Design and Operation of Advanced Control Centers€¦ · high-technology centers and how the...

Comparable Systems Analysis:Design and Operation of Advanced Control Centers

Publication No. FHWA-RD-94-147 August 1995

U.S. Department of Transportation

Research and DevelopmentTurner-Fairbank Highway Research Center6300 Georgetown Pike

Federal Highway Administration McLean, Virginia 22010-2296

FOREWORD

This research explored lessons that have been learned involving human factors in the design andoperation of control centers that were similar to a generic, IVHS-class traffic management center(TMC). During an initial phase, brief visits were made at 10 existing operation control centers. Three ofthese were selected for more detailed study, along with eight additional centers that were stronglyrecommended by center managers and other researchers. During the second phase, structuredinterviews were completed with operators and managers at the 11 centers located in the United States,Canada, and Europe.

This report summarizes the critical input, data processing, and output functions that are expected to beperformed by IVHS-class TMC’s. It describes how these functions are currently performed by thesample of existing high-technology centers and how the functions might evolve with near-termtechnology advancements and automation. It addresses a series of important TMC designconsiderations, including the user-centered design process, operator selection and training, systemevolution and advanced automation, user interface design, and job performance aids.

Lyle SaxtonDirector, Office of Safety and Traffic OperationsResearch and Development

NOTICE

This document is disseminated under the sponsorship of the Department of Transportation in theinterest of information exchange. The United States Government assumes no liability for its contents oruse thereof. This report does not constitute a standard, specification, or regulation.

The United States Government does not endorse products or manufacturers. Trade or manufacturers’names appear herein only because they are considered essential to the object of this document.

1. Repor t No . 2. Government Accession No. 3. Recipient's Catalog No.

FHWA-RD-94-1474. Title and Subtitle 5. Report Date

COMPARABLE SYSTEMS ANALYSIS:August 1995

DESIGN AND OPERATlON OF ADVANCED CONTROL CENTERS 6. Performing Organization Code

7. Author(s) 8. Performing Organization Report No.

Michael J. Kelly, Jeffrey M. Gerth, and Christopher J. Whaley9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)

Electronic Systems Laboratory 3B1E1012Georgia Tech Research InstituteGeorgia Institute of Technology 11. Contract or Grant No.

Atlanta, Georgia 30332 DTFH61-92-C-00094

12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered

Office of Safety and Traffic Operations R&D Task Report (June 93 to November 93)Turner-Fairbank Highway Research Center6300 Georgetown Pike 14. Sponsoring Agency Code

McLean, Virginia 22101-2296

15. Supplementary Notes

Contracting Officer’s Technical Representative (COTR): Nazemeh Sobhi (HSR-30)

The authors are grateful to the management and operators of the pioneering facilities we visited for their time,insight, and candor. By sharing with us both their successes and missteps, they will make the path toward IVHSsmoother for those who follow.

16. Abstract

This research explored lessons that have been learned involving human factors in the design and operation ofcontrol centers that were similar to a generic, IVHS-dass traffic management center (TMC). During an initialphase, brief visits were made at 10 existing operation control centers. Three of these were selected for moredetailed study along with 8 additional centers that were strongly recommended by center managers and otherresearchers, During the second phase, structured interviews were completed with operators and managers at theeleven centers located in the United States, Canada and Europe.

This report summarizes the critical input, data processing, and output functions that are expected to be performedby IVHS-class TMC’s. lt describes how these functions are currently performed by the sample of existinghigh-technology centers and how the functions might evolve with near-term technology advancements andautomation. It addresses a series of important TMC design considerations including the user-centered designprocess, operator selection and training, system evolution and advanced automation, user interface design, and jobperfomance aids.

17. Key words 18. Distribution Statement

Comparable systems, IVHS, also, Traffic ManagementCenter (TMC), Advanced Traffic Management

No Restrictions. This document is available to the publicthrough the National Technical Information Service,

(ATMS) Springfield, Virginia 22161.

