incident management: detection, verification, and traffic ... · incident detection, surveillance,...

Intelligent Transportation SystemsField Operational TestCross-Cutting Study

Incident Management:Detection, Verification,

and Traffic Management

September 1998

Prepared for: By:U.S. Department of Transportation Booz·Allen & HamiltonFederal Highway Administration Highway & Vehicle Technology GroupWashington, D.C. McLean, Virginia

Notice

This document is disseminated under the sponsorshipof the Department of Transportation in the interest of

information exchange. The United States Governmentassumes no liability for its contents or use thereof.

Technical Report Documentation Page��

�� !�"�#

#� $��%&'(�� )� ��*$+,,*�+$$��& !�$�$*!&-&$+.&�*+,�! +��$*!�,$+&$ &��(��//0

��!&&��1$$*�&$1�-*�*�+$.��+.+$2�+$+�$*!3 4� ��5��!��6��%�7+�*�*��$*!3��$��*�.��+.+$ 8��6•��9��*��

:� �'�;��<�= 0� ��5��!��6��

7�� 3&��&'(��

/� ��5��!��6��%�%%�� "� ��>1��<$��*&=

8��6•��9��*��

0�"��(�� &'��4"/.�,��37��"� �$��4��/#��""�":

�� &��?��%�%%�� $?��5��% ��%�� %��'��&�'%?

1�&��5$�� (��//:�&��(��//0��%��;@�?��%��

*��$��&?�� #� &��?��%�� !55��#""&� ��;&��3&�3��#�� 7��;��3��")/"

�)� &'��?��

��!55��$��;�� <�!$�=�� 3��#�4

�4� �(��

��

�� !�� !��"��#��

��$��#�� !��#�� %�&��&��#��%�&��

'&��%��(��)��%��'%��*%�!)��(��+,��-��%��%��

��.��/��#��+,��)��

%��&��#��#��)��%��%�&��"��/�&��%��

��&��(�,��*��)%'��!0��

��$��#�� !��#��&��&��#��1��-��1��

��&��$��#��#��

��#��#��$��#��

��#��"��-��1��#��&��&��#��

��&��$��#��#��#��#��#��1��#��

��-#��&��#��#��$

�:� A�?��%� �0� ��('��&��

��'��3*��%��.��3$��55��.��3 ��$;��%��'�� (��;�7��5��3��3*$&,��,��%3�!$�3��% �'(��;��'�;�;��$��;��*�5��!��$��3&'� �� &�� 3&��5��%37��4��

�/� &��'��?��5�<�5�;��= �"� &��'��?��5�<�5�;��= �� 5 ��

1��5��% 1��5��% �4

�� %'��5��%��'�;��6�%

TABLE OF CONTENTS

1.0 EXECUTIVE SUMMARY .................................................................................................1

1.0 REPORT BACKGROUND ................................................................................................2

2.0 INTRODUCTION .............................................................................................................. 2

WHAT ARE INCIDENT DETECTION , VERIFICATION , AND TRAFFIC MANAGEMENT ? ....... 3INFORMATION SOURCES ..................................................................................................... 4FIELD OPERATIONAL TESTS CONSIDERED IN THIS ANALYSIS .......................................... 5

Advanced Driver and Vehicle Advisory Navigation Concept ................................. 5Alternate Bus Routing ................................................................................................ 5Atlanta NAVIGATOR ................................................................................................ 5Borman Expressway ATMS........................................................................................ 6Cellular Applied To Intelligent Transportation Systems Tracking and Location ................................................................................................................ 6During Incidents Vehicles Exit to Reduce Time ....................................................... 6Faster And Safer Travel Through Traffic Routing and Advanced Controls ................................................................................................................ 6Multijurisdictional Virginia Live Aerial Video Surveillance System ..................... 6San Diego Smart Call Box .......................................................................................... 7 TransGuide ................................................................................................................ 7TRANSCOM’s Systems For Managing Incidents and Traffic ............................... 8

3.0 FINDINGS ........................................................................................................................ 8

IMPACTS ............................................................................................................................. 8

Incident Detection ....................................................................................................... 9

Vehicles As Probes .............................................................................................. 9

Incident Verification ................................................................................................... 9Incident Management ............................................................................................... 11Technologies Supporting Detection, Verification & Traffic Management .......... 12

USER RESPONSE ............................................................................................................... 13

TABLE OF CONTENTS (Continued)

TECHNICAL LESSONS LEARNED ....................................................................................... 15

Infrastructure Based Detection ....................................................................... 13

INSTITUTIONAL CHALLENGES AND RESOLUTIONS ........................................................... 16COST TO IMPLEMENT ....................................................................................................... 18

4.0 SUMMARY ...................................................................................................................... 18

5.0 BIBLIOGRAPHY ............................................................................................................ 19

U.S. Department of Transportation September 1998Federal Highway Administration Booz· Allen & Hamilton

Incident Management: Detection, Verification, and Traffic Management Field Operational Test Cross-Cutting Study1

EXECUTIVE SUMMARY

�Traffic incidents are obstructions or restrictions totraffic flow, such as stalled vehicles, accidents,construction and maintenance activities, adverseweather conditions, or special events. Expertsestimate that 65 percent of traffic congestion iscaused by incidents. Incident detection,verification, and traffic management is the processof identifying, verifying, responding to, andclearing the incident and then restoring normaltraffic flow. This report summarizes and interpretsrecent findings from Intelligent TransportationSystems (ITS) Field Operational Test (FOT)projects in the field of incident management.

�Incident detection, verification, and trafficmanagement FOTs have provided some importantfindings in testing technical alternatives that can beused to implement core functions rapidly and withless expense than can be achieved throughconventional means. ITS technologies such as theuse of vehicles as probes, wirelesscommunications, and a variety of nonintrusivevehicle detection technologies present promise infully deploying the intelligent transportationinfrastructure. These tests have proven that severalmanagement concepts are essential to realizing thebest results from the partners’ investments. Theseconcepts include allowing adequate time forevaluation, and for the complexities of systemsintegration and testing.

�The FOTs discussed here involved a significantdegree of partnership or teaming, often betweenpublic-private organizations. As with most team-focused undertakings, institutional issues havepresented challenges equal to or greater than thetechnical issues. The greatest challenges havecentered on procurement and contracting, and onthe dynamics of working in a team environment.�The traveling public is the ultimate beneficiary ofincident management services. These tests haveshown that dealing appropriately with the public iscritical to achieving optimal results. Several FOTsimplemented highly successful outreach programs,both assessing public needs and concerns, and

addressing the issues that arose from the publicinteraction.

