F066-B10-004 © 2010 The MITRE Corporation. All rights reserved. Modeling NextGen with systemwideModeler Pete Kuzminski Stéphane Mondoloni, PhD January

F066-B10-004© 2010 The MITRE Corporation. All rights reserved.

Modeling NextGen with systemwideModeler

Pete Kuzminski

Stéphane Mondoloni, PhD

January 2010

F066-B10-004© 2010 The MITRE Corporation. All rights reserved.2

For Release to All FAA. This document has been approved for public release. Distribution is unlimited.

This is the copyright work of The MITRE Corporation and was produced for the U.S. Government under Contract Number DTFA01-01-C-00001 and is subject to Federal Aviation Administration Acquisition Management System Clause 3.5-13, Rights in Data-General, Alt. III and Alt. IV (Oct. 1996). No other use other than that granted to the U.S. Government, or to those acting on behalf of the U.S. Government, under that Clause is authorized without the express written permission of The MITRE Corporation. For further information, please contact The MITRE Corporation, Contract Office, 7515 Colshire Drive, McLean, VA 22102, (703) 983-6000.

The contents of this material reflect the views of the author and/or the Director of the Center for Advanced Aviation System Development, and do not necessarily reflect the views of the Federal Aviation Administration (FAA) or Department of Transportation (DOT). Neither the FAA nor the DOT makes any warranty or guarantee, or promise, expressed or implied, concerning the content or accuracy of the views expressed herein.

2010 The MITRE Corporation. As such, the Government retains a nonexclusive, royalty-free right to publish or reproduce this document, or to allow others to do so, for “Government Purposes Only”.


Uses of systemwideModeler to Model the NAS

3

Benefits Estimation Operations Analysis

• Future airspace bottlenecks

• Traffic management initiatives

• NextGen operational improvements

• Data communications

• New runways

• Airspace design

• … and more

2007

2009

2011

2013

2015

2017

Fiscal Year

Av

era

ge

An

nu

al D

ela

y

pe

r F

ligh

t* (

min

)


CAASD’s System-wideAnalysis Capabilities

4

systemwideModeler

base future

visualizationand analysis

experimentation

scenariogeneration

operationalimprovementabstraction

• Base scenarios data-driven• Future scenarios grown via projection, linking/trimming, airframe routing algorithms• 30+ sample days per treatment• Future airspace/routes

• Fast-time simulation at flight-level• 5-10 minute runtime

• Typically 300-600 runs

• 1-2 days runtime

• Database-based• Visualization

package• Measurement tools

• Operational concept definition, benefits mechanism identification and quantification, model parameterization• NAS EA → operational scenarios → influence diagrams• Some higher resolution modeling for airports and sectors


systemwideModeler Approach

5

Flights plansplans

plans

Resources

plansplans

constraints

• Start with initial trajectories and airframe assignments

• Change plans to respect constraints• Delay• Swap airframe• Cancellation• Re-route (research)

• Characterize use by a flight

• Monitor flight plans/progress

• Anticipate resource condition, e.g., occupancy

• Formulate response

• Issue constraints to flights

• Most influential resources– Airports

– Sectors

– Airframes

• For each resource we model– Use (and load in aggregate)

– Acceptable use/load (e.g., capacity)

– Anticipation of use/load

– Response


Recent Improvements to systemwideModeler

Document Number Here© 2009 The MITRE Corporation. All rights reserved.6


Airport Runway Congestion Model

7

• Use is landing/takeoff; load is throughput

• Acceptable load is throughput over short time periods– Proxy configuration for each weather condition

• Tactical response is queueing

• Strategic response is selection of operating point and called rates

runway layout

runway usage

fleet mix

weather

rules

simulation

interface

separations andrequirements libraries

aux

Tactical Model

Arrivals queue, landingspaced at max rate

Departures queue, takingoff spaced at rate feasiblewrt imminent arrivals

Demand Management

• Monitors departure queue and anticipated traffic• Will call Arrival Acceptance Rate (AAR) to thin arrival stream if departure delays unacceptable

runwaySimulator systemwideModeler


Arrival Flow Management Model

8

• minimum spacing over nodes (including AAR)• limits to delay absorption between nodes• planned passage times

scheduling algorithm

spacing definition algorithm

Node detection inpre-processing

• To place realistic loads on TRACONs and en route sectors, systemwideModeler distributes delay absorption for airport arrival congestion

• Maintains anticipated landing schedule; respects AAR

• Solves for node passage times (including pushback and landing)– Limits delays in arrival TRACON, between airborne nodes, and in air before

merge structure

• Spacing definition algorithm used to endogenously define in-trail restrictions over nodes

• Ground delay modeled


En Route Sector Congestion Model

9

sector workload

t

delayed entry

threshold

flight workload

tEntry Exit

flight-attributableworkload

tEntry Exit

anon. workload

t

old new

lookahead

• Upon entering sector, flight plans delay absorption

• Generally plans to take fair share in each sectoro “Fair” by relative increase in transit times