19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Price

Unclassified Unclassified 77

Form D0T F 1700.7 (8-72) Reproduction of completed page authorized.

inftydmi

inchesfeetyardsmiles

LENGTH25.40.3050.9141.61

AREA

millimeters mmmeters mmeters mkilometers km

LENGTHmmm

km

millimeters 0.039 inches inmeters 3.28 feet ftmeters 1.09 yardskilometers

yd0.621 miles mi

AREA

in2

acmi2

square inchessquare feetsquare yardsacressquare miles

645.20.0930.8360.4052.59

VOLUME

square millimeters mm2 mm2

square meters m2 m2

square meters m2 m2

hectares ha hasquare kilometers km2 km2

square millimeters 0.0016 square inchessquare meters 10.764 square feetsquare meters 1.195hectares

square yards2.47 acres

square kilometers 0.386 square miles

VOLUME

fl oz fluid ounces 29.57 millilitersgal gallons 3.785 Iiters

cubic feet 0.028 cubic meterscubic yards 0.765 cubic meters

NOTE: Volumes greater than 1000 I shall be shown in m

MASS

mLLm3

m3

mLLm3

m3

milliliters 0.034 fluid ouncesIiters 0.264 gallonscubic meters 35.71 cubic feetcubic meters 1.307 cubic yards

MASS

ozlbT

ounces 28.35 gramspounds 0.454 kilogramsshort tons (2000 lb) 0.907 megagrams

(or “metric ton”)TEMPERATURE (exact)

gkgMg(or "t")

gkgMg(or “t”)

grams 0.035 ounces ozkilograms 2.202 pounds lbmegagrams 1.103(or “metric ton”)

short tons (2000 lb) T

TEMPERATURE (exact)

°F Fahrenheit 5(F-32)/9 Celciustemperature or (F-32)/1.8 temperature

ILLUMINATION

°C °C Celciustemperature

1.8C + 32

ILLUMINATION

fcfl

foot-candles 10.76 luxfoot-Lamberts 3.426 candela/m2

FORCE and PRESSURE or STRESS

lxcd/m2

lxcd/m2

lux 0.0929 foot-candlescandela/m2 fc

0.2919 foot-Lamberts fl

FORCE and PRESSURE or STRESS

IbfIbf/in2

poundforcepoundforce persquare inch

4.45 newtons N N newtons 0.2256.89 kilopascals k P a kPa kilopascals 0.145

in2

acmi2

fl ozgal

Fahrenheittemperature

°F

poundforcepoundforce persquare inch

IbfIbf/in2

*SI is the symbol for the International System of Units. Appropriaterounding should be made to comply with Section 4 of ASTM E380.

(Revised September 1993)

2

yd 2

m

ft

3

3

yd 3ft 3

yd 3ft

2

yd 2ft

Symbol When You Know Multiply By To Find Symbol

APPROXIMATE CONVERSIONS TO SI UNITSSymbol When You Know Multiply By To Find Symbol

APPROXIMATE CONVERSIONS TO SI UNITS

SI* (MODERN METRIC) CONVERSION FACTORS

TABLE OF CONTENTS

Section 1. BACKGROUND AND SUMMARY ..................................................................................... 1

TRAFFIC MANAGEMENT CENTER OBJECTIVES ................................................................ 1

DESIGNING THE TMC ........................................................................................................... 2

