before

(%) case (%)

Before incident Direct route 28.4 84.4Diverted route 61.0 115 145.0 72

After incident Fixed time control 105.6 271 189.6 125(diverted route) SCOOT responsive 44.4 56 128.4 52

control

Roberg (1995) investigated several dynamic strate-gies for controlling and dispersing traffic jams on anidealized one-way grid network. The strategiesinclude the application of restricted movements on anumber of critical junctions in the network. An inci-dent was placed on a link in the network to studytraffic congestion evolution and migration. Controlstrategies in this study were a combination of turnbans, ahead bans, and gating. Traffic was notallowed to turn into selected links to avoid gridlock.An ahead ban was imposed around the envelope ofthe congested area to reduce input into the criticalsections of road. The ahead ban forced traffic toreroute away from the jam. In some experiments,instead of forcing vehicles to turn away from thecongested region via ahead ban, vehicles werequeued on the approaches to the jam without beingdiverted. This technique is called gating or externaltraffic metering.

Roberg and Abbess (1994) reported that gating gen-erally yields higher delay than rerouting becausetraffic, which is stored on the approach links, maycreate some secondary gridlock.

The results of the study indicate the operationaldomain, in which the chosen strategies have success-fully eliminated the jam, with respect to differentlevels of demand, turning percentage, and lane utili-zation at the intersection. The increase in delay fromthe normal situation was presented.


9

This study did not have a routing procedure fordiverted traffic and there was no guarantee that vehi-cles returned to their original routes. The study alsoneglected to account for the longer journey time dueto alternate routes and thus the indicated reduction indelay was exaggerated.

Van Vuren and Leonard (1994) studied four scenar-ios of different control strategies on incident conges-tion; signal optimization, gating, diversion, anddemand restraint. They tested these techniques onan irregular shaped network of Kingston-upon-Thames. A congestion index, defined as the ratio ofthe travel time to the free-flow travel time, was usedas a measure of effectiveness. The signal down-stream of the incident was timed so that allapproaches had equal saturation. Two main entrypoints to the incident were metered and gated trafficwas stored in the peripheral region. Drivers werediverted to alternate routes when they faced an unex-pected queue, either primary or secondary, based ontravel time along the links. This means the drivershad no prior knowledge of congestion on thediverted routes. In this study, traffic was forced toreroute at selected intersections.

The results show that the equisaturation policy atintersections downstream from the incident lowersthe congestion index, a surrogate measure of delay,by 14 to 22 percent. Benefits from the gating strat-egy were estimated to be up to 7 percent. The resultsuggests that rerouting before reaching the con-

~

-

_

_~

_-

-

-

gested area has substantial potential benefit. Thishowever depends on the incident location, demandpattern, position of traffic signals in the system. andlink storage capacity. The secondary jam due tostored vehicles was not addressed. It was impliedthat diversion becomes potentially damaging whenthe overall level of recurrent saturation in the net-work increases.

Rakha et. al. (1989) examined the interactionsbetween route guidance and signal control strategieson a freeway with a parallel signalized arterial. Traf-fic demand is assumed to be 1800 vehicles per houron the 4-lane freeway and 150 vehicles per hour onthe 2-lane arterial, in the direction incidents are intro-duced. The spacing between junctions on the arte-rial is approximately 2300 meters. Two types ofincidents were investigated, one was a one-laneblockage on the freeway and the other was a block-

FAST-TRAC

age of 0.75 of a lane on the arterial. The authors didnot consider oversaturation and total blockage dur-ing the analysis. Two signal control strategies wereevaluated, namely an off-line signal optimization byholding the signal setting as it was, and an on-linesignal optimization by adjusting signal timings atperiodic intervals. For the on-line signal optimiza-tion strategy the signal splits and the cycle lengthwere optimized as isolated intersections every threeminutes without considering signal offset.

Table 2.2 displays the results from this study for thearterial incident scenarios. In the no incident situa-tion, the on-line signal optimization reduced theaverage trip travel time over the off-line control byfour percent. However, signal optimizationincreases trip time when a higher percentage of vehi-cles with route guidance capability is assumed. The

Table 2.2 Percentage Average Trip Time Relative To Base Case: Arterial Incident(Rakha et. al., 1988)

Scenario

(Duration ofincident)

No signal optimizationNo incident5 minutes

10 minutes15 minutes

On-line optimizationNo incident5 minutes

10 minutes15 minutes

Percentage of route guidance (%)

0 20 40 60 80 100

100 82 79 79 79 79102 83 80 79 79 79105 85 81 80 79 79109 88 82 81 79 79

96 81 79 79 79 7996 82 82 80 80 8098 87 82 82 80 79

100 89 83 81 81 79

authors gave the reason that the signal optimizer was 2.5 Summaryset to calculate timings based on the non-incidentflow rates rather than the reduced flow rates follow- Most traffic control strategies, developed for

ing the incident. The conclusion of this study was recurrent traffic conditions and employed in current

that the combined effect of this signal optimization control systems, may not be effective in an incident

strategy and route guidance was not positive on arte- situation. Incident congestion requires a responsive

rial incidents because the signal control strategy signal control system that can adapt the signal to

failed to improve traffic performance. changing traffic patterns during the incident. More-over, other alternatives, such as traffic restraints andtraffic diversion, can be applied to alleviate the con-


10

FAST-TRAC

gestion. Limited number of studies were conducted dures were tested, they have not been tested in theto evaluate the effectivenesses of these traffic control same street network. Furthermore, a number of -strategies for incidents. The primary focus of the assumptions were imposed and a detailed analysis ofprevious studies was to estimate the impact of one the impact of these assumptions has not been con- -~control type on incident congestion. Although van- ducted.ous signal control policies, as well as routing proce-

The Model Analysis Report on the Benefits of SCATS in Alleviating the impacts of lncidentsFAST-TRAC - Phase IIB #20

11

FAST-TRAC

3. DESIGN OF STUDYThis study was designed to test different traffic

control strategies under a controlled environmentusing the NETSIM traffic simulation program. Thetesting was based on a hypothetical network with abase demand. The normal traffic operations wereassumed to be similar to a peak traffic period, withthe optimal signal timing plan developed by an off-line calculation. Three types of incidents were con-sidered; a one-lane closure, a two-lane closure, and areduction of the two-lane capacity to 15 percent ofthe original capacity. Three incident durations withvarious control strategies applied to these scenarioswere tested.

3.1 NETSIM simulation modelSimulation is a standard tool of engineers in

studying existing systems and in predicting thebehavior of projected systems (Gerlough, 1965).Traffic simulation provides the mechanism for test-ing theories, modeling concepts, control strategies,and new ideas prior to field demonstration. For thisresearch, given that incidents are rare and unplannedevents, it is impossible to study the effect of trafficcontrol strategies empirically. Therefore, simulationis suitable for this research in that it provides anopportunity to experiment with possible policiesunder predetermined incident and controlled trafficconditions.

The NETSIM traffic simulation model was cho-sen for this study. NETSIM was developed by theFederal Highway Administration for simulating traf-fic operations on a surface street system. This

microscopic model is based on the behavior of indi-vidual vehicles, and the newest version (5.0)includes detailed features on signal and routing con-trols. The model is accepted as an official tool fortraffic analysis. The model has recently been modi-fied to incorporate advanced signal and routing con-trol logic, although these options are not yetavailable for public use.

3.2 Hypothetical network formu-lation

There is a general resistance among transportmodellers to use hypothetical networks for simula-tion work because of the potential for introducingunrealistic features (Van Vuren and Leonard, 1994).Nonetheless, a theoretical network is essential whenthe objective is to test a series of options. includinginteractions among features which have not beenapplied in an existing network. The network can bedesigned to obtain a direct measure of the effect of achange by controlling other characteristics thatmight impact on the designated measure. The net-work thus can be used to obtain performance mea-sures under certain characteristics for several controlstrategies. Moreover, the results of one case candirectly be compared with another because the net-work environment is identical.

A postulated surface street network system shown inFigure 3.1 was constructed for this study. The net-work is grid system, with each intersection one-quarter mile apart.


12

FAST-TRAC

Figure 3.1 Street Network in this Study

Incident Takes Place At Node 111

Streets are all two-way with a left turn pocket of 200ft at each intersection approach. Specification of thenetwork and demand is illustrated in Figure 3.2.

3.3 Traffic demand and initial trafficsignal setting

Traffic volumes were selected to be similar tothe peak-hour traffic condition. The traffic volumesand the associated level of service (LOS) is shown inTable 3.1.

The initial traffic signal settings were determinedfrom the TRANSYT signal optimization program.The program selected signal splits, offsets, and cyclelength which minimizes a performance index. Theperformance index for this study was system-widedelay in the network. These initial signal conditionsrepresent the optimal signal setting for normal trafficcondition using the off-line technique.


13

FAST-TRAC

3.5.2 ATMS onlyTraffic metering was chosen to represent the

ATMS only scenario. It is assumed that signals onthe main arterial are responsive to the queue length.The traffic signal timing at an intersection is modi-fied when there is a possibility that this intersectionwill be blocked in the next cycle. With the signalmodification, traffic is released to approach the inci-dent at the same rate that traffic is discharged at theincident location. This technique ensures that thequeue on the link does not result in intersectionblockage. In the both-lane blockage incident, trafficis prohibited from entering that link. The time toimplement this control is determined by calculatingthe growth of the queue, a special sensor situated atan upstream location on the link to detect the pres-ence of queue, or manual monitoring from surveil-lance cameras.

3.5.3 ATIS onlyTraffic diversion represents the ATIS only sce-

nario. Traffic diversion is a method to reduce thedemand approaching an incident-caused congestedarea by rerouting traffic away from the incident.Diversion is the only demand restriction solution thatcan be applied to traffic already on the street at thetime an incident takes place.

There are many mediums to implement traffic diver-sion. The route information can be disseminated viaa dynamic route guidance system, a changeablemessage sign, or even police regulation. With adynamic route guidance system, the percent of com-pliance to the information is one major issue. How-ever, an assumption of 100 percent compliance isused in this study.