Several technical lessons were learned from thetests. A relatively small population of probevehicles (5 percent) is sufficient to provideadequate traffic information, if the probes arereasonably distributed in the traffic stream.Vehicle sensors located over the roadwayencountered problems of undercounting whenattempting to distinguish vehicles traveling at highspeed and low vehicle headway or tall vehicles.Video surveillance surfaced as the best method toremotely verify incidents. Projects using RadioFrequency (RF) communications encounteredsome problems with interference but theseproblems were overcome through tuning andrefining the devices. Most RF communicationsproved as reliable as those using wireline systemsdid. Diversion routes can be an effective methodof managing the traffic congestion produced by anincident as long as the capacity of the diversionroutes is adequate and traffic flow is controlled bydynamic signal timing adjustments to maintainservice levels. The sharing of incident informationamong partnering agencies, particularly lawenforcement, emergency services, trafficmanagement, and transit, produced unanticipatedbenefits to the involved agencies. Two effectivecost saving strategies emerged: reusing existingresources and avoiding cable burial costs by usingwireless technology.

The findings summarized from these projects canhelp ITS professionals move rapidly to developand deploy state-of-the-art incident detection,verification, and traffic management systems. Thisreport highlights the successes and problems thesetests encountered while attempting to develop thetechnologies and systems to support incidentmanagement.

REPORT BACKGROUND

In 1991, the U.S. Department of Transportation(USDOT) initiated a new program to address the


Incident Detection, Surveillance, and Management Field Operational Test Cross-Cutting Study2

needs of the emerging ITS field. This programsolicited and funded projects, called FOTs. Thetests were sponsored and supported by severaladministrations of the Department, including theFederal Highway Administration (FHWA), theFederal Transit Administration (FTA), and theNational Highway Traffic Safety Administration(NHTSA).

The FOTs demonstrated potentially beneficialtransportation products, technologies, andapproaches. The FOTs implemented theseproducts, technologies, or approaches on alimited scale under real-world operationalconditions. These tests were an interim stepbridging the gap between conventional researchand development (that formed the idea), andfull-scale deployment (that would seewidespread use of the idea). FOTs typicallyincluded a local or regional transportationagency, as well as the FHWA, as partners in theproject. The partners often included privatesector providers of the equipment, systems, andservices interested in demonstrating their idea.The FOTs concentrated on user service areasneeding a “proof of concept” in order to achievedeployment goals.

A fundamental element of each test was anindependent, formal evaluation. The evaluationwas designed to produce a final report thatdetailed the test’s purpose, methods, andfindings. The evaluation aspect of the testintended to assess whether the product,technology or approach provided effectivesolutions at acceptable levels of cost, schedule,and technical risk.

As the sponsoring organization and a partner inmany of the FOTs, the FHWA played a centralrole. FHWA supported the tests by providing astandardized set of evaluation guidelines and byhelping coordinate and promote the relationshipsamong test partners. The FHWA also acted asthe communications clearinghouse collecting

reviewing, and disseminating information aboutthe tests.

Among the more than 80 FOTs, several testsencompassed the same or similar areas ofinterest. FHWA has prepared several “cross-cutting” studies that compare or synthesize thefindings of multiple tests within a particular areaof interest. The purpose of this series of studiesis to extract from the separate tests the commoninformation and lessons learned that are ofinterest to ITS practitioners. These are lessonsthat could improve the testing and deployment offuture applications of the subject technology.

This report is one of the series described above.It focuses on the topic of Incident Detection,Verification, and Traffic Management using ITStechnologies.

INTRODUCTION

It is estimated that over half of the trafficcongestion in the U.S. is caused by incidents.Incidents such as accidents, construction andmaintenance activities, adverse weatherconditions, parades, sporting events, touristattractions, or other events can cause congestionby temporarily increasing demand or reducingthe capacity of the transportation network. Evenminor incidents, such as a disabled or abandonedvehicle on the shoulder can reduce roadwaycapacity and create a potential safety hazard.

The problem is growing worse. As presented inExhibit 1, the percentage of congestion due toincidents is expected to increase to 70 percent bythe year 2005. Incident related traffic congestionwill cost the U.S. public over $75 billion in theyear 2005 in lost productivity assuming anaverage rate of $10/hour lost due to congestion,and also results in the waste of over 8.4 billiongallons of fuel. These losses, in addition to theair quality and environmental impacts, clearly



indicate the need for immediate measures tomitigate incident occurrence and impact.

Over the past 35 years, incident managementprograms have been implemented in variouslocations throughout the U.S. as a systematicapproach to minimizing the traffic congestion,air quality, economic productivity, and safetyimpacts of incidents. This report presents theresults of nine FHWA ITS FOTs and the ITSdeployment in Georgia for the 1996 SummerOlympics, and discusses the possibleimplications of these findings for furtherdeployment of incident detection, verification,and traffic management services.

WHAT ARE INCIDENT DETECTION ,

VERIFICATION , AND TRAFFIC

MANAGEMENT ?

The National ITS Program Plan stated that “TheIncident Management user service will useadvanced sensors, data processing, andcommunications to improve the IncidentManagement capabilities of transportation andpublic safety officials. The service will helpthese officials to quickly and accurately identifya variety of incidents, and to implement a set ofactions to minimize the effects of those incidentson the movement of goods and people. In

70%

60%

50%

40%

30%

20%

10%

0%

1984 1987 2005

61% 64%70%

Perc

ent o

f Con

gest

ion

from

Inci

dent

s

Exhibit 1: Traffic Congestion Caused by Incidents [Lindley 1989]



addition, the service will help officials toidentify or forecast hazardous weather, traffic,and facility conditions so that they can takeactions in advance to prevent incidents orminimize their impacts.”

Incident management involves five majorphases. They are:

•� Incident Detection

•� Incident Verification or Surveillance

•� Incident Response

•� Incident Clearance

•� Queue Dissipation

�

�As presented in Exhibit 2, these steps occursequentially. The schematic below presents thetemporal distribution of the five phases anddescribes the key time points during the incidentmanagement process.�

�Beside these five steps, on-scene trafficmanagement and motorist informationdissemination commences during the incidentresponse phase and continues throughout theincident impact period.�

�INFORMATION SOURCES�

�This report was prepared using material gatheredas part of Booz·Allen & Hamilton’s work toprovide evaluation oversight support of ITSFOTs. This material includes published andunpublished reports prepared by the testpersonnel and evaluators as well as informationgathered in meetings and conversations with testpersonnel. Booz·Allen was not directly involvedin the conduct of the tests. The reports preparedby the test personnel and evaluators present thefindings, results, and conclusions of the teststhemselves. This report interprets the results ofa group of tests that have a common theme in anattempt to extract lessons that cut across thegroup of tests. Because it draws from the resultsof the tests as a group, this report may offerlessons and conclusions that are not found in thematerial from the individual tests.