• Ability to absorb delay adjustable by sector-pair

Representing Operational Improvements in systemwideModeler

Document Number Here© 2009 The MITRE Corporation. All rights reserved.10


Aircraft EquipageRNAV/RNPVNAVCurved path capability (radius to fix)RNP ARLPVEFBData Communication (FANS 1/A+, ATN Baseline 1)Flight Information Services - BroadcastData Communications (ATN Baseline 2)GNSS Landing SystemADS-B outADS-B inCDTIGuidance DisplayPaired Approach Guidance

Transformational ProgramsADS-BSWIMNextGen Network Enabled WeatherNAS Voice SwitchData Communications

Airfield DevelopmentRunways, Taxiways & Airfields

Initiate TBODelegated Responsibility for SeparationOceanic In-trail Climb & DescentAutomation Support for Mixed EnvironmentInitial Conflict Resolution AdvisoriesFlexible Entry Times for Oceanic TracksPoint-in-space MeteringFlexible Airspace ManagementIncreased Capacity and Efficiency Using RNAV/RNP

Increase Arrival/Departures at High Density Airports

Improve Operations to CSPRInitial Surface Traffic ManagementTime-Based Metering using RNP and RNAV Route AssignmentIntegrated Arrival/Departure Airspace Management

Increase Flexibility in the Terminal EnvironmentWTMDGBAS Precision ApproachesUse Optimized Profile DescentProvide Full Surface Situation InformationEnhance Surface Traffic Operations

Improve Collaborative Air Traffic ManagementContinuous Flight Day EvaluationTMI with Flight Specific TrajectoriesImproved Management of Airspace for Special UseTrajectory Flight Data ManagementProvide Full Flight Plan Constraint Evaluation with Feedback

Reduce Weather ImpactTrajectory-Based Weather Impact Evaluation

Improve Safety, Security and Environmental Performance

Safety Management System ImplementationSafety Management Enterprise ServicesAviation Safety and Information Analysis and SharingOperational Security Capability for Threat Detection & Tracking NAS Impact Analysis and Risk-based AssessmentSSA and ISS Integrated Incident Detection and ResponseInformation on System Security and Surveillance Integration/ProtectionEnhanced Air Traffic Procedures, Improved Environmental Technologies and Sustainable Alternative Aviation Fuels, and Integrated Environmental ModelingEMS Implementation and Environmental Policy Support

Transform FacilitiesIntegration, development and operations analysis CapabilityNextGen FacilitiesNet-Centric Virtual Facility

11

Images: source FAA

Document Number Here© 2010 The MITRE Corporation. All rights reserved.

Decisions, decisions…

Many improvements with implementation decisions & architectural alternatives.

Many candidate decisions to equip


Modeling Performance Impact of Decisions

12

40 Independent decisions 1 trillion possible outcomes…start modeling!


Discussion focuses here

Evaluating Alternatives

13

Estimate the service provider and user life-cycle costs

Assess the benefit performance of alternatives

Define, Refine, Decompose Operational Concept

Determine feasible alternative evolution paths

Coordinated, iterativeConsistent

Documented in NAS EA • OV-6c• OV-5

Benefit Mechanisms

Used to develop

Influence Diagrams

Documented in

Down-selection

Reduced Influence Diagrams & Timeline

Analysis & Quantification

High-level Operational Requirements

“Functional Clusters”

Influence Diagrams

Used to develop

Consistent with

Down-selection

Timelines

Analysis & Quantification


Influence Diagrams – The Basics

• Four types of nodes† used:– Decision Node

– Metric

– Random Variable

– Key Performance Area

• Influences described with arrows• Arrows with dotted lines prevent “cycles”

Decision

Metric

Random Variable

KPAKey Performance Areas (11) from ICAO

†Different tools use different symbols

Plant Locust-

Resistant Crops

Likelihood of Locusts

Crop Yield

14


NextGen Example

Inter-departure spacing – parallel

runwaysWTMD Departure Capacity

Taxi Delays

Favorable wind

conditions

Number of par. runways <2500feet apart

Fleet-mix

Fuel Consumed

Gate-to-gate time

Operating Costs

Schedule Predictability

Emissions Environment

Capacity

Efficiency

Predictability

• Use to obtain agreement on single mechanism• Provides line-of-sight with interim metrics• Transparent linkages to corresponding costing elements

15


Dependencies & Shared Benefits

• Multiple factors influence the same mechanisms

16

Number of aircraft tactical maneuvers

Number of restrictions

Lateral precision

Route Density

Along-track predictability

Trajectory prediction accuracy

Resolution look-ahead time

Metering Planning Accuracy

… Efficiency

Visualize Dependencies

Many decisions can lead to same impact on interim measures

Arrival Flow Gaps

Some paths provide additional mechanisms


Each “Influence” requires modeling E.g. FMS Offsets

17

Controller Workload

Sector Capacity

En Route Delay

Gate-to-gate time

Fuel Consumed

Emissions Costs

Airline Schedule

Predictability

Number of aircraft flow maneuvers

Offset resolutions


FMS Route Offsets Time/Fuel

• Quantification– From assumptions: 6.4 NMI additional for lead

– Trailing aircraft saves time, but likely incurs a fuel penalty (from operating at cost index > 0)