The diversion paths can be determined in manyways. If no information on travel time on possiblediversion routes is available, then traffic can beequally distributed to parallel arterials. The simplestpoint of diversion is at an intersection upstream tothe incident. A more sophisticated method includesthe calculation of dynamic travel time on all possiblepaths and dynamic traffic assignment. However, thismethod requires information on dynamic signal con-trols. To date, there is no known diversion method

that considers the combined effect of the adaptivesignal control and dynamic traffic assignment. Thediversion plan can also be determined from theresults of simulation runs. The percentage of the dis-tribution to adjacent parallel arterials can be set toyield the minimum total travel time.

3.5.4 ATIS/ATMSThe traffic diversion combined with the signal

modification on the diversion route is selected forthis category. While there are several methods to ini-tiate a signal modification, the signal setting basedon degree of saturation is chosen in the initial case.The degree of saturation control variable is a well-accepted method of signal control setting. In theory,Webster used this criterion for his signal split deter-mination. In practice, several urban traffic controlsystems such as SCATS and SCOOT have adoptedthis method in their signal modification.

At each intersection green splits can be determinedindependently according to the local demand. Inmany “reactive” traffic controllers such as theSCATS strategic control level, local traffic patternsin a previous cycle are used to determine the splitsettings. Traffic counts are then translated to thedegrees of saturation (DS). The degree of saturationis the ratio of the actual amount of green time uti-lized by traffic to the ideal (minimum) amount ofgreen time which could serve the same amount oftraffic. The DS can also be viewed as a measure ofunused green time and the minimization of theunused green time means the lowest intersectiondelays.

3.6 Measures of effectiveness(MOEs)

Several MOEs that are appropriate for use incharacterizing the control of incident congestionwere used. To be useful, the MOEs must be under-standable and easy to obtain. Moreover, the MOEsmust reflect the objectives.


16

Indicators related to stops, delays, travel time, andproductivity are used for this incident congestionmanagement. These MOEs can be drawn from gen-eral MOEs suggested by Pignataro et. al. (1978).

The MOEs used in recurrent and nonrecurrent trafficconditions are different due to different desiredobjectives. The incident congestion also requiressome MOEs which may not be the same as recurrentconditions.

The MOEs selected for this study are:

a. Total travel time;

b. Delay time;

c. Queue time;

d. Time to dissipate the congestion; and

e. Duration of spillbacks.

3.6.1 Total travel timeTotal travel time is the sum of the travel time for

all individuals completing their trips. This measureconsiders number of motorists, distance of travel,speed, and delay associated with the travel. In thediversion versus non-diversion cases, this MOE canbe used to determine whether the time saved due tolower delay on the diversion route exceeds theincreased travel time due to the longer distance onthe diversion route. The total travel time may be themost appropriate indicator for a system operator,who seeks the minimum overall system-wide traveltime.

3.6.2 Delay timeDelay is a very good MOE in that it reflects the

traffic operation as perceived by system users. Over-all delay over a period of time reflects the efficiencyof the system without considering the impact onindividuals. Some control strategies such as gatingare designed to sacrifice some traffic movements inorder to yield higher overall productivity and theimpact of this strategy can be reflected in this MOE.This MOE, however, may not be appropriate for

comparing diversion and non-diversion control strat-egies as the total delay does not consider theincreased distance on the diversion routes.

3.6.3 Queue TimeQueue time is similar to delay time except it

considers only the period of time vehicles spend in astanding or moving queue. Normally the queuetime is a proportional to the delay time. However, inNETSIM, queue time is recorded at any timewhereas the delay time is collected only when avehicle departs a link. With this reporting character-istic, the queue time is a good representation of thecongestion currently on the streets.

3.6.4 Time to dissipate congestionThis MOE is useful for describing the effect of

congestion on traffic conditions as a whole. One ofthe objectives of congestion management is to clearthe congestion as soon as possible. This indicatorreflects the impacts of a control strategy on conges-tion duration. However, this MOE is very difficult tomeasure in the field due to the fact that it depends onhow congestion is defined. In the NETSIM model,this MOE is not directly available but a surrogatemeasure can be determined from the increasedqueue time on each link. The queue time as a resultof an application of a control strategy can be com-pared with the queue time under normal traffic con-dition to find the duration of congestion due to theincident and control strategy.

3.6.5 Duration of spillbacks

The duration of spillbacks roughly indicates theduration time of the congestion. When spillbacksoccur following an incident situation, these cause abreakdown in the system. Spillbacks generate inter-section blockages and make the congestion spreadvery quickly. The end of spillback duration in thesystem is a measure of the quickness of the action torelieve congestion.

The Model Analysis Report on the Benefits of SCATS in Allevrating the Impacts of IncidentsFAST-TRAC - Phase IIB #20

17

3.7 Study planThe study scheme for this research was designed

to test the impact of the control strategies under tech-nological scenarios and variations in traffic flows. Aset of control strategies were selected to representthe technological groups and a detailed analysis ofthe effect of these controls on the traffic were per-formed. Subsequently, alternate signal control strat-egies as well as diversion alternatives were includedin the experiment to seek better results underselected incident situations. Sensitivity analyses ofthe base case traffic parameters were conducted todetermine the impact of these controls on changingtraffic characteristics. The study plan is shown inFigure 3.3.

3.8 SummaryThe experiment was designed to test control

strategies on a hypothetical network using theNETSIM program. The base traffic demand repre-sents the peak period, with the approximate overallnetwork LOS C. The normal traffic was operatingwith optimal signal timing obtained from an off-linecalculation. Incidents at mid-block with one-laneclosure, both-lane closure, and two-lane closurewhich reduces the capacity to 15 percent of its origi-nal were introduced in this experiment. Four tech-nological scenarios depicting no-change in trafficcontrol, the deployment of ATMS alone, ATIS alone,and the combined ATMS and ATIS were studied.Traffic metering, traffic diversion, and traffic respon-sive control strategies were designed in the analysis.

The Model Analysis Report on the Benefits of SCATS in Alleviating the impacts of IncidentsFAST-TRAC - Phase IlB #20

18

The Model Analysis Report on the Benefits of SCATS in Alleviating the impacts of Incidents

FAST-TRAC - Phase IlB #20

FAST-TRAC

Figu

re 3

.3 S

tudy

Pla

n

Test

ing

cont

rol s

trat

egie

s in

diff

eren

t tec

hnol

ogic

al s

cena

rios

Do-

noth

ing

ATM

S/A

TIS

sign

al re

spon

sive

/tra

ffic

div

ersi

on

ATM

S al

one

Traf

fic m

eter

ing

ATI

S al

one

Traf

fic d

iver

sion

one-

lane

85%

cap

acity

redu

ctio

nbo

th-l

ane

05 m

in10

min

15 m

in

one-

lane

85%

cap

acity

redu

ctio

nbo

th-l

ane

05 m

in10

min

15 m

in

one-

lane

85%

cap

acity

redu

ctio

nbo

th-l

ane

05 m

in10

min

15 m

in

one-

lane

85%

cap

acity

redu

ctio

nbo

th-l

ane

05 m

in10

min

15 m

in

Sen

sitiv

ity to

dem

and

chan

ge10

-min

ute

both

-lane

clo

sure

Hig

hly-

cong

este

d15

per

cent

incr

ease

Off

-pea

k30

per

cent

dec

reas

ePe

ak p

erio

dN

orm

al

Dem

and

vari

atio

n

The Model Analysis Report on the Benefits of SCATS in Alleviating the impacts of Incidents

FAST-TRAC - Phase IlB #20

FAST-TRAC

Figu

re 3

.3 (c

ont’d

)

ATI

S r

outin

g al

tern

ativ

es10

-min

ute

both

-lane

clo

sure

ATI

SD

iffe

rent

rout

ing

sche

me

At u

pstre

am in

ters

ectio

nEq

ual d

istri

butio

nto

par

alle

l are

teria

ls

At u

pstre

am in

ters

ectio

nEq

ual d

istri

butio

nA

t ups

tream

inte

rsec

tion

Opt

imal

dis

trib

utio

n

ATM

S c

ontr

ol a

ltern

ativ

es10

-min

ute

both

-lane

clo

sure

ATM

S/A

TIS

Traf

fic

resp

onsi

ve c

ontr

ol v

aria

bles

Max

imum

que

ue le

ngth

Deg

ree

of sa

tura

tion

Stop

tim

e

ATM

S/A

TIS

Sign

al c

hang

e co

vera

ge

On

dive

rsio

n ro

utes

At a

ll in

ters

ectio

n

4. SIMULATION ANALYSIS AND RESULTSThe NETSIM model was used to simulate traffic

conditions under several control scenarios. Mea-sures of effectiveness (MOEs) from the programwere collected and compared. The first sectiondescribed traffic conditions when incidents wereintroduced. Then various control strategies weretested on the same incident situations and theirMOEs were compared.

of incidents have a significant difference in theireffect on traffic congestion as measured by traveltime, intersection blockage and congestion dura-tion. They also have different impacts on differenttraffic streams.

4.1 Effect of type and duration ofincident

Table 4.1 shows the effect of different types anddurations of incidents on traffic operations whenthere is no change in traffic control. The three types

4.1.1 Total travel timeAs shown in Table 4.1, the one-lane closure inci-

dent produces no discernible impact on traffic opera-tions at this demand level because traffic on theblocked lane is able to switch to the other lane.

Table 4.1 Total Travel Time In The Network Under Incidents

Total travel time in one hour (veh-hrs)

Incident type Incident duration (minutes)

0 5 10 15(no incident)

One-lane closure 774.3 774.3 774.2 774.3

85% reduction in 774.3 779.2 793.8 841.4capacity

Both-lane closure 774.3 788.3 882.5 1193.3

In the 85 percent reduction of capacity and both-laneblockage situations, total travel time increases expo-nentially as the duration of the incident increases.Traffic passing an incident which reduces the capac-ity at the bottleneck to 15 percent of its originalcapacity must stop and wait in a queue before pro-ceeding past the incident. Traffic proceeding throughthe both-lane closure incident must stop until theincident is cleared.

As expected, traffic experiences longer delay whenthe duration and severity of an incident increase.