�FOTS CONSIDERED IN THIS ANALYSIS�

�Advance Driver and Vehicle AdvisoryNavigation Concept (ADVANCE)�

�The primary focus of the ADVANCE FOT wasto address the concept of dynamic routeguidance. ADVANCE utilized travel time datareported by traffic probes and loop detectors and

�

��

��

��

��

��

��

��

��

��

��

��

Incident Response

Incident Clearance

Queue Dissipation

On Scene Management

Motorists Information Dissemination

Incident Detection

Verification

Detection Time

VerificationTime

DispatchTravel Time

ClearanceTime

GetawayTime

Incident Occurrence

IncidentConfirmed

Restorationto Normal

Traffic Flow

Incident Cleared

Response at Location



anecdotal incident information to providetravelers with the quickest route to theirdestinations in the northwest suburbs ofChicago. The ADVANCE Traffic InformationCenter (TIC) processed and transmitted traveltime data to vehicles equipped with MobileNavigation Assistants (MNAs), which providedroute guidance and dynamic route guidanceinformation directly to drivers through an in-vehicle mounted graphic user interface. TheMNA’s vehicle location system used GlobalPositioning Systems (GPS), RF-baseddifferential GPS, dead reckoning, and mapmatching using a CD-ROM map database.�

�ADVANCE utilized vehicles as probes formonitoring traffic conditions on freeways andarterials in a large geographic area (incidentdetection), and provided diversionary routing tovehicles carrying the onboard vehicle navigationunit (incident management). ADVANCEoperations have been completed.�

�Alternate Bus Routing (ABR)�

�The (New Jersey) ABR FOT combinesinformation from vehicle probes, conventionalvehicle detection technologies, and anecdotalreports to determine optimal routing for buseswithin a defined corridor. The system applies atravel time algorithm to recommend a preferredroute for each bus, and provides the trafficoperations operator-approved routerecommendation to the vehicle operator via anonboard annunciator.�

�The ABR test addresses several areas directlyimpacting incident management. It serves as anadditional test of the ability of vehicle probes toprovide traffic condition information. Itintegrates this probe information withinformation from conventional vehicle detectors.It tests an algorithm that recommends diversionrouting, which is a significant method ofreducing the impact of an existing incident, and

determines the overall travel time impact ofvehicles taking the diversion route.�

�ABR has completed its first phase, and work isunderway on a final evaluation report. Theagencies have decided not to proceed with thesecond project phase.

�Atlanta NAVIGATOR�

�The Atlanta NAVIGATOR ITS deployment (notan FOT) is a massive integrated AdvancedTraffic Management Systems (ATMS)-Advanced Traveler Information Systems (ATIS)deployment in the greater Atlanta metropolitanregion. The system includes several incidentmanagement features including real-time trafficparameter data collection, congestionmonitoring, automatic and manual incidentdetection, freeway safety patrols, Closed CircuitTelevision (CCTV) cameras, changeablemessage signs, highway advisory radio and ahost of interconnected ATIS features todisseminate traveler information. In addition tomanaging traffic conditions during normalAtlanta travel periods, the system was calledupon to manage the special traffic conditionsbrought about by the 1996 Summer OlympicGames.�

�The system continues in operation, withadditional equipment installation, softwareimprovements, and additional integrationplanned or underway.

Borman Expressway ATMS

This FOT developed and evaluated the use ofATMS along a three-mile section of the BormanExpressway in northern Indiana. This test wasPhase I of a project to establish an ATMS for theentire 16-mile Expressway. In Phase I, theIndiana Department of Transportation evaluatedthe use and viability of a variety of above-road



vehicle sensing technologies and an advancedcommunications system.

�Cellular Applied to IntelligentTransportation Systems Tracking andLocation (CAPITAL)�

The purpose of the FOT was to demonstrate theability to locate and monitor cellular phoneequipped vehicles, to gather data about trafficflow, volume, speed, and detection of probableincidents. CAPITAL tested, on Washington,DC, metropolitan area freeways and arterials, amethod of obtaining real time traffic data whichcould be implemented rapidly and with moderatecost of operation and maintenance whencompared to conventional vehicle detectiontechnologies. CAPITAL operations have beencompleted.

�During Incidents Vehicles Exit to ReduceTime (DIVERT)�

�The purpose of the DIVERT ITS FOT is to makeeffective use of parallel arterials adjacent to acongested and incident-prone segment ofinterstate highway in St. Paul, Minnesota.DIVERT provides an opportunity to analyze theutility of diversion routing over arterials, and ofthe impact of diversion on local traffic flow.The system provides demand-responsive signaltiming along the arterials to accommodate theadditional demand, and monitors both freewayand arterial traffic conditions with detection andvideo surveillance equipment. DIVERT iscurrently in operation, with data collectionongoing. DIVERT operations are scheduled forcompletion in the second half of 1998.

�Faster and Safer Travel Through TrafficRouting and Advanced Controls (FAST-TRAC)�

�FAST-TRAC implemented a real time, adaptivetraffic control system integrated with a traveler

information system using in-vehicle display unitsin Oakland County, Michigan. These unitsprovide real time route guidance informationfrom a central routing system that utilizesvehicle location information as traffic probedata. The FAST-TRAC system also uses videoimage detection systems to provide vehicledetection at intersection.�

�FAST-TRAC tested a relatively new vehicledetection technology — video image detection.It also utilized vehicles as probes to determinetravel conditions and it had the capability toprovide diversion routing to vehicles carryingthe onboard traveler information device. FAST-TRAC is now assessing the challenges ofintegrating dissimilar systems into an operatingTraffic Information Management System. TheSCATS and AUTOSCOPE systemsimplemented as part of FAST-TRAC continue inoperation and represent the core of OaklandCounty’s transportation managementinfrastructure. Data collection, analysis, andpreparation of a case study are ongoing.

�Multijurisdictional Live Aerial VideoSurveillance System – Virginia�

�The Multijurisdictional Live Aerial VideoSurveillance FOT assessed the effectiveness andlimitations of traffic surveillance from ahelicopter. This concept could potentiallyprovide video surveillance over a largegeographic area with a low initial cost, andwithout significant communicationsinfrastructure. The surveillance resource couldbe used to gather traffic condition informationover a planned route, and could also be targetedat known problem areas and existing incidentsites. The video received from the system couldbe used to identify and resolve recurringproblems, for traffic studies, to documentconditions for presentation to officials, and inareas not warranting full time video surveillanceinfrastructure. The test also demonstrated anapproach where multiple agencies benefited



from one another’s investments in both capitalresources and in operations and maintenance ofthose resources.�

�The project outfitted an existing policehelicopter with a video camera and microwavetransmission system. The helicopter flew 90minutes during AM and PM peaks, providingvideo images to Virginia Department ofTransportation (VDOT), law enforcement, andemergency services, via links to the FairfaxCounty Public Safety Communications Center.A mobile law enforcement van was also able toreceive video directly from the helicopter.�

�The FOT has been completed and the systemcontinues to operate as part of the VDOT’sNorthern Virginia Transportation ManagementSystem.