• Could add many smaller influences

18

SLOWFAST

SLOWSLOWf

FASTFASTff

V

X

V

XT

V

XVW

V

XVWW

)(

8 NMI45ºX

Change in Fuel (Wf) and time (T) of trailing aircraft


FMS Route Offsets Time/Fuel (cont’d)

19

Only consider circumstances where cost to lead < benefit to trailing

Speed Distribution

VSLOWVFAST

Compute Implied CI:

Sample:

60% of Conflicts†

Total Conflicts

Number of Overtaking

TWTCICostFactor f )100(

Duration Distribution

Number of beneficial options (50%)

~N(450,18)

100

12

2

qS

WCqSCCI DDo

Sample Aircraft Types

Average benefit equivalent to X lbs of fuel per eventApplies in 30% of all conflicts (60%*50%)Benefit of 0.3X lbs per conflict

† From Bilimoria, K.,D., Methodology for the Performance Evaluation of a Conflict Probe, J. of Guidance, Control and Dynamics, Vol. 24, No.3, May-June 2001


Linked Benefit Mechanism Influence Diagrams (In Progress)

20

Surface

Arrival/Departure

CATM, Wx, Airspace

RNAV/RNP, TMA

En Route


Parameter Setting - Example

• Influence diagrams documents parameter setting in systemwideModeler

21

Number of aircraft conflict

maneuvers

Trajectory Prediction Accuracy

Use of RTA

Establish Metrics & Relationships

Other things

Affects systemwideModeler Parameters

Model the influence

. . . Further influences


Accuracy Affects Conflicts

Look-ahead

• Detect using 5 NMI + Buffer• Buffer selected to get very few missed alerts† at 5 minutes • Begin responding to alerts at a look-ahead > 5 minutes• False alerts* result in additional conflict resolution

workload• Applied Monte Carlo simulation to obtain buffers• With radar-level accuracy, required buffer = 3 NMI

0

0.5

1

1.5

2

2.5

0 0.1 0.2 0.3 0.4

FA R

ate

MA Rate

CD SOC Curves - Radar

20 MIN

15 MIN

10 MIN

5 MIN23

45

Look-ahead

Increasing buffer

*False alerts = A detected conflict at a specified look-ahead that does not result in loss of separation†Missed alert = A loss of separation that is not detected at a specified look-ahead time‡ Monte-Carlo compared against: Bilimoria, K.D., Lee, H.Q., Properties of Air Traffic Conflicts for Free and Structured Routing, AIAA-2001-4051, GN&C Conference, Montreal, PQ, August, 2001

0%

10%

20%

30%

40%

50%

60%

0-60 60-120 120-180

Perc

ent o

f Con

flict

s

Encounter Angle

Conflict Geometries

Simulation

Measured

‡

22


RTA – Improves Prediction Accuracy

-30

-20

-10

0

10

20

30

40

0 2000 4000 6000 8000 10000

Spee

d (f

ps)

Time (sec)

Open loop speed dynamics

Air

Ground

Wind

Wind variability†

Aircraft dynamics

† Mondoloni, S., A multiple-scale model of wind-prediction uncertainty and application to trajectory prediction, AIAA-2006-7807, ATIO 2006, Wichita, KS

Only controls to time as approach RTA point, speed limited, infrequent speed target changes

-30000

-20000

-10000

0

10000

20000

30000

40000

0 200 400 600 800 1000

Alon

g-tr

ack

erro

r (fe

et)

Look-ahead (sec)

Along-track prediction error (open-loop)

RMS

-10000

-5000

0

5000

10000

15000

0 200 400 600 800 1000

Alo

ng-t

rack

Err

or (f

eet)

Look-ahead (sec)

Along-track prediction error (w/ RTA)

RMS

Design choice

23


Improved Conflict Detection(based on Monte Carlo Simulation)

• ADS-B Improves over radar (better current speed estimation & position error)• RTA Improves further through closed-loop control

0%10%20%30%40%50%60%70%80%

20 15 10 5

Rate

Look-Ahead (minutes)

Missed Alert Rate

Radar

ADS-B

RTA

0%

5%

10%

15%

20%

25%

30%

20 15 10 5

Perc

enta

ge R

educ

tion

Look-Ahead (minutes)

Reduction in Total Alerts from Radar

ADS-B

RTA

†

† Validated against FAA-2007-29305-0012.1 (supporting material to NPRM for ADS-B out)

False Alert Rate

24


Summary

• Improving both models and analysis process– Improving and expanding modeling features of

systemwideModeler

– Applying a structured process from concept to simulations

– Improves capturing of benefit dependencies

– Process reflects benefit and costs dependencies

25

Operational Concept

Benefits Mechanisms

Benefits Quantification

sM Parameter Setting

Operational Concept


Operational Concept


Benefits Mechanisms

Benefits Quantification

26

Documents

F066-B10-004 © 2010 The MITRE Corporation. All rights reserved. Modeling NextGen with systemwideModeler Pete Kuzminski Stéphane Mondoloni, PhD January