4.1.2 Intersection blockageThe increase in total travel time is exacerbated

by spillback (intersection blockage) on approachinglinks and adjacent arterials. The spillbacks and dura-tions are shown in Table 4.2.

In the 85 percent capacity reduction incident situa-tion, an intersection blockage occurred at the nearestupstream intersection at 326 seconds after the inci-dent started. The


~-

21 )--

-

--

-

FAST-TRAC

Table 4.2 Start And End Time Of Intersection Blockage:various incident types and durations

409 n/a n/a n/a 879 108914 n/a n/a n/a 721 98113 n/a n/a n/a 824 898

Note: * The intersections are periodically cleared during the spillback periods.

blockages at this intersection ended in 9, 310, and610 seconds after the initial blockage for 5, 10, and15-minute incidents respectively. Blockage at theprior upstream intersection also occurred in the 15-minute incident situation, starting at 917 secondsafter the incident was introduced and lasting for 215seconds. However, these upstream intersectionswere intermittently clear as traffic was released at theincident location or at the downstream intersection.The total duration of blockage was 8, 118, and 237seconds at the nearest intersection for 5, 10, and 15-minute incidents and 46 seconds at the second near-est intersection for the 15-minute incident situation.

In the 5-minute both-lane closure incident situation,the blockage at the nearest upstream intersectionstarted at 212 seconds after the incident was intro-duced, and lasted for 125 seconds. In the lo-minuteincident situation, the blockage started at the nearestupstream intersection at the same time as in the 5-minute incident situation but lasted for 436 seconds.and the blockage at the prior upstream intersectionbegan at 602 seconds after the incident started andended 123 seconds later. The blockage at the nearestupstream intersection in the 10-minute incident situ-

ation lasted longer than in the 5-minute incident situ-ation because of the presence of the traffic queuewaiting at the entry of the link.

The 15-minute both-lane closure incident causedspillbacks on five road sections, three of which werelocated on the street where the incident occurred andtwo of which were situated on a parallel arterial. Onthe arterial leading to the incident, the blockageswere initiated at 212,602, and 879 seconds after theincident, and lasted for 762,418, and 210 seconds,respectively from the nearest to the furthest intersec-tion. One intersection on a side street (intersection14 in Figure 3.1) was blocked starting 72 1 secondsafter the incident and lasting for 260 seconds. Thissecondary congestion is the result of the blockage atthe nearest upstream intersection to the incident(intersection 10). One intersection on a parallel arte-rial (intersection 13) experienced spillback causedby the blockage at intersection 14. The blockage onthis intersection started 824 seconds after the inci-dent and lasted for 74 seconds.

The Model Analysis Report on the Benefits of SCATS in Alleviating the impacts of IncidentsFAST-TRAC - Phase IIB #20

22

4.1.3 Congestion durationThe time period and lateral extent of congestion

in the network can be determined from an analysisof queue time statistics available from the standardNETSIM outputs. While delay is the primary MOEused in this analysis, the output from the NETSIMmodel is not suitable for using this measure as abasis for tracking the spread of congestion throughthe network. Figure 4.1 demonstrates why delayfrom the standard NETSIM outputs cannot be used.This is because the delay for individual vehicles arecollected when they leave a link. Thus, during anincident that closes a link, the delay would go tozero, even though vehicles are queued on that link.This output attribute results in high delay beingreported after the incident is cleared. As shown inFigure 4.1, the IO-minute both-lane closure incidentwas cleared at 900 seconds after the simulationstarted. The incident caused a blockage at the down-stream intersection and hence no traffic departed thelink between the 902nd and 1020th second aftersimulation. The delay was reported when the spill-back dissipated as shown by the high delay betweenthe 1000th and 1500th second. Travel time fromNETSIM also has the same reporting characteristicas the approach delay.

A more appropriate measure for determining thespread of congestion is stopped delay. The stoppeddelay curve in Figure 4.1 was calculated from queuelengths obtained every four seconds. The queuelength is illustrated in Figure 4.2. The queue timefrom the standard NETSIM output is found to behighly correlated with the queue length and stoppedtime delay and thereby suitable for identifying thecongestion period. It is noted that the shape of thethree curves shown in Figure 4.1 are differentbecause of their definitions.

The effect of incident type and duration on thelength of the network-level congestion periods isshown in Figure 4.3. All incidents started at 300 sec-onds after the beginning of the simulation. In the 85percent capacity reduction incident situation, thetraffic operations returned to normal at time 400,900, and 1400 seconds after the incident wascleared, for the 5, 10, and 15-minute incident respec-tively. With the same order of incident duration, theboth-lane closure incident congestion periods lasted600, 1900, and 3900 seconds after the incidents wereremoved.


23

The spread and duration of congestion in the net-work can be presented as illustrated in Figure 4.4.The queue time of all vehicles in a link is used todetermine the congestion period. The congestedlinks in the network are plotted over time to deter-mine the congestion coverage and duration. In thelo-minute both-lane closure incident situation, inter-section blockages created congestion on both thestreet where the incident occurred and the nearbynorth-south streets. After the incident is cleared andthe spillback dissipated, links downstream from thecongested links receive heavy traffic and becomecongested. This is illustrated in Figure 4.4a.

Figure 4.4b shows the congestion caused by the 15-minute both-lane closure incident. Congestionoccurs on the mainstream approaching the incident

FAST-TRAC

as well as on a parallel arterial. Although the con-gestion starts to dissipate on the upstream linkswithin 5 minutes after the incident is cleared. thespillbacks produce major congestion on approachinglinks as well as downstream links well after the inci-dent is cleared. The congestion in the network lastedfor 80 minutes after the incident was introduced.

4.1.4 Delay on different traffic streamsTo better understand the effects of the various laneblockages, an analysis was made to determine therelative delay to various traffic streams in the net-work. This information is essential in determiningthe control strategies to be used in reducing theimpact of an incident. Table 4.3 shows total traveltime of through traffic on the path

a. 10-minute both-lane closure incident

5 minutes after incident


10 minutes after incident 15 minutes after incident

0 Incident location


Figure 4.4 Affected links and duration of congestion: no control change

27The Model Analysis Report on the Benefits of SCATS in Alleviating the Impacts of IncidentsFAST-TRAC - Phase IIB #20

FAST-TRAC




b. 15-minute both-lane closure incident



Figure 4.4 (cont’d)



[] Incident location

Table 4.3 Total Travel Time Of Traffic Passing The Incident(traffic heading from intersection 409 to intersection 12)


Incident type

One-lane closure

85% reduction in capacity

Both-lane closure

Incident duration (minutes)

0 5 10 15(no incident)

62.4 62.4 62.3 62.4

62.4 64.9 77.3 102.7

62.4 78.7 115.7 152.8

approaching the incident. The results reveal the were some differences in the impact on this trafficsame trend as Table 4.1 with travel time increasing stream versus the impact on the entire network.as the incident duration increases. However, there Table 4.3 indicates that the journey time on this

The Model Analysis Report on the Benefits of SCATS in Alleviating the Impacts of Incidents 28FAST-TRAC - Phase IIB #20

FAST-TRAC

through route in the IO-minute 85 percent capacityreduction incident situation is lower than the journeytime for a 5-minute both-lane closure incident (77.3vehicle-hours compared with 78.7 vehicle-hours).However, the total travel time of all traffic in the 10-minute 85 percent capacity reduction incident situa-tion (793.8 vehicle-hours) is higher than that of the5-minute both-lane closure incident (788.3 vehicle-hours). This means that the 10-minute 85 percentcapacity reduction incident causes greater trafficinterruption in non-affected traffic streams than the5-minute both-lane closure incident, although thespillback duration in the first situation is shorter.

As shown in Table 4.3, the 15-minute both-lane clo-sure incident creates widespread congestion in thenetwork. For this incident, the increase in travel timeon the path passing the incident contributes a smallerproportion of the overall network travel time than inthe other cases.

4.2 Effect of different control strat-egies

After determining the effect of an incident ontraffic with no change in traffic control, differenttypes of control strategies were tested and theimpacts were measured. This simulation experimentallows the testing of these control strategies alone aswell as the combination of two or more. controls. In

this study, the impact of the control strategies inthree different ITS technological groups were deter-mined. The experiments were performed on severalincident types and durations.

The 85 percent capacity reduction and both-lane clo-sure incidents were selected as the base cases fordetermining the effect of different control strategiesbecause they cover a wide range of congestion levelsand offer the opportunity to experiment with variouscontrols. Since one of the objectives of this study isto compare alternative control strategies without traf-fic diversion, with diversion alone, and a combina-tion of diversion and traffic control, total travel timeis a good indicator of performance because itincludes the effect of longer route distances resultingfrom diversion.

-I

I

I-

-

i

’-

4.2.1 Total travel timeFour major traffic control scenarios were tested

and compared. The total travel time of all links inthe network for these different scenarios is shown inTable 4.4.

I_

Traffic metering in the partial lane closure incidentsituation consists of gating at the intersection imme-diately upstream from the incident. Traffic isreleased to approach the incident at a rate equal to 15percent of the link capacity so that no growing queue

Table 4.4 Total Travel Time For Four Control Strategy Scenarios:lo-minute both-lane closure incident

Scenario Incident duration (minutes)

5 15

No traffic control change

ATMS: traffic metering

ATIS: traffic diversion

ATIS/ATMS: traffic diversion withsignal change (equalization of degreeof saturation)

779.30*

78 1.79

783.56

789.19

10

793.80

792.99*

803.99

804.76

841.40

821.74

820.00*

823.27


c.-

The Model Analysts Report on the Benefits of SCATS in Alleviating the Impacts of lncidentsFAST-TRAC - Phase IIB #20

29

FAST-TRAC

u

b. Both-lane closure incident

Scenario Incident duration (minutes) I5 10 I 15 I




ATIS/ATMS: traffic diversionwith signal change (equalizationof degree of saturation)

788.30 882.50 1193.30

787.55” 847.68 942.20

788.45 826.01” 869.76”

794.96 839.75 931.83


Note: * indicates the lowest total travel time for an incident situation

developed. In the total-lane closure incident, trafficmetering is set at a rate which keeps the intersectionfrom being blocked by a queue. The green phase isskipped when the queue stored on any receiving linkis full.