�San Diego Smart Call Box�

�The purpose of the San Diego Smart Call BoxFOT was to demonstrate the possibility of usingexisting roadside call boxes as communicationlinks to additional transportation managementinfrastructure, such as vehicle detectors, videocameras, Variable Message Signs (VMS), andweather information stations. Thus, Smart CallBox tested a technology that supports thedeployment of incident detection, verification,and traffic management resources.�

�Successful application of this project’s approachwould allow implementation of transportationmanagement infrastructure more rapidly and atsignificantly reduced cost, by eliminating theneed for fixed communication infrastructure tosupport the roadside equipment. The San DiegoSmart Call Box FOT report of final results ispresently in preparation and review. CalTrans ispursuing a small-scale pilot deployment of someSmart Call Box functions in order to further testand refine the system.

�TransGuide�

�TransGuide is the ITS deployment in the greaterSan Antonio, Texas area. In conjunction withthe implementation of the first phase of theTransGuide ATMS, an ITS FOT was initiatedwhose purpose was to provide a detailedexplanation of the project technology designdecisions. This included the selection of a fullyredundant fiber optic communications network,all-digital communications, and highmagnification video cameras. The operationaltest also documented how the system metstrenuous objectives for the time to detect,verify, and respond to freeway incidents.�

TransGuide demonstrated the ability to applyadvanced technology in a conventional ATMSarchitecture to achieve significant improvementsin safety based on rapid detection, verification,and response to freeway incidents. The systemwas implemented in a multi-agency cooperativeenvironment, and was sized for expansion toprovide regionwide intelligent transportationinfrastructure.

�The TransGuide ITS FOT examined issueswithin the context of the first phase of thedeployment of the TransGuide advancedtransportation management system, whichcovered the interstates within San Antonio’scentral business district. Additional centerlinemiles of TransGuide are being implemented.Simultaneously, modifications to and expansionsof system functionality have been implemented.�

�TransGuide is also one of the first four ITSMetropolitan Model Deployment Initiative sites.This expansion of the FOT will evaluate systemperformance across a larger area, along thenorthern crest of San Antonio’s beltway.

�TRANSCOM’s Systems For ManagingIncidents and Traffic (TRANSMIT)�



�The purpose of the TRANSMIT FOT was toevaluate the ability to use existing vehiclescarrying electronic toll collection tags as trafficprobes as they approached the New York Cityarea. The travel times of these vehicles could beanalyzed in order to reliably detect incidents andto monitor traffic conditions.�

�The TRANSMIT approach of expanding uponthe use of existing infrastructure could provide amoderate cost method for rapidly expanding theability to monitor traffic conditions and toidentify incidents on both freeways and arterialsin a large geographic area.�

�Following measurement of initial results, theagencies involved in TRANSMIT areimplementing the system on additional roadwaysin the New York/New Jersey area. The initialproject area continues to be operated as part ofTRANSCOM’s transportation managementinfrastructure.

�FINDINGS�

�This section presents the comparison of thesimilarities and differences of these tests and aninterpretation of the results. Findings areorganized into five categories:�

•� Impacts—whether the results of the tests

caused changes

•� User Response—how test participants

reacted

•� Technical Lessons Learned—conclusions

about the ease of use, applicability,transferability, and safety of the testedtechnologies

•� Institutional Challenges and

Resolutions—conclusions about the

relationships among the test partners,institutional barriers, and legal issues

•� Cost To Implement—how the costs may

affect the potential development anddeployment of the technologies

�IMPACTS�

�The nine ITS FOTs involving incidentmanagement issues have primarily focused onsix major goals:�

•� Reducing the cost to deploy (and then

operate and maintain) the infrastructurenecessary to monitor traffic conditions andto detect, verify, and respond to incidents

•� Reducing the time necessary to deploy or

expand deployment of intelligenttransportation infrastructure

•� Exploring opportunities for agencies to work

together to implement intelligenttransportation infrastructure and thenoptimize coordinated incident detection,verification, and response activity

•� Effectively utilizing freeways and arterials in

a mutually supportive manner to mitigate theeffects of incident-related congestion

•� Documenting the impacts that incident

management systems have on congestion,safety, and air quality, and documentingpublic acceptance of and compliance withtraveler information delivered by incidentmanagement systems

•� Exploring the effectiveness of several new

technologies for the basic ATMS functionsthat may provide additional functionality(e.g., vehicle classification) beyond thoseneeded by an ATMS or may avoid problemssuch as unreliability, pavement intrusion,



and traffic disruption which are associatedwith conventional solutions

�Findings in each of these areas have providedcritical information necessary to determine if thenew approaches are practical for widespreaddeployment. Several tests also identifiedproblems that are likely to occur with earlydeployments of the new technologies orapproaches, and examined possible solutions.Several tests generated conclusions on benefitsof various incident management approaches.These conclusions are summarized below in fourcategories: incident detection, incidentverification, traffic management, and thetechnologies that support them.

�Incident Detection�

�Vehicles As Probes�

�Several tests have proven the ability of vehiclesto act as probes, providing sufficiently accurateand timely information on traffic conditions(typically link travel times) and likely incidents(typically derived from comparing actual linktravel times to statistical norms) for use in bothtransportation management and travelerinformation. Vehicle probe data has beengenerated via location information derived fromonboard sensors, from vehicle location sensed byan infrastructure-based vehicle navigationsystem, from detecting the RF toll tags carried byvehicles, and from interaction between cellulartelephones used by vehicle passengers and thecellular phone roadside infrastructure. Over 25percent of the incidents on the New York State(NYS) Thruway during TRANSMIT’s testperiod were detected first by TRANSMIT, withhalf of these detected at least 11 minutes earlierthan by conventional vehicle detectors. Insituations where such conventional detectionsystems detected incidents first, nearly 65percent were detected less than 10 minutesbefore they were noted by TRANSMIT. Of

equal importance, TRANSMIT experienced verylow adjusted false alarm rates.

�Several tests have successfully integrated probe-derived link travel time data with point oraverage speed readings from conventionalvehicle sensors. This integration demonstratesthe ability to expand upon existing conventionalvehicle detection infrastructure with lessinfrastructure-intensive traffic monitoring. TheNew Jersey Alternate Bus system foundacceptable results weighting probe data twice asheavily as conventional detector data in itsalgorithm.

�The net effect of these demonstrations is thatseveral other metropolitan areas, such as SanAntonio and Houston, are now using vehicleprobes as an accepted method for rapidly andinexpensively extending the “reach” of theirconventional traffic monitoring environment.Similarly, TRANSMIT is expanding its coveragearea significantly using vehicle probes.�

�Incident Verification�

�Relatively few tests have addressed innovationin video surveillance. Virginia’s Live AerialVideo test addressed the challenge of developingless infrastructure-intensive incident verificationmechanisms, by using an aerial video platform.Performance of the aerial video approach provedin some ways superior to fixed videosurveillance because aerial video:�

•� Quickly scans overall incident scene

•� Quickly detects secondary incidents

•� Can provide specific details via request to

pilot

•� Is useful for monitoring special events that

cover a large area over an extended period



•� Is useful for a variety of objectives, beyond

simple incident verification, such as trafficstudies or corridor analyses

�The TransGuide FOT investigated a verydifferent approach to conventional videosurveillance: the use of very high quality videoto improve the agencies’ ability to respondeffectively to incidents. TransGuideimplemented video cameras at one-mileintervals. The cameras were near-broadcastquality, with an unusually high (48:1)magnification lensing, on pan/tilt mountings.Through use of TransGuide’s high-bandwidthfiber optic communications network, the videoimages were compressed only minimally fortransmission to the traffic operations center. Theobjective of this approach was for the trafficsystems operator to be able to better identify theresources required to manage the incident, sothat they could be immediately dispatched.Although no statistical results are available, testsconfirmed excellent performance in areas suchas nighttime visibility and identification ofsupport needs at the incident site, in comparisonto a variety of other camera technologies,including near-infrared, black/white, lower cost,and lower magnification cameras. TransGuideoperations personnel report that the system hasprovided excellent results under a wide range oflighting and weather conditions, and that theyfeel that their understanding of incidents is wellsupported by the high quality video images.�

�Experience in Maryland with lower resolutionvideo images has also proven to be satisfactoryfor incident verifications, at lower cost.

�Incident Management�

�Trials have verified the potential benefits ofactive incident management to the travelingpublic. Although actual travel time-savings aredifficult to isolate and are often “contaminated”by the benefits of other improvements made in asingle project, safety improvements such as

reductions in secondary incidents, total numberof incidents, and incident response time wereclearly proven. TransGuide experienced a 15percent reduction in injury related accident rates,and projected that the improvement would growto 21 percent based on trends during the testperiod. It also achieved an overall reduction intime to detect, verify, and respond to incidents of20 percent. The test achieved response timesbelow 20 minutes for major and minor incidents,as measured by the San Antonio PoliceDepartment.

�The willingness of motorists to respond toincident-related travel information from bothwide area (VMS, Highway Advisory Radio --HAR) and in-vehicle sources was confirmed byseveral tests. In its first year of operation,TransGuide noted a significant increase in theportion of motorists who had noticed andcomplied with the VMS messages. This portionrose from 33 percent to 80 percent. Further, thenumber of San Antonio motorists stating thatthey used alternate routes during incidentsincreased from 45 percent to 71 percent.�

�In systems with well prepared diversion routes,the primary benefits from incident managementseem to accrue not from diversion but from rapiddetection, effective response, and timelyclearance of incidents. Although TransGuideincluded provisions to modify signal timing onthe arterial “frontage” roads adjacent to theinterstates, this capability was seldom used.Although not statistically verified, anecdotalevidence such as from TransGuide, and anearlier implementation in Fort Worth indicatesthat directional lane control devices can be aneffective support to incident management,assisting in directing travelers into availablelanes and away from incidents and debris, wellin advance of the danger zone.

Not surprisingly, the effectiveness of diversionroutes is highly dependent upon the route’scondition. In order to assure that the diversion



route is capable of effectively supporting its role,most tests monitored the diversion route. Theyalso took steps such as changing signal timingand adding temporary signage to improve theroute’s capacity to carry the temporary additionalload, and to assure that motorists understoodhow to proceed along the route and how andwhere to re-enter the primary roadway. TheDIVERT test provides an excellent example.DIVERT not only performed initial capacityanalysis on the diversion routes for both peakand off-peak periods, but also studied rampcapacity. The system monitored severalmeasures of effectiveness while in the diversionmode, and adjusted signal timing when capacitywas available to keep the level of service fromdropping. In the future, the city also plans toimplement video surveillance along thediversion routes. The ABR system also monitorsand compares the available routes, and onlyrecommends diversion if the alternate route isbetter than the primary route by a predeterminedthreshold.

�An additional benefit from incident managementachieved by several tests accrued from sharingincident information between agencies. Themost common relationships were between lawenforcement, emergency services, trafficmanagement, traffic control, and transit. Insome cases this sharing was achieved byelectronically transmitting the informationbetween separate locations, by collocatingpersonnel from different agencies in a singleTransportation Management Center (TMC), orby other means such as voice communications,pager, or fax. TransGuide furnishes oneexample of how this works. As the sensornetwork detects a suspected incident, a TxDOTshift supervisor is notified. If the supervisorconsiders it to be worthy of verification, theincident alarm occurs not only at trafficoperations, but also at the San Antonio PoliceDepartment officer station in the control center.Personnel from Via Metropolitan Transit, SanAntonio’s transit property, are also onsite within

the control center, and have access to theinformation over identical workstations. Aremote workstation connection has also beenprovided to Via’s downtown headquarters. Inanother test, the Virginia Live Aerial Videodemonstration shares video images betweenVDOT and local law enforcement andemergency services units. In Atlanta, agenciesworked closely, both in instances wherepersonnel were collocated in the TMC andwhere information was distributed to the city andcounty traffic control centers.

�In all cases, the incident management processkept “the person in the loop,” never allowing thesystem to deliver critical traveler informationwithout approval of an experienced individual.TransGuide, NAVIGATOR, DIVERT, and ABRsystem are highly automated, but none allows itssystem to take action without human approvalbeforehand. These projects achieved maximumcost savings, however, by automating thedetection, verification, and solution developmentto the greatest extent possible.

�Technologies Supporting IncidentDetection, Verification, and TrafficManagement�

�The most common solution found to reduce thecost of incident management systemimplementation was to reuse existing resources.Since communication was often half of the totalcost of deploying an incident managementsystem, using existing communication resourceswould be a significant cost saving strategy. Asecond cost saving strategy would be avoidingthe intensive investment of cable burial by usingwireless communication. Some Smart Call Boxsites avoided both the cost of communicationcable burial through the use of RFcommunications. They also avoided the cost ofpower cable burial by using low-powerequipment and solar power receivers withbattery storage. This arrangement can, however,



present problems in applications with constantdemand (i.e., 30-second polling of detectors), atnight, or in inclement weather situations. Asecond alternative, contracting forcommunications services (“leased lines”) canalso provide the required connectivity withsignificantly lower capitol cost, but will incur anon-going stream of payments for the duration ofthe service.