ATIS traffic diversion is the assignment of traffic toadjacent arterials to avoid the incident. Traffic is dis-tributed equally between two adjacent parallel routesat the nearest upstream intersection to the incident.The middle through lane can be used by both leftand right-turn traffic. No signal adjustment is madein this scenario.

In the ATIS/ATMS scenario, signals along the diver-sion routes are modified every cycle, using trafficdata of the previous cycle. The signal controlchange in the ATIS/ATMS scenario attempts toequalize degrees of saturation of all approaches ateach intersection. This logic is similar to a SCATSnetwork with no system level control. No offsetchange was made in these initial simulation runs.

The results from Table 4.4a indicate that trafficmetering control yields longer total travel time thanthe do-nothing condition in the 5-minute incident sit-uation. The increase in travel time is caused by thefact that gated traffic has to wait at the upstream link

for approximately one cycle (100 seconds), whilethe spillback occurs for only 9 seconds (Table 4.2).In this situation, it is better to let the spillback occursfor a short time period. Traffic metering in the 10 and15-minute incident situations results in an improve-ment over the do-nothing cases.

The traffic metering strategy in the 5 and lo-minuteincident situation gives shorter travel time than traf-fic diversion. This implies that the longer travel dis-tance caused by the diversion is greater than the timewaiting in the queue at the upstream link. In the 15-minute incident situation, the traffic metering givesslightly higher travel time than traffic diversionalone.

In the 85 percent capacity reduction incident cases,traffic diversion in the 5 and lo-minute incident situ-ations does not improve the overall travel time. Thisis because the diverted traffic has a greater increasein total time than the delay savings to the trafficimpacted by the incident if they had remained ontheir original travel path in these incident situations.Since traffic can be partially released at the incidentlocation, the system performs better if vehicles waitto get through the incident than if they reroute toadjacent streets. Although the 5 and lo-minute inci-dents create periodic blockage at the upstream inter-


30

FAST-TRAC

section(s), spillbacks are relatively short. Thererouting produces traffic disruption at other inter-sections, creating higher delays to traffic on thediversion paths, and on overall network perfor-mance. In the 15-minute incident, however, diver-sion leads to lower network travel time as theblockage and waiting time to pass the incidentadversely affect the network operation.

Traffic diversion alone yields the lowest travel timefor the 15-minute incident. The addition of ATMScontrol increases total travel time slightly over thetraffic diversion only case.

The ATIS/ATMS control creates higher total net-work travel time than traffic diversion alone for allthree incident durations. In fact, in the 5 and 10-minute incident situations, changing signal timingcauses higher delay than the no-control change sce-narios.

For the both-lane closure incident (Table 4.4b), traf-fic metering improves the traffic operations over theno control change situation. The level of improve-ment increases as the duration of the incidentincreases.

Traffic diversion does not lower total travel time forthe 5-minute incident situation. However, for the 10-minute and 15-minute incident, this controldecreases the total travel time. The reduction is 27percent from the no-control change situation for the15-minute incident since the diversion eliminates allintersection blockages.

The addition of ATMS to the ATIS does not reducethe total travel time for any of the incident durations.

4.2.2 Intersection blockageAll traffic control strategies eliminated intersec-

tion blockage. The traffic metering at the upstreamintersection cut off the green light of the congesteddirection before the back of the queue reached. Inthe diversion situation, traffic was re-routed beforethe queue blocked the upstream intersection.

4.2.3 Congestion durationThe duration of congestion is another measure

used to express the impacts of these control strate-gies. As discussed in Chapter 3, the queue time wasan appropriate measure to identify the beginning andthe end of the congestion period for the no controlchange scenarios. However, because there isincreased travel distance in the diversion controlstrategies, queue time may be the best measure todetermine the congestion period, but the total traveltime is a measure of traffic efficiency for these con-trol strategies.

The 10 and 15-minute both-lane closure incidentswere selected for this analysis. The effect of differ-ent control strategies on congestion duration isshown in Figure 4.5. In the lo-minute incident, theoverall congestion ends 1900 seconds after the inci-dent is cleared if there is no control change. Trafficmetering causes longer congestion duration than theno control change situation by 200 seconds. This isbecause traffic metering, although it successfullyeliminates the spillback, stores traffic in the networkto postpone the surge of heavy congestion. It spreadsthe congestion peak but does not reduce the conges-tion duration. On the other hand. the ATIS alone andthe ATIS/ATMS control strategies alleviate the con-gestion 800 seconds sooner than the do-nothingcase.

In the 15-minute incident, the no control change cre-ates congestion until 3500 seconds after the incidentis removed. The traffic metering shortens the con-gestion duration over the do-nothing case by 700seconds. The ATIS/ATMS shortens the congestionperiod by 1600 seconds over the no control changecase. Diversion alone eliminates the congestionperiod 500 seconds sooner than the ATIS/ATMS, or2100 seconds sooner than the do-nothing case.

The extent of the congestion is shown in Figure 4.6.The IO-minute both-lane closure is chosen in theanalysis. The traffic metering strategy producesheavy traffic on the link approaching the incident, asthese links are designated to store the spillback (Fig-ure 4.6a). Traffic diversion creates heavier traffic ondiversion routes during the rerouting, but the dura-


31

i

I

r

a.

Traf

fic

met

erin

g

5 m

inut

es a

fter

inci

dent

20 m

inut

es a

fter

inci

dent

35 m

inut

es a

fter

inci

dent

10 m

inut

es a

fter

inci

dent

25 m

inut

es a

fter

inci

dent

40 m

inut

es a

fter

inci

dent

15

min

utes

afte

r in

cide

nt

30 m

inut

es a

fter

inci

dent

45 m

inut

es a

fter

inci

dent

[] In

cide

nt l

oca

tio

n

Figu

re 4

.6 A

ffect

ed li

nks

and

dura

tion

of c

onge

stion

: di

ffere

nt c

ontro

l stra

tegi

es

4.2.4 Delay on different traffic streamsAn analysis of the impact of these control strate-

gies on different traffic groups was performed sinceeach control strategy provides different treatments totraffic groups. It is possible that a control treatmentwould give an advantage to a specific traffic groupand sacrifice others, and total travel time alone can-not distinguish the impact on different traffic groups.Therefore, the data for MOEs of specific trafficgroups were obtained and analyzed.

The traffic was divided into three groups, based onthe potential for different impacts resulting from thecontrol strategies. They were:

a. Rerouted traffic;

b. Traffic competing with the rerouted traffic ondiversion route(s); and

FAST-TRAC

c. Traffic at other intersections.

MOEs for these traffic groups were measured todetermine the impact of ATIS/ATMS and ATIS onlycontrol strategies. Table 4.5 displays the results ofthis analysis. As expected. the delay in making leftturns for the diverted traffic is reduced under ATMScontrol as the queue of these vehicles increased thedegree of saturation for this movement, resulting inan increased allocation of green time. However. thedelay to traffic in other directions, which are compet-ing with the diverted traffic, increases as a result ofthis reallocation of green times. Traffic in other loca-tions also suffer higher delay due to the interruptionof progression caused by the signal adaptation. Thisinterruption of traffic in other directions is the maincontributor to longer overall travel time in the ATIS/ATMS scenario than in the ATIS alone scenario.

Table 4.5 Total Travel Time Of Particular Traffic Groups: Both-lane Closure Incident

Link Group

Links on diversionroutes where trafficmakes

Left turnRight turn

All links approachingintersections at whichtraffic is diverted

All other links

Total 788.45 794.96 826.0 1 839.75 869.76 931.83


Diver-sion only

9.03 8.93 18.82 15.33 38.29 24.7 17.85 8.42 11.33 11.55 17.55 19.69

342.17 347.62 377.53 379.59 376.55 447.67

429.40 429.98 418.33 433.28 437.37 439.76

5

With signalchange ondiversion

routes

Incident duration (minutes)

Diver-sion only

10


routes

15

Diver-sion only


routes


37

FAST-TRAC

4.3 SummaryWith the initial traffic and network condition in

this study, the one-lane closure did not affect trafficperformance because the level of the demand waswell below the reduced capacity at the incident loca-tion. However, the 85 percent capacity reductionand the both-lane closure did affect traffic opera-tions. The longer the incident duration and the moresevere the incident, the more impact on the networktravel time. The traffic stream passing the incidenthad the greatest increase in delay, although otherdirections were delayed by the spread of congestion.

Control strategies tested in this research had differenteffects on the incident-based congestion. Trafficmetering reduced travel time in the both-lane closureincident situation and the longer duration of the 85percent capacity reduction scenario. However, thiscontrol did not have a major impact on the reductionin the length of the congestion period.

Traffic diversion did not improve the traffic opera-opera-tion in the incident situations which had short dura-tion (5 and 10 minutes in the 85 percent reduction,and 5 minutes in the both-lane closure). Trafficdiversion was effective in the 15-minute partial laneclosure, and the 10 and 15-minute both lane closurescenarios. It reduced the network total travel timedespite the fact that some vehicles had to travellonger distances. Diversion shortened the conges-tion period compared to the do-nothing case.

The traffic diversion with signal change along thediversion routes was not effective in the 5 and 10-minute 85 percent capacity reduction, and the 5-

. . .minute both lane closure incident situation.Although this control tended to favor the divertingtraffic and resulted in an improvement in total traveltime over the no control change scenario in otherscenarios, it did not provide congestion reductionbeyond the traffic diversion alone strategy.


38

FAST-TRAC

5. SENSITIVITY ANALYSISThe analyses in the previous chapter were based

on a common set of traffic parameters and a speci-fied set of control strategies. In this chapter, theimpact of the control strategies under different trafficconditions and control concepts were investigated.The traffic volumes were altered to depict higherdemand, as might be found in the peak period, aswell as a low demand scenario representing the off-peak period. Alternate diversion schemes weredeveloped and examined. Different treatments forsignal timing, including the use of different variables(other than degree of saturation) as a basic for modi-fying the phase plan, were considered. Finally, thenumber of signals considered in the ATMS strategieswere varied.

5.1 Sensitivity to traffic demandchange

To determine the effectiveness of the controlstrategies under diverse traffic demand conditions,two traffic volumes were introduced to represent thetraffic demand in a highly-congested peak periodand an off-peak period. The new traffic volumes andlevels of service (LOS) are shown in Table 5.1. A 15percent increase in traffic volume was used to repre-sent a higher traffic demand and more congested net-work condition. A 30 percent decrease in trafficvolume was selected to represent an off-peak trafficcondition.

Table 5.1 New Traffic Demand Used In The Sensitivity Analysisa) Low demand--off-peak period

Traffic direction

East Bound (EB)- Thru and Right turn- Left turn

West Bound (WB)- Thru and Bight turn- Left turn

North Bound (NB)- Thru and Right turn- Left turn

South Bound (SB)- Thru and Right turn- Left turn

Traffic volume(vph)

945105

47253

63070

31535


Level ofservice*

(intersection 10)

AE

CD

B’D

BE

--

-

-

-

-

39 -~

FAST-TRAC

b) High demand--highly-saturated system

Traffic direction Traffic volume(vph)

East Bound (EB)- Thru and Right turn- Left turn

West Bound (WB)- Thru and Right turn- Left turn

North Bound (NB)- Thru and Right turn- Left turn

South Bound (SB)- Thru and Right turn- Left turn

1553 F172 E

776 B86 D

1035115

51758

Level ofservice*

(intersection 10)

DE

CC

Note: Levels of service are determined at optimal progression based on Synchro 2 . 0 T M program (1995) andHighway Capacity Manual (1994)

For each demand case, signal timings were recalcu-lated using the TRANSYT program. The optimalsignal timings were then used for normal signaloperations.

The new demand levels were input into theNETSIM model. A lo-minute both-lane closureincident was selected for this analysis.

5.1.1 Impact on total travel timeThe impacts on travel time of new volumes are

shown in Table 5.2. In all traffic volume conditionswith the lo-minute both-lane closure incident, trafficmetering reduces the travel time over the no trafficcontrol change. The metering is more effective inthe high demand situation, reducing 56 percent ofthe increased travel time due to incident under heavytraffic, compared to 36 percent in the low demandsituation.


40

Table 5.2 Total travel time for three volume scenarios:lo-minute both-lane closure incident

Scenario

Normal traffic (no incident)




ATIS/ATMS: traffic diversion withsignal change (equalization ofdegree of saturation)


Volume scenarios

30% decrease Normal 15% increase(off-peak) (peak) (highly-congested)

486.79 774.30 1005.10

523.29 882.50 1287.95

510.24 847.68 1128.66

502.36” 826.01* 1115.09

505.84 839.75 1076.09*

Note: * indicates the lowest total travel time for an incident situation

Traffic diversion alone also improves traffic opera-tions over the do-nothing situation in all demandranges. However, in the off-peak and normal peakperiod, the ATIS/ATMS strategy does not increasethe benefits. In fact, modification of signal timingincreases the travel time by 3.48 vehicle-hours in theoff-peak demand case. In the highly congested traf-fic condition, the addition of traffic signal controlmodification reduces the travel time over diversiononly, from 1115 vehicle-hours to 1076 vehicle-hours.

The results shown in Table 5.2 suggest that signaltiming modification along with the diversion isdesirable in high traffic demand conditions. Thismay be because the high volume of rerouted trafficcontributes to longer overall delay in the diversiononly situation. The benefits of signal timing modifi-cation favoring the rerouted traffic exceeds theincrease in delay incurred on other directions,thereby reducing the overall delay in this scenario.

5.1.2 Impact on intersection blockageThe three traffic demand levels created diverse

results when measured by intersection blockages.The results were shown in Table 5.3.

Spillback occurs at all demand levels for the no con-trol change scenarios. The traffic volume determinesthe start time and duration of the spillback( Thehigher the volume, the sooner the blockage begins.The start time of the blockage at intersection 10starts at 248,2 12, and 208 seconds after the incidentoccurs for off-peak, peak, and highly-congested traf-fic conditions, respectively. The incident in thehighly-congested condition creates intersectionblockages on the north-south arterial, as the result ofthe spillback on the arterial on which incidentoccurs. All control strategies successfully eliminatethe intersection blockages.


41

-

FAST-TRAC

Table 5.3 Start And End Time Of Intersection Blockages:various traffic demand levels, 10-minute both-lane closure incident

Time of spillback occurrence after incident (sec)

Intersectionnumber

Traffic demand level

Off-peak Normal peak Highly-congest-ed

109

1413

Begin End Begin End Begin End

248 634 212 648 208 631n/a n/a 602 725 554 1015n/a n/a n/a n/a 497 710n/a n/a n/a n/a 1075 1086

Note: The intersections are periodically cleared during the spillback periods.

5.1.3 Impact on congestion durationThe congestion durations as determined from

the queue time are illustrated in Figure 5.1. Controlstrategies in each of the traffic demand ranges havesimilar potential in reducing the congestion duration.The traffic metering strategy does not have mucheffect on congestion duration over the do-nothingcases. Two diversion options, with and without sig-nal timing modification, have similar impacts. In the

highly-congested traffic condition, however, diver-sion with signal modification clears the congestion300 seconds sooner than the diversion alone.

A comparison of the reduction in the congestionperiod for various traffic demand levels is shown inFigure 5.2. The congestion duration in the highly-congested condition is projected from Figure 5.1 (c)because the simulation approached the NETSIMmaximum number of vehicles in the system. Figure5.2 shows that the application of control


42

Off-peak Peak Highly-congested

Traffic demand scenario

Figure 5.2 Comparison of reduction in congestion duration in different demand levels

strategies is more effective in reducing the conges-tion period at higher demand levels. The ATIS/ATMS control reduces the congestion period to lessthan a half of the time required under the no controlchange scenario.

5.2 Sensitivity to diversion controlalternatives

The analyses in the preceding chapter werebased on diverting traffic equally onto the two adja-cent parallel arterials at the nearest upstream inter-section. Two new diversion plans were developedand compared in this section. Three diversion alter-natives considered here were:

a. Traffic diversion at the nearest upstream inter-section, equally distributed to two adjacentarterials (the alternative plan from Chapter4);

b. Traffic diversion at the nearest intersection,optimal distribution to the arterials; and

c. Traffic diversion at the two nearest intersec-tions, equally-distributed.

Under alternative b, the diversion percentage is cho-sen to yield the lowest network travel time. This dis-tribution was obtained by trial-and-error, becauseNETSIM does not have internal logic for dynamictraffic assignment. The optimal distribution sent 56percent of eastbound through traffic at intersection10 to the north arterial (passing through intersections10-6-7-l 1) and 44 percent of the traffic to the southarterial (passing through intersections 10- 14- 15-11). The last alternative was to divert half of the traf-fic at each of the two nearest upstream intersec-tions. At each intersection the traffic was equallydistributed to the two parallel arterials.

5.2.1 Impact on total travel timeThe total travel time employing these diversion

plans is shown in Table 5.4. The optimal distributionplan gives the shortest travel time, although thereduction is only 5.18 vehicle-hours or 0.6 percent ofthe total travel time from the equal distribution diver-

The Model Analysis Report on the Benefits of SCATS in Alleviating the Impacts of Incidents 46FAST-TWC - Phase IIB #20

FAST-TRAC

sion plan. The diversion at two upstream intersec- 5.2.2 Impact on congestion durationtions gives only a slight improvement over the equal Figure 5.3 shows the congestion duration of thediversion at the nearest intersection. three diversion procedures. All three diversion plans

create similar incident durations. However, the two-intersection diversion has 200 seconds longer con-gestion than each of the one-intersection diversions.

Table 5.4 Total Travel Time Of Alternate Diversion Plans10-minute both-lane closure incident

Diversion plan Total travel time in one hour(veh-hrs)

Equally distributed at thenearest upstream intersection

Optimal distribution at thenearest upstream intersection

Equally distributed at twoupstream intersections

826.45

821.27

825.20

In contrast with the congestion duration, the spreadof the congestion on the network was significantlydifferent. The effect of congestion spread is illus-trated in Figure 5.4. The optimal distribution diver-sion results in a slightly longer congestion period onthe north parallel arterial, to which more traffic isassigned, and a slightly shorter congestion period onthe diversion route passing the south parallel arterial.The two-intersection diversion causes congestion atthe second upstream intersection. The congestion atthe nearest upstream intersection clears sooner than

the one-intersection rerouting since less traffic isapproaching this intersection (half of the throughtraffic is diverted at the prior intersection).

5.2.3 Impact on different traffic streamsTo understand the impact of the control on dif-

ferent traffic movements, traffic was arranged intothree groups. The first group was the traffic on thediversion routes including the paths passing the sec-ond nearest intersection. The second group was thetraffic competing with the diversion traffic, at theintersections on which the diverted

The Model Analysis Report on the Benefits of SCATS in Alleviating the Impacts of IncidentsFAST-TRIG - Phase IIB #20

47

FAST-TRAC

traffic traversed. The last group was all other traffic.The effect of the alternate diversion plans on differ-ent traffic groups is shown in Table 5.5.

Comparing the alternate rerouting plans at the near-est intersections, the optimal distribution plan resultsin lower travel time on the diversion routes. More-over, the optimal plan produces the minimal inter-ruption to other traffic directions competing with thediversion routes. By sending more traffic to thenorth arterial, the rerouting produces less impacts ontraffic in other directions as traffic diverting to thenorth arterial makes two right turns, compared withtwo left turns on the south arterial diversion path.

Comparing the equally-distributed diversion at thenearest intersection and the diversion at twoupstream intersections, the latter plan produces

lower travel time (delay) on the diversion routes.The travel time of diverted traffic is shorter by 3.58vehicle-hours in the two-intersection diversioncase. The two intersection diversion results in lowercongestion at each intersection. Although the two-intersection diversion creates higher delay in non-diverted traffic, the overall travel time (delay) isshorter by 1.65 vehicle-hours.

5.3 Sensitiity to signal controlvariable alternatives

In the previous analysis, the equalization ofdegree of saturation was used as a criterion for signalmodification. In this section, two other control vari-ables were introduced and tested. The controlparameters were maximum queue length and stoptime.