�A related lesson is that, when making thecommunication investment, there can be a greatlong term benefit to sizing the infrastructure tohandle the eventual bandwidth demand.�

�Experiences with incorporating newtechnologies into the intelligent transportationinfrastructure have been mixed. As many testsattempted to quickly implement and integratemultiple technologies with limited funds, itshould not be surprising that problems arosewith the technologies. Most technologiesdemonstrated in the tests proved adequatelysuccessful, or could be made so with furthertesting, debugging, and integration.�USER RESPONSE�

�All tests evaluating user response have arrived ata uniform conclusion that travelers will act basedupon information from a trusted and well-understood system that provides information ofvalue. TransGuide’s survey of motorist reactionto the information demonstrated a high level ofacceptance of the system (80 percent followedits instructions), and a significant level ofappreciation (71 percent felt they saved time) forthe benefits they derived from it. No test hasquantified isolated travel time benefits fromincident management information alone, butmotorists clearly felt that they derived anappreciable benefit from acting on theinformation, even in situations whereinformation was only available en-route, such asin TransGuide Phase I, and where diversionaryrouting was not utilized. Acceptance data wasexcellent from TransGuide’s implementation of

directional lane control signals, although theimpact of these devices could not be isolatedfrom the impact of the remainder of the incidentmanagement system.�

�FAST-TRAC experienced negative publicperception due to problems with project/systemimage management. TransGuide attributes partof its success to a significant investment inoutreach; DIVERT also executed an extensiveoutreach program, and identified projectobjectives in both travel behavior and useracceptance categories.

�Concerns over privacy, including videosurveillance by “big brother,” and tracking ofmovement of individuals by cellular phone orelectronic toll tag, continue to arise periodically,but much has been done to assuage such fears.CAPITAL directly addressed these issuessuccessfully through a planned media programworking with cellular communicationsproviders. Privacy policies, such as thatdeveloped by ITS America in cooperation withUSDOT and many of its member agencies havemade clear the issues related to privacy whentechnologies being used in incident managementsystems have the potential for violation ofprivacy. Specific measures within a project suchas assignment of random numbers to toll tagtransaction records when they are being usedonly for monitoring traffic flow and componentsin the outreach program have been shown tocalm fears of motorists and area residents.

�Although concerns often arise over the impact ofdiverting traffic from major roadways ontosurface streets, no test has yet demonstrated anysubstantial negative impact from suchdiversions. Projects that included planneddiversions carefully prepared incidentmanagement solutions, such as TransGuide’s“scenarios” which were tailored to each potentialincident type and location. Both DIVERT andTransGuide developed these solutions inconsensus with agencies responsible for



transportation throughout the affected area. Thiscareful preparation has thus far prevented thefloods of high-speed traffic through residentialareas that were once feared. Other keycomponents of responsibly planning to preventsuch negative impact have been:�

•� Careful traffic engineering analysis of the

alternate route

•� Identification and implementation of

possible improvements to the diversion route(such as signage, lane marking, andgeometrics)

•� Effective coordination of the diversion

route’s traffic control system toaccommodate the increased volumegenerated when it is in use as a diversionroute.

�Together, these and other measures haveprevented situations that could have created asignificant public outcry against the use ofdiversion routing as a component in incidentmanagement.

�TECHNICAL LESSONS LEARNED�

�Infrastructure Based Detection�

�TransGuide implemented a relativelyconventional automated incident detectionsystem, using pairs of inductive loop vehicledetectors in each lane at ½-mile intervals.Properly installed loops are typically botheconomical and reliable in the south Texas roadenvironment. The Texas Department ofTransportation (TxDOT) initially developed andimplemented an algorithm based only on vehiclespeeds. Although this algorithm worked well atreasonable vehicle flow rates, it produced erraticresults late at night with few vehicles travelingthe roadways. An alternative algorithm wasdeveloped, tuned, and implemented for these

periods. In later extensions of the regionaldeployment, passive acoustic detectors havebeen added to the network, and have alsofunctioned adequately. TxDOT reportsconfidence that the system consistently meetstheir target of detecting incidents within twominutes of occurrence.�

�The Borman (Indiana) Expressway FOTimplemented 5 sensor technologies, with a totalof 6 different types of devices, including 12inductive loops. The sensor suite includedmicrowave continuous beam/Doppler shift,microwave frequency modulated/continuouswave, active ultrasonic, active infrared, andpassive infrared. Both overhead and side-mounted configurations were implemented.Conventional inductive loops were also installedfor comparison. The loops exhibited poorreliability and were subject to constructionlosses.�

�The Borman test sought detectors that wouldprovide speed, volume, and occupancyinformation. These parameters are the corecomponents of most computerized incidentdetection algorithms. In addition, the testevaluated one detector that provided vehicleclassification. None of the detectors as testedmet all of the test’s performance criteria. Thebest detector (as installed during the test) missed10 percent-15 percent of the vehicles. Primaryproblems in sensor performance were attributedto low vehicle headway and high vehicle speeds,which typically resulted in undercounting. Inparticular, tall vehicles presented problems foroverhead sensors. Many sensor reliabilityproblems were due to fixable installationdifficulties. The test concluded that, in eachdeployment, sensor adequacy should beevaluated based on the needs of chosenalgorithm. Regarding installation, the testconcluded that if overhead sensors areimplemented, a procedure should be developedto adjust, tune, and repair the sensors thatreduces the difficulty, cost, and danger of such



work, and that eliminates the need for laneclosures.�

�The Atlanta NAVIGATOR experience provideslessons in incident detection. NAVIGATORimplemented a massive traffic detectioninfrastructure – over 315 Autoscope video imagedetection cameras, over 50-radar speed sensors,and 67 CCTV cameras with pan-zoom-tiltcapability. During the 1996 Summer OlympicGames, however, incidents were primarilydetected through manual efforts augmented bythe ITS resources, but without a workingalgorithm. Manual detection was necessarybecause the computerized automatic incidentdetection algorithm was not yet operational intime for the Olympic Games. Sixty percent ofthe incidents were detected through manualmonitoring of CCTV cameras, by MetroNetwork primarily using aerial surveillance, bythe HERO traffic safety patrols, and by *DOTtoll free cellular calls. This mix of methodsproved effective in rapidly detecting incidents,although it was labor intensive and thereforeexpensive, and human factors concerns existregarding the constant scanning of banks ofmonitors by control center personnel.

�The computerized incident detection system wasbrought online in late 1996, using two incidentdetection algorithms. However, due to dataaccuracy problems at a few Autoscopes causinghigh levels of false alarms at these locations, theautomated incident detection system was notused until the end of the year.�

�Several of the key technical lessons learned fromthese projects relate to the use of vehicle probesfor gathering traffic condition data and locatingincidents. A key early finding was that relativelysmall populations of probes (5 percent), ifreasonably distributed over time, are sufficient togenerate traffic information adequate foreffective traffic management. Only when thepopulation of probes becomes exceedingly smallor traffic flow sparse does accurate information

become a problem. CAPITAL noted thisproblem on arterials in its test area, and duringlate PM/early AM hours. The level of probesnecessary seems greater if the number of readerstations is fewer (greater distances apart) and ifprobes are involved in stop-and-go traffic suchas on signalized arterials.

�Data regarding the time to detect incidents withprobes is still thin, but TRANSMIT resultsindicated that probes may perform as well as,and in some instances better than, conventionalvehicle detector-based incident managementsystems.�

�Adequate integration of probe data with datafrom vehicle detectors has been achieved withfairly simple algorithms. ABR added 1/3 probedata to 2/3 detector data. Not much detailedresearch has been performed on the algorithms,and it is likely that improved results will beobtained from refined algorithms.

TRANSMIT FOT logged detection rates rangingfrom 67 percent to 95 percent. The false alarmrate varied from 0 to 3.4 times per 100,000 timesthe algorithm was polled. These results put theTRANSMIT algorithm’s performancecomparable in detection rates, and superior infalse alarm rates to the California algorithm, theHigh Occupancy Algorithm (HIOCC), and thedouble exponential smoothing model.

�The CAPITAL project tested an alternate probetechnology, tracking of vehicles via signals fromcellular telephones. Although troubled byrapidly changing cellular technology, at itscompletion CAPITAL was able to locate probevehicles within a range of 24 to 185 meters, andwas able to detect 93 percent of the incidentslogged by local traffic management agencies.The test, however, experienced an 80 percentfalse alarm rate. The system was comparable incost to loop-based systems, at $25,000 per milefor implementation.�



�None of the tests has identified any better waythan video surveillance to verify incidents andgather the detailed information necessary forincreased effectiveness in incident response.Aerial video surveillance worked quite well inVirginia and Maryland, although no formalevaluation was performed in Maryland. Bothprojects have continued using video surveillanceafter the testing concluded. Technicalchallenges in aerial video were primarily inantenna positioning and transmission path.Virginia achieved best results after adjustingantenna positioning at both transmitting andreceiving ends, and by adding appropriate signalfiltering.�

Other tests involving airpath communication(Smart Call Box and Borman ExpresswayATMS) have encountered antenna aiming andtransmission path problems, as well as problemswith frequency utilization. Frequency problemsare most common in industrially active urbanareas, and may be overcome by using systemsthat offer several frequencies. None of the testsencountered any transmission problems that didnot seem possible to overcome with thoroughtuning and refinement. The Borman ATMSproject found that the primary interference wasfrom other project radios. Most tests of RFcommunications, once fully debugged andoperational, have been at least as reliable asthose using wireline communications, andwithout the exposure to disruption fromconstruction activity. Borman found its fixedsite links had negligible error rates, and werereliable. Communication with mobile platforms,however, varied due to conditions andmovement. In general, though, these tests werenot undertaken until careful analysis of thecommunications environment had beenperformed to avoid any major communicationconflicts and disturbances.�One finding regarding the incident managementprocess has been that incident managementworks best if the incident is monitored and theresponse adjusted after an initial solution has

been initiated. Incidents appear to be dynamic,as are motorists’ responses to them. Thus themost effective response needs to also beappropriately dynamic. Both TransGuide andDIVERT implemented this approach.�

�Many tests have experienced the technicalchallenges of system testing and integration.Smart Call Box encountered particular problemsin using commercial products with off-the-shelfsoftware for new purposes. NAVIGATORcontinued bringing new systems onlinethroughout the Summer Olympics, and lost somecommunication links due to system conflictsduring the Games. Other tests have encountereda variety of “bugs” in both commercial andcustom software. Incident management systemsdepend upon many thousands of lines ofsoftware code, often significantly customized tothe local conditions and operations concept. Insystems implementing new approachesdependent upon algorithms, adequate timeshould be allowed for ongoing refinement of thealgorithm(s). Multiple algorithms may benecessary in order to deal with different road andtraffic conditions (peak/off-peak,weekday/weekend), as was the case inTransGuide.

�Equally challenging has been making the manydifferent devices involved in a comprehensiveincident management system work together asplanned. Key lessons have been:�

•� Retain properly qualified and experienced

organizations for both software and systemsintegration.

•� Develop and execute proper quality

assurance and testing programs.

•� Implement careful documentation and

configuration management programs.



•� Allow adequate time and budget for the

unforeseen challenges that are likely to arisein this key program phase.

�A related finding has been that system repair andrefinement, as with scientific experimentation,suffers if too many changes are made betweentests. Several tests made multiple modificationsand adjustments between tests of operationalcondition, which resulted in greater confusionand no improvement in operational status.DIVERT’s geometric improvements, made at thesame time as the system was implemented,appear to have significantly reduced the numberof situations in which diversion will be needed,although there were fewer than a dozen duringthe test period, making it difficult to determinethe effectiveness of the diversion system.

�INSTITUTIONAL CHALLENGES AND

RESOLUTIONS�

�A wide variety of institutional issues have arisenin the incident detection, verification, and trafficmanagement ITS FOTs. These same issues arecommonly found in larger scale systemimplementations, and the solutions explored inthe FOT program may be useful to thoseprograms.�

�A primary area of difficulty has beenprocurement and contracting. As most FOTshave involved some portion of private sectorfunding, issues have included:�

•� Privacy of sensitive corporate information

•� Intellectual property rights

•� Rights to review of results before

publication

•� Valuation of non-cash contributed resources.

�Since many ITS projects include public-privatepartnerships such as are common in FOTs, FOTexperience with partnerships may also bevaluable. In more than one case, the definitionof “partnership” between the public-privatesector participants has been confusing, withproject partners unsure how a “partnership”differed from a more conventional contractingarrangement. Unfortunately, no universalsolutions have emerged, primarily because eachsituation differs, particularly with regard to theregulations of the involved agencies.�

�Partnership issues have become furthercomplicated in situations where multipleagencies have teamed. Such teaming has thepotential to create a situation where, not only arepartners unaware of one another’s requirements,but also where funding may flow from oneinstitution, through a second, to the final userinstitution, as was the case in FAST-TRAC.This creates a cascade of possibly evenconflicting requirements, and can slow theproject approval process to a point where anymeaningful progress is difficult.

�A complicating result of slow project progress isthe impact of ongoing technologicaladvancement. Several projects have foundthemselves implementing devices or approacheswhich, when the system was conceived, werewell ahead of their time, but which werepossibly commonplace or even outmoded whenthe work was completed. CAPITALencountered major technical challenges due totechnical evolution that altered some of theassumptions on which it was based.Simultaneously, however, it found that changesin the regulatory environment were likely tomake the concept (although not its technicalapproach) highly promising. The scope oftesting undertaken by ADVANCE wassignificantly reduced when its core technologybecame increasingly commerciallycommonplace.