Table 5.5 Total Travel Time Of Traffic Streams In Alternate Diversion Plans:lo-minute both-lane closure incident

-


Link Group Diversion plan

Equally distributed at Optimal distribu- Equally distributedthe nearest upstream tion at the nearest at two upstream

intersection upstream intersec- intersectionstion

Links on diversion routes’ wheretraffic makes

Left turnThroughRight turn

All links approaching intersectionsat which traffic is diverted2 (exceptthe diverted traffic direction)

25.57 24.5 1 23.0538.19 38.01 37.8815.24 16.24 14.49

513.62 507.08 515.27

All other links3

Total

233.83 235.43 234.5 1

826.45 821.27 825.20LNote:

1. This traffic group contains:

intersection 5, NB-RT

intersection 6, EB-THRU, NB-RT

intersection 7, EB-RT

The Model Analysis Report on the Benefits of SCATS in Alleviating the impacts of IncidentsFAST-TRAC - Phase IIB #20

52

FAST-TRAC

intersection 9, EB-LT, EB-RT

intersection 10, EB-LT, EB-RT

intersection 11, SB-LT, NB-RT

intersection 13, SB-LT

intersection 14, EB-THRU, SB-LT

intersection 15. EB-LT

2. This traffic group contains all traffic direc-tions at intersection 5, 6, 7, 9, 10, 11, 13, 14,15 excluding diverted traffic movementsabove

3. This traffic group contains all traffic exceptthe two traffic groups above.

The maximum queue length in each traffic move-ment is a good indicator of traffic demand at theintersection. The signal setting strategy used in thiscase is to balance the maximum queue lengths in allapproaches, which leads to the postponement of

spillback and the elimination of secondary conges-tion. The spillback can be either left-turn queueoverflow to the through traffic lane or the intersec-tion blockage at an upstream intersection.

The maximum queue length in each cycle is not pro-vided in the NETSIM standard output. However, itcan obtained by collecting queue data every fourseconds from the NETSIM intermediate statistics.

The stop time delay is another possible control vari-able available from NETSIM. The equalization ofstop time delay in all competing approaches impliesequality of the level-of-service (LOS). as defined inthe Highway Capacity Manual. in all directions.

5.3.1 Impact on total travel time

Table 5.6 shows the total travel time for the threesignal control alternatives. The signal setting byequalizing of the stop time yields the lowest totaltravel time. It improves this MOE by 2.4 percent of

Table 5.6 Total Travel Time Of Alternate Signal Control Variables:10-minute both-lane closure incident

Signal control variable

Degree of Saturation

Maximum Queue length

Stop time

the total travel time beyond the setting by degree ofsaturation for the lo-minute both-lane closure sce-nario.

5.3.2 Impact on congestion durationThe congestion duration for each of these three

signal control settings is shown in Figure 5.5. Thesignal settings using the degree of saturation andstop time alleviate the congestion 300 secondssooner than that using the maximum queue length.The control based on the stop time, however, pro-duces the least time in queue during and after theincident, as reflected in the minimum total traveltime measure.

Total travel time in one hour(veh-hrs)

5.3.3 Impact on different traffic streamsTable 5.7 shows the effect of these signal set-

tings on different traffic groups. Among these threecontrols, the equalization of stop time concept is thebest because it reduces travel times at intersectionson the diversion routes over the degree of saturationand the maximum queue length criteria. The reduc-tion in travel time on diversion routes is greater thanthe increase on other links, resulting in the loweroverall travel time.


53 -

Table 5.7 Total Travel Time Of Traffic Groups Of Alternate Signal Control Variables:1 O-minute both-lane closure incident

Link Group

Links on diversion routeswhere traffic makes

Left turnRight turn

All links approachingintersections at which traf-fic is diverted


Signal control variable

Degree of Maximum Equalization ofSaturation Queue Stop time

Length

15.33 14.46 13.4411.26 10.26 9.98

379.88 374.44 360.72

All other links 433.28 433.87 435.39

Total 839.75 833.33 8 19.53

5.4.1 Impact on total travel time 5.4.2 Impact on congestion durationTable 5.8 shows the total travel time for two sce- Figure 5.6 shows the congestion durations of

narios with different signal coverage. The modifica-tion of signal timings at all intersections results in

these two signal coverage alternatives. Although the

lower total travel time. By making the signals atsignal modification at all intersections produces

these intersections more responsive to the traffic, itslightly lower delay (queue time) than the signalchange on diversion route case, the network recovers

results in better progression and lower total travel from congestion at the same time.time.

Table 5.8 Total travel time of alternate signal control coveragelo-minute both-lane closure incident

Signal control coverage Total travel time in one hour (veh-hrs)

On diversion routes 839.75

All intersections 826.48

5.4.3 Impact on different traffic streams streams reveals that the benefit comes from theThe preceding section shows that the larger cov- lower delay at competing approaches on the diver-

erage area of signal modification yields better traffic sion routes. The results are shown in Table 5.9. Onoperation, as indicated by shorter overall travel time. the diversion route, the signal modification at allThe analysis of travel time on particular traffic intersections slightly reduces the travel time in the


55

diverted traffic directions. However, most of the progression on these parallel arterials. The signaldelay reduction occurs on traffic in other directions. modification at all intersections reduces the delay ofe.g. through traffic on the parallel arterials. This is competing movements to diverted traffic frombecause the adaptation of signals at intersections 379.88 vehicle-hours to 361.74 vehicle-hours,upstream and downstream from the intersections although it increases the travel on other links fromreceiving diversion traffic results in more response to 433.28 vehicle-hours to 438.43 vehicle-hours.the cycle-by cycle traffic situation, leading to better

Table 5.9 Total Travel Time Of Traffic Groups Of Alternate Signal Control Coverage10-minute both-lane closure incident

Link Group


Diversion only With signal change Withon diversion routes signal change at all

intersections

Links on diversion routeswhere traffic makes

LefttumRight turn

All links approaching inter-sections at which traffic isdiverted

All other links

Total

18.82 15.33 15.0511.33 11.26 11.26

377.93 379.88 361.74

418.33 433.28 438.43

826.01 839.75 826.48

-


56

FAST-TRAC

5.5 SummaryThe control strategies used in Chapter 4 were

tested to determine their effectivenesses under differ-ent demand conditions. The "traffic metering” and“diversion only” alternatives were based on the opti-mal fixed time signal plans for each demand level.The impact of these control strategies under trafficdemand representing off-peak, peak, and highly-congested conditions were obtained for the 10-minute both-lane closure incident situation. Thetraffic metering produces lower total travel time(delay) than the no control change in each demandrange. However, the traffic metering does not havemuch impact on the congestion duration. Trafficdiversion is more effective than traffic metering, anddiversion with signal control change is the mosteffective at the high demand level. For both diver-sion with and without signal modification, the reduc-tion in congestion period increases as the trafficdemand increases.

Three different diversion and signal control settingalternatives were analyzed. The optimal distributionat the nearest upstream intersection yields the lowestdelay. The distribution at two upstream intersections

creates less congestion for each intersection but in awider area. The duration of overall congestion in thethree diversion plans lasts approximately the same.The optimal distribution produces minimum inter-ruption to traffic competing with the diverted trafficdirections.

Signal control modifications using three control vari-ables were tested. The signal setting using the stoptime produces the lowest total travel time, although itdoes not shorten the congestion period beyond thesetting using degree of saturation. The reduction indelay occurs at intersections along diversion routes.Analysis on signal control coverage indicates thatthe signal modification at all intersections results inlower delay than the signal change on diversionroutes since it increases progression on the parallelarterials.


58

6. CONCLUSION AND RECOMMENDATION ~_

6.1 ConclusionThis research examined the effectivenesses of

several control strategies on various incident condi-tions. The aim of these controls was to alleviate thecongestion caused by these incidents. In this study,the control strategies considered were traffic meter-ing, traffic diversion, and signal timing modification.The research was conducted using the NETSIMsimulation program. A hypothetical network of asurface street system was used. Using the same plat-form, several incident situations as well as controlstrategies were tested. The measures of effective-ness used to identify the performance of each controlstrategy were total travel time (delay and queuetime), congestion duration, and spillback duration.

The impacts of three incident types, each with threeincident durations, were studied under the assump-tion of no control change in traffic control system.The investigation showed that, in the severe incidentconditions, the congestion lasted up to 68 minutesand produced 419 vehicle-hours of delay if no spe-cial control strategy was applied.

The impact of the control strategies on the variousincident situations was analyzed. It was found thatdifferent control strategies had different levels ofeffectivenesses in specific incident conditions.These results are shown in Table 6.1. In a lesssevere incident situation, such as the one-lane clo-sure and the 5-minute 85 percent reduction in capac-ity, none of the control strategies offered anyimprovement in traffic operations over the do-noth-ing case. In the IO-minute partial-lane and 5-minuteboth-lane closure incidents, only the traffic meteringstrategy reduced network travel time. In the 10-

minute partial-lane closure situation, traffic meteringdid not have a major impact on the congestionperiod.

Traffic diversion was effective when the severity ofthe incident increased (15-minute 85 percent capac-ity reduction, and 10 and 15-minute both-lane clo-sure). This control strategy substantially reduced thelength of the overall congestion period. Althoughdiversion results in an increase in congestion on thediversion routes for a short time period, it reducesthe congestion on the affected traffic and the overallnetwork. The diversion with signal timing modifica-tion strategy did not offer any improvement overtraffic diversion alone.

The limits of effectiveness of these control strategieswere examined by conducting a sensitivity analysison their effectiveness at different demand levels.The control strategies were tested using volumesrepresenting off-peak, peak, and highly-congestedtraffic conditions. The results indicated that, whenthe demand level increases, the control strategies aremore effective in reducing both total travel time andcongestion duration.


59 -

FAST-TRAC

-

Note: 1. the negative signs indicate the improvement over the no control change

2. bold numbers indicate the situations where the control strategies are effective

3. the * means this control strategy produces the greatest improvement in that

incident type and duration.