Cost Item Description TRANSMIT(AVI Based)

ILDS VIDS MRDS

Capital Cost:Hardware CostsInstallation Costs

$14,700$21,700

$4,100$50,560

$24,500$45,100

$26,500$25,200

Maintenance Costs/Year $2,900 $7,950 $3,300 $2,900Operations Costs/Year $2,040 $2,040 $2,040 $2,040

Table 1: Comparative Costs Per Detection Site (Six lane))i h )

[after NJIT, 1997]



�Challenges continue during projectimplementation. Extended review processes,differences in the partners’ goals, and slowconsensus-based decision making can add to thedifficulty of moving the project forward.Assurance of continued funding for multi-yearprojects has increased the level of riskexperienced by both partners and the projectsthemselves even once work is well underway.�

�A core message from every project has been thata thorough and effective communicationprogram between the partners in the project canachieve wonders in supporting projectadvancement. Key elements in suchcommunication programs have included:�

•� A thorough knowledge of each partner’s

roles and responsibilities

•� Understanding each partner’s capabilities/

limitations

•� Managing expectations

�The formal partnering process, pioneered by theU.S. Army Corps of Engineers, was viewed inTransGuide as having resulted in saving TxDOTmillions of dollars by project completion.�

�Another institutional issue that is related toworking in a “team” environment is dealing withoutside organizations. FAST-TRACexperienced the challenge of determining whoofficially spoke (primarily to the media) for theproject, and to what extent other partners couldinteract with the media.�

�COST TO IMPLEMENT�

�The major costs associated with incidentdetection include the capital costs of trafficdetection and surveillance deployment andsystem installation, the operation andmaintenance cost, and the system operations

costs. Table 1 presents comparative costs forfour incident detection systems – TRANSMITusing a Vehicle Probe-Based System, InductiveLoop Detection System (ILDS), Video ImageDetection System (VIDS) and Microwave RadarDetection System (MRDS).

�The TRANSMIT system is considerably cheaperthan the other three incident detection systems.While the operations costs are the same for all,TRANSMIT deployment and maintenance costsare equal to or lower than the other three makingit considerably less expensive overall. TheTRANSMIT hardware costs include allhardware required to read and process ElectronicToll and Traffic Management (ETTM) tag datafor estimating travel times. The incidentdetection computer costs are also included. Thecalculation assumes that thousands of usersalready have ETTM toll tags in their vehicles –the costs of outfitting these vehicles with tags (atapproximately $50 to $100 per vehicle) is notincluded as part of the TRANSMIT capital costs.This reduces the system cost since the systemuses existing infrastructure (RF tags) alreadydeployed for a different need – electronic tolling.

�SUMMARY�

�The results of the ITS projects summarized hereshow that the tools and technology for incidentdetection, verification, and traffic managementcan provide significant benefits to userorganizations. These tools and technology havebeen shown to be practical and efficient. Thecombination of technologies has already had apositive impact on the metropolitan areas inwhich they have been implemented. Theexamination of user reaction to the technologieshas shown that they will change their drivingpatterns if they trust and understand the sourceof incident information. These projects havehighlighted several technical lessons that canbenefit organizations that are in the process ofdeveloping and implementing similar system.



The projects have shown that institutional issuesremain one of the most significant challenges;also that these issues can be overcome when thepartners work together. The limited informationabout implementation costs indicates that thenew technologies can be substantially lessexpensive than existing technologies.�

�This report was based on the results of incidentmanagement operational tests in different stagesof completion, some completed, others still inprogress. Considering the results from theseprojects, it is reasonable to expect that thecharacter of the next generation of incidentmanagement systems is likely to demonstrateadditional benefits.�



BIBLIOGRAPHY

Hughes Transportation Management Systems,New Jersey Alternate Bus Routing System PhaseI Test Midterm Review, December 1997

Rutgers University Departments of Civil andIndustrial Engineering, Evaluation of theAlternate Bus Routing Project – Phase 1,December 1997

Booz Allen & Hamilton, Final Report – 1996Olympic and Paralympic Games Event Study,November 1997.

Booz·Allen & Hamilton, FHWA Federal-Aid ITSProcurement Regulations and ContractingOptions, August 1997

Rutgers University, Alternate Bus RoutingSystem Project Overall Description andEvaluation Plan, August 1997

Transportation Studies Center, University ofMaryland, Final Evaluation Report for theCapital-ITS Operational Test anddemonstration Program, May 1997

L.S. Gallegos & Associates, InnovativeContracting Practices for ITS, January 1997

Texas Transportation Institute, Before and AfterAnalysis of the San Antonio TransGuide SystemPhase I, January 1997

U.S Department Of Transportation, IntelligentTransportation Systems (ITS) ProjectsHandbook, January 1997

Intelligent Transportation Society of America,Technology Acquisition for the IntelligentTransportation Initiative, December 1996

DeLeuw, Cather & Company for ADVANCE,The ADVANCE Project: Insights andAchievements, November 1996

Institute for Transportation, New Jersey Instituteof Technology, Summary of the IncidentDetection System Evaluation, September 1996

Institute for Transportation, New Jersey Instituteof Technology, TRANSMIT System EvaluationTest Plan, October 1995

University of Michigan, Michigan StateUniversity, et. al., FAST-TRAC Phase IIA FinalEvaluation Reports, August 1995

Westwood Professional Services, Ind. (forMinnesota Department of Transportation),DIVERT Operational Test Detailed EvaluationPlan, July 1995

San Diego State University, Evaluation PlanSmart Call Box Field Operational Test,November 1994

San Diego State University, Final Report SmartCall Box Field Operational Test, November1994

Edwards & Kelcey, for the MinnesotaDepartment of Transportation, St. Paul IncidentManagement (SPIM) Operational Test ConceptDefinition and Preliminary System Design, July1994

Science Application International Corporation(for the Volpe Transportation Research Center),TRANSCOM/ TRANSMIT Case Study, January1994

University of Michigan for Roads Commissionof Oakland County, FAST-TRAC EvaluationPlan, September 1993



S. Subramaniam, Literature Review of IncidentDetection Algorithms to Initiate DiversionStrategies, Working Paper, Virginia Tech Centerfor Transportation Research, 1991

J. Lindley, Urban Freeway CongestionProblems and Solutions: An Update, ITEJournal, December 1989.

Virginia Transportation Research Council,Evaluation of Live Aerial Video for TrafficManagement, no date

SouthWest Research Institute, TransGuide ITSDesign Report, no date

To access an electronic version of this publicationand other ITS related publications visit the

ITS Electronic Document Librar y (EDL):http://www.its.fhwa.dot.gov/cyberdocs/welcome.htm

EDL Document Number 6328

Visit Our I TS WEB SiteITS Joint Program Office: http://www.its.dot.gov

Publication No. FHWA-J PO-99-034HVH-1/9-98

incident management: detection, verification, and traffic ... · incident detection, surveillance,...

Documents