Table 6.1 Impact Of Alternative Control Strategies On Various Incidentsa. Difference in total travel time from the no control change scenarios (percent)

Incident type and duration

Control strategy 85 percent capacity reduction Both-lane closure

5 minutes 10 minutes 15 minutes 5 minutes 10 minutes 15 minutes

ATMS: traffic metering + 50” - 4” -29 - 6* - 32 - 60

ATIS: traffic diversion + 85 +52 - 32” + 1 - 52” - 77”

ATIS/ATMS: traffic +198 +56 -27 +48 -40 - 62diversion with signaltiming modification

b. Difference in congestion duration from the no control change scenarios (percent)

(both-lane closure incident)

Control strategy Incident duration

10 minutes 15 minutes

ATMS: traffic metering + 8 - 16

ATIS: traffic diversion - 32” - 36”

ATIS/ATMS: traffic diversion with - 32* -25signal timing modification

Moreover, in the highly-congested condition, diver- For each control strategy, variations in the tech-sion with signal modification became the most effec- niques to accomplish the controls were studied. Fortive. The results are shown in Table 6.2. the signal timing modification (ATMS), three con-

Three diversion plans were compared: equal distri- trol variables were used to modify the signal timing;

bution at the nearest intersection, optimal distribu- the degree of saturation, the maximum queue length,

tion at the nearest intersection, and equal distribution and the stop time. The use of stop time yielded the

at two upstream intersections. As shown in Table lowest total travel time. The results of this analysis

6.3, the optimal distribution diversion plan results in are shown in Table 6.4.

a slight improvement over the equal distribution The signal coverage experiment indicated that theplan. signal modification at all intersections was better

than only changing the signal on the diversionroutes.


60

FAST-TRAC

6.2 Recommendation time decisions on alternative control strategies

This research has contributed to the understand- depending on the traffic condition existing at the

ing of the impact of several control strategies on inci-time of the incident and the severity of the incident.

dent congestion. It also identifies several areas The same study framework can be used to experi-where there is a need for further research. The study ment the impact of the incident on different networkcan be extended to cover additional incident types geometric configuration to uncover an absoluteand durations under various levels of demand. This effectiveness of acould be used to develop a set of rules to make real

Table 6.2 Effectiveness Of The Control Strategies Under Different Demand Levelsa. Difference in total travel time from the no control change scenarios (percent)

(lo-minute both-lane closure incident)

Control strategy



ATIS/ATMS: traffic diversion with signaltiming modification

Off-peak

- 36

- 57*

- 48

Demand level

Peak

- 32

- 52*

- 40

Highly-con-gested

- 56

- 61

- 75*

b. Difference in congestion duration from the no control change scenarios (percent)

(lo-minute both-lane closure incident)

Control strategy Demand level

Off-peak Peak Highly-con-gested

ATMS: traffic metering + 0 + 8 - 21

ATIS: traffic diversion - 23” - 32* - 49

ATIS/ATMS: traffic diversion with signal - 23” - 32” - 55*timing modification

Note: 1. the negative signs indicates the improvement over the no control change

2. the * means this control strategy produces the greatest improvement in that demand situation.


61 me

FAST-TRAC

Measure of effectiveness

Difference in total travel timefrom the no control change

Difference in congestion dura-tion from the no control change

1 O-minute both-lane closure incident

Diversion plan

Equally distributed atthe nearest upstream

intersection

- 52

- 32

Optimal distributionat the nearest

upstream intersection

-57

-32

Equally distributedat two upstream

intersections

-53

-24

Table 6.4 Comparison Of Alternate Signal Control Policies For Signal Modificationlo-minute both-lane closure incident

Measure of

effectiveness

Difference in total travel timefrom the no control change

Difference in congestion dura-tion from the no control change

Control variables for signal timing modification

Degree of saturation

-40

- 32

Maximum queuelength

-45

-20

Stop time

- 58

- 32

1

control strategy under different network circum- Since the study was designed to test certain selectedstances, in addition to the different volume and inci- control strategies for a wide range of incident condi-dent conditions. tions and one incident situation was selected for the

The control strategies should be incorporated into acomplete simulation analysis tool. One of the limi-tations of this research is the lack of simulation toolswhich integrates control strategies. Although theNETSIM simulation program was used in this study,it lacks of capability to incorporate an adaptive sig-nal logic as well as dynamic traffic routing. Thedevelopment of these features will be a valuable con-tribution to the extension of these experiments oncontrol strategies.

sensitivity analysis, the results from the sensitivityanalysis demonstrate the effectivenesses of variouscontrol strategies only in that situation. A full facto-rial analysis of these control strategies may be onlyway to reveal their effectivenesses in all range oftraffic and incident conditions. It is recommendedthat the full factorial design of the experiment shouldbe conducted to determine the most appropriate con-trol strategy for a particular incident and traffic con-dition.


62

FAST-TRAC

BIBLIOGRAPHYAl-Deek, H. and A. Kanafani (1993) Modeling the

benefits of advanced traveler informationsystems in corridors with incidents. Trans-portation Research C, Vol 1, No 4, pp 303-324.

Allsop, R.E. and J.A. Charlesworth (1977) Traffic ina signal-controlled road network: an exam-ple of different signal timings inducing dif-ferent routing. Traffic Engineering andControl, May, pp 262-267.

Ancker Jr., C.J. and V. Gafarian (1967) A simplerenewal model of throughput at an oversatu-rated signalized intersection. TransportationResearch A, Vol 1, pp 57-65.

d’ Ans, G.C. and D.C. Gazis (1976) Optimal controlof oversaturated store-and-forward transpor-tation networks. Transportation Science, Vol101, pp l-19.

Beaubien, R.F. (1994) Incident management notes.Traffic engineering enforcement committee.Hubbell, Roth & Clarks Inc.

Belcher, P.L. (1990) The effects of traffic manage-ment and control on vehicle capacity in Lon-don. Third International Conference onRoad Traffic Control, Institute of ElectricalEngineers, Conference Publication No 320,pp 67-7 1.

Bell, M.C. and PT. Martin (1990) Use of detectordata for traffic control strategies. Third Inter-national Conference on Road Traffic Con-trol, Institute of Electrical Engineers,Conference Publication No 320, pp 86-89.

Bretherton, R.D. et. al. (1986) The use of SCOOTfor traffic management. Second Interna-tional Conference on Road Traffic Control,Institute of Electrical Engineers, ConferencePublication No 260, pp 8 l-84.

Bretherton, R.D. and G.T. Bowen (1990) Recentenhancement to SCOOT - SCOOT version2.4. Third International Conference on RoadTraffic Control, Institute of Electrical Engi-neers, Conference Publication No 320, pp95-98.

Castling, I. (1994) Advanced technology for roadtransport: IVHS and ATT. Artech House,Boston.

Church, R. and C. Revelle (1978) Modelling anoversaturated intersection. TransportationResearch A, Vol 12, pp 185-189.

Cohen, S. (1986) Traffic control strategies duringpeak hours: The case of Aix-En-Provence.Second International Conference on RoadTraffic Control, Institute of Electrical Engi-neers, Conference Publication No 260, pp133-136.

Cracknell, J.A. et. al. (1975) Physical restraint:Greater London’s cordon restraint study.Traffic Engineering and Control, Septem-ber, pp 368-372.

Davies, P. et. al. (1991) Assessment of advancedtechnologies for relieving urban traffic con-gestion. NCHRP Report 340, TransportationResearch Board, Washington, DC.

Drenick, R.F., S.A. Ahmed, and W.R. McShane(198 1) Decentralized control of congestedstreet networks. Transportation ResearchRecord 795, National Academy of Sciences,Washington, DC, pp 24-27.

Drew, D.R. and C. Pinnell (1965) Some theoreticalconsiderations of peak-hour control of arte-rial street systems. Traffic control: Theoryand instrumentation, Instrument Society ofAmerica, Plenum Press, NY, pp 9 I- 123.

The Model Analysts Report on the Benefits of SCATS in Alleviating the Impacts of lncidentsFAST-TRAC - Phase IIB #20

63

FAST-TRAC

Dudek, C.L. (1976) Better management of trafficincidents: Scope of the problem. ibid.

Dudek, C.L. (1988) Solutions to problems generatedby incidents and special events. Strategies toalleviate traffic congestion, Institute ofTransportation Engineers, Washington, DC,pp 24-31.

Easa, S.M. and A.D. May (1981) Evaluation of traf-fic management strategies in central busi-ness districts. Transportation ResearchRecord 816, National Academy of Sciences,Washington, DC, pp 1- 10.

Federal Highway Administration (1995) TRAF userreference guide. Office of Safety and TrafficOperations, March.

Gartner, N.H. and D.L. Hou (1992) Comparativeevaluation of alternative traffic control strat-egies. Transportation Research Record1360, Transportation Research Board,Washington, DC, pp 66-73.

-

Gazis, D.C. (1964) Optimum control of a system ofoversaturated intersections. OperationsResearch, Vol 12, pp 815-831.

Gerlough, D.L. (1965) Simulation as a tool in trafficcontrol system evaluation. Traffic control:Theory and instrumentation, InstrumentSociety of America, Plenum Press, NY, pp71.

Gray, B.M. and L. Ibbetson (199 1) Operational traf-fic control strategies for congestion. TrafficEngineering’and Control, February, pp 60-66.

Hall, R.W. (1993) Non-recurrent congestion: Howbig is the problem?, Are traveler informationsystems the solution?. TransportationResearch C, Vol 1, No 1, pp 89-103.

Hobeika, A.G., et. al. (1993) Real-time traffic diver-sion model: Conceptual approach. Journalof Transportation Engineering, Vol 119, No4, pp 5 15-534.

Holmes, G.N., and D.R. Leonard (1993) “The fre-quency and importance of incidents whichcause congestion in urban areas.“, TransportResearch Laboratory, Contractor Report342.

Hunt, PB. and T.R. Holland (1985) The effects of anincident in SCOOT system - in simulationand on-street. Traffic Engineering and Con-trol, February, pp 55-58.

Institute of Transportation Engineers (1989) Toolboxfor alleviating traffic congestion. Washing-ton, DC.

Kagolanu, K. (1994) A comparative study of trafficcontrol system ITE Compendium of Techni-cal Papers, Institute of Transportation Engi-neers, pp 374-378.

Gerlough, D.L. and EA. Wagner (1967) Improved Kay, J.L., J.C. Allen, and J.M. Bruggeman (1975)criteria for traffic signals at individual inter- Evaluation of first generation UTCS/BPSsections. NCHRP Report 32, Transportation control strategy. Report FHWA-RD-75-26,Research Board, Washington, DC. Federal Highway Administration, March.

Gordon, R.L. (1969) A technique for control of traf- Lindley, J.A. (1987) Urban freeway Congestion:fic at critical intersections. Transportation Quantification of the problem and effective-Science, Vol 3, No 4, pp 279-288. ness of potential solutions. ITE Journal, Jan-

uary, pp 27-32.


64

Lindley, J.A. (1989) Urban freeway congestionproblems and solutions: An update. ITEJournal, December, pp 2 l-23.

Little, J.D.C. (1966) The synchronization of trafficsignals by mixed-integer linear program-ming. Operation Research, Vol 14, No 4, pp568-594.

Longley, D. (1968) A control strategy for a con-gested computer-controlled traffic network.Transportation Research, Vo12, pp 39 l-408.

Lowrie, P.R. (1982) Sydney co-ordinated adaptivetraffic control system-Principle, methodol-ogy, algorithm. International Conference onRoad Traffic Signaling, Institute of Electri-cal Engineers, Conference Publication No207, pp 67-70.

Lowrie, P.R. ( 1990) SCATS - Sydney co-ordinatedadaptive traffic system: A traffic responsivemethod of controlling urban traffic. Roadsand Traffic Authority, Sydney.

Luedtke, P.M. and J.T. Short (1994) User benefitfrom an arterial incident detection andresponse system. ITE Compendium of Tech-nical Papers, pp 523-527.

Maher, M.J. and R. Akcelik (1975) The re-distribu-tional effects of an area traffic control policy.Traffic Engineering and Control, Septem-ber, pp 383-385.

May, A.D. and D. Westland (1979) Transportationsystem management TSM-type projects insix selected European countries. TrafficEngineering and Control, February, pp 18-20.

May, A.D. and F.O. Montgomery (1986) Avoidingthe need for manual intervention in signalcontrol: A case study of Bangkok. SecondInternational Conference on Road TrafficControl. Institute of Electrical Engineers.Conference Publication No 260.

May, A.D. and F.O. Montgomery (1993) The assess-ment of fixed-time signal control in highly-congested conditions. Traffic Engineeringand Control, pp 176- 179.

McShane, W.R. and L.J. Pignataro (1978) Someresults on guidelines for treatment of trafficcongestion on street networks, ibid., pp 19-25.

McShane, W.R. and R.P. Roess (1990) Traffic Engi-neering. Prentice Hall, Englewood Cliffs,NJ.

Michalopoulos, P.G. and G. Stephanopoulos (1977a)Oversaturated signal systems with queuelength constraints-I: Single intersection.Transportation Research A, Vol 11, pp 413-421.

Michalopoulos, P.G. and G. Stephanopoulos(1977b) Oversaturated signal systems withqueue length constraints-II: Systems ofintersections. Transportation Research A,Vol 11, pp 423-428.

Michigan Department of Transportation ( 1992) Met-ropolitan Detroit Advanced Traffic Manage-ment System (ATMS). MetropolitanTransportation Center, Traffic&Safety Divi-sion.

Miner, R.P. (1990) After the incident management:Reducing the impact of traffic incidentsthrough planned and coordinated response.Technical Papers of ITE's 1990, 1989, and1988 Conferences, Institute of Transporta-tion Engineers, Washington, DC, pp 78-89.


65

,A

‘_-

-

I

-

I

I-

-

--

--

-a

‘.L_

i

.-

d

+-

FAST-TRAC

-

Montgomery, F.O. and D.J. Quinn (1992) Queuemanagement strategies using fixed time sig-nal control. IEE Conference 355. April, pp80-84.

National Cooperative Highway Research Program( 1967) Improved criteria for traffic signalsat individual intersections. Report No 32,Highway Research Board, Washington, DC.

. (1969) Improved criteria for trafficsignal systems on urban arterials. ReportNo 73, Highway Research Board, Washing-ton, DC.

. (1971) Improved criteria for trafficsignals in urban networks. Report No 124,Highway Research Board, Washington, DC.

. (1978) Traffic control in oversatu-rated street networks. Report No 194, Trans-portation Research Board, Washington, DC.

. (1991) Freeway Corridor Manage-ment, Synthesis of Highway Practice 177,Transportation Research Board, Washing-ton, DC.

Nip, K.F. (1975) A comparative study of traffic con-trol schemes at an isolated intersection. Traf-fic Engineering and Control, pp 264-269.

OECD (1985) Dynamic traffic management inurban and suburban road systems. RoadTransport Program, Reported by ScientificExpert Group MC3, Paris, France.

OECD (1981) Traffic control in saturated condi-tions. Road Research, Paris, France.

Pignataro, L.J. (1973) Traffic engineering: Theoryand practice. Prentice-Hall, EnglewoodCliffs, NJ.

Pigantaro, L.J. (1974) Oversaturation terms andmeasurements seem adequate after field val-idation. Traffic Engineering, Vol44. No 9,pp 20-24.

Pooran, F.J. and H.C. Lieu (1994) Implementationof operating strategies for integrated free-way/arterial control systems. ITE Compen-dium of Technical Papers, pp 364-368.

Quinn, D.J. (1992) A review of queue managementstrategies. Traffic Engineering and Control,November, pp 600-605.

Rakha, H. et. al. (1989) Evaluating the benefits andinteractions of routes guidance and trafficcontrol strategies using simulation.VNIS’89, IEEE-VTS-V399, pp 296-303.

Rathi, A.K. (1991) Traffic metering: An effective-ness study. Transportation Quarterly, July,pp 421-440.

Rathi, A.K. and E.B. Lieberman (1989) Effective-ness of traffic restraint for a congested urbannetwork: A simulation study. TransportationResearch Record 1232, TransportationResearch Board, pp 95- 102.

Rathi, A.K. and A.J. Santiago (1990) The newNETSIM simulation program. Traffic Engi-neering and Control, pp 3 17-320.

Rathi. A.K. (199 1) Traffic metering: An effective-ness study. Transportation Quarterly, Vol45, No 3, pp 42 l-440.

Reid, D.A. and M.C. Bell (1990) Junction designand congested route choice. Third Interna-tional Conference on Road Traffic Control,Institute of Electrical Engineers, ConferencePublication No 320, pp 8 l-85.


66

FAST-TRAC

Roberg, P ( 1995) A distributed strategy for eliminat-ing incident-based traffic jams from urbannetworks. Traffic Engineering and Control,June, pp 348-354.

Roberg, P. and CR. Abbess (1994) Diagnosis andtreatment of congestion problems in centralurban areas. Presented at EURO 13/ORConference, Strathclyde.

Robertson, D.I. and R.D. Breherton ( 199 1) Optimiz- Transportation Research Board (1987) Managementing networks of traffic signals in real time - of freeway emergencies and special events.The SCOOT method. IEEE Transactions on Transportation Research Circular 326.Vehicular Technology, Vol 40, No 1, pp ll- National Research Council. Washington,15. DC.

Robertson, G.D. (1986) Dual strategy U.T.C. - vehi-cle actuation under computer control andSCOOT. Second International Conferenceon Road Traffic Control, Institute of Electri-cal Engineers, Conference Publication No260, pp 95-97.

Transportation Research Board (1989) Traffic man-agement for freeway emergencies and spe-cial events. Transportation ResearchCircular 344, National Research Council,Washington, DC.

Robertson. G.D. (1987) Handling congestion withSCOOT. Traffic Engineering and Control,April, pp 228-230.

Ross, D.W., R.C. Sandys, and J.L. Schlaefli (1971)A computer control scheme for critical-intersection control in a urban network.Transportation Science, Vol 5, No 2, pp 14 l-160.

Rothenberg, M.J. (1985) Urban congestion in theUnited States: What does the future hold?ITE Journal, Vol 55, No 7, pp 22-39.

Sarakki, V. and J. Kerr (1994) Effectiveness of IVHSelements on freeway/arterial capacity-con-cept and case studies. ITE Compendium ofTechnical Papers, pp 303-307.

Shepherd, S.P. (1992) A review of traffic signal con-trol Institute of Transport Studies, the Uni-versity of Leeds, January.

Taylor. I.G., M.C. Bell, and G.M. Geary (1987)Queue volume: A measure of congestion.Traffic Engineering and Control November.pp 582-585.

Transportation Research Board ( 1984) Developmentin traffic signal system. TransportationResearch Circular 282, National ResearchCouncil, Washington, DC.

Van Aerde, M. and H. Rakha (1989) Developmentand potential of system optimized routeguidance strategies. VNIS’89, IEEE-VTS-V399, pp 304-309.

Van Aerde, M.W. and S. Yagar (1990) Combiningtraffic management and driver informationin integrated traffic networks. Third Interna-tional Conference on Road Traffic Control,Institute of Electrical Engineers, ConferencePublication No 320, pp 1 l- 15.

Van Vuren, T. and D. Leonard (1994) Urban conges-tion caused by incidents. Traffic Engineer-ing and Control, Vol 35, No 7/8.

Waugh, M.R., J.E. Clark, and A. Kanaan (1994)Using simulation models to aid in determin-ing traffic control strategies for IVHS/ATMS signal systems. ITE Compendium ofTechnical Papers, Institute of TransportationEngineers, Washington, DC, pp 470475.


67

FAST-TRAC

Wood, K. (1993) Urban traffic control, systemsreview. Project Report 4 1, TransportResearch Laboratory, Crowthorne, UK.

Wright, C. and K.W. Huddart (1989) Strategies forurban traffic control. Discussion Paper 6,Middlesex Polytechnic, November.

-7’


68

Documents

Incident Management, FAST-TRAC Phase IIB Deliverables€¦ · Van Aerde and Plum ( 1988). The effectiveness of each traffic control strategy depends on demand, the network. and control