Section 2. THE COMPARABLE SYSTEMS ANALYSIS METHODOLOGY ........................................ 5

PHASE I. IDENTIFYING COMPARABLE SYSTEMS ............................................................. 6

PHASE II. IDENTIFYING LESSONS LEARNED..................................................................... 7

Section 3. SYSTEM OVERVIEWS FOR SITE VISITS ........................................................................ 9

AIR COMBAT MANEUVERING INSTRUMENTATION ............................................................ 9

AIR MOBILITY COMMAND TANKER AIRLIFT COMMAND CENTER ..................................... 9

ALLIED SERVICES CORPORATION DISPATCH CENTER.................................................. 10

ANAHEIM TRAFFIC MANAGEMENT CENTER .................................................................... 10

AMSTERDAM PUBLIC TRANSIT CONTROL CENTER ........................................................ 11

AMSTERDAM SIGNAL CONTROL CENTER ........................................................................ 11

AUTOMATED TRAFFIC SURVEILLANCE AND CONTROL SYSTEM .................................. 12

CALTRANS DISTRICT 7 TRAFFIC OPERATIONS CENTER ............................................... 12

CHICAGO IDOT COMMUNICATIONS CENTER................................................................... 13

CHICAGO IDOT TRAFFIC INFORMATION (ADVANCE) CENTER ....................................... 13

CHICAGO IDOT TRAFFIC SYSTEMS CENTER ................................................................... 14

FAA AIR TRAFFIC CONTROL SYSTEM COMMAND CENTER............................................ 14

FLORIDA FREEWAY MANAGEMENT SYSTEM .................................................................. 14

MINNEAPOLIS DOT TRAFFIC MANAGEMENT CENTER .................................................... 15

METRO ORLANDO TRAFFIC MANAGEMENT CENTER ..................................................... 15

NETHERLANDS MOTORWAY CONTROL AND SIGNALING SYSTEM ............................... 16

TORONTO 401 FREEWAY TRAFFIC MANAGEMENT CENTER ......................................... 17

Section 4. STATE-OF-THE-ART OF IVHS TMC TECHNOLOGIES ................................................. 19

OPERATIONAL REQUIREMENTS ....................................................................................... 19MONITOR CURRENT TRAFFIC CONDITIONS .................................................................... 19

MONITOR CURRENT ROADWAY CONDITIONS ................................................................. 20IDENTIFY LOCATION OF INCIDENTS ................................................................................. 20

CLASSIFY TYPE AND SEVERITY OF INCIDENTS .............................................................. 21IDENTIFY LOCATION OF EMERGENCY RESOURCES ...................................................... 22

MONITOR CURRENT AND PREDICTED WEATHER ........................................................... 22

PREDICT DEMAND .............................................................................................................. 23MONITOR PUBLIC CONFIDENCE ....................................................................................... 23PREDICT TRAFFIC CONDITIONS ....................................................................................... 23DETERMINE MAINTENANCE NEEDS.................................................................................. 23SELECT TRAFFIC MANAGEMENT TACTICS ...................................................................... 24

SELECT INCIDENT MANAGEMENT TACTICS .................................................................... 24

SELECT DEMAND REGULATION TACTICS ........................................................................ 25

SELECT PUBLIC RELATIONS TACTICS.............................................................................. 25INFLUENCE ROUTE SELECTION ........................................................................................ 25INFLUENCE VEHICLE MOTION ........................................................................................... 26CONTROL ROADWAY ACCESS .......................................................................................... 27DISPATCH EMERGENCY AND MAINTENANCE RESOURCES ........................................... 27INFLUENCE TRIP DECISIONS AND INFLUENCE MODE SELECTION ............................... 28PROVIDE PUBLIC INFORMATION ....................................................................................... 28

i i i

TABLE OF CONTENTS (CONTINUED)

Section 5. USER-CENTERED AUTOMATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

VMS AUTOMATION PHILOSOPHIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30AUTOMATED RESPONSE PLANNING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31AUTOMATED LANE CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31FLEXIBLE FUNCTION ALLOCATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32AUTOMATION TO SUPPORT THE HUMAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Section 6. SYSTEM EVOLUTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

PLAN FOR EVOLUTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35INTEGRATION OF THE OLD WITH THE NEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36MORE IS BETTER? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Section 7. USER SYSTEM INTERFACES AND WORKSTATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Section

Section

Section

Section

RAPID PROTOTYPING OF DISPLAYS DURING TMC DESIGN........................................... 40THE CARDINAL RULE: HUMANS MAKE ERRORS ............................................................. 40HOW MANY OPERATORS? ................................................................................................. 41BIG BOARD DISPLAYS ........................................................................................................ 41CCTV SYSTEMS .................................................................................................................. 44VARIABLE MESSAGE SIGNS .............................................................................................. 46VIRTUAL DISPLAYS ............................................................................................................. 46

8. STAFFING, SELECTION, AND TRAINING ..................................................................... 49

WHO’S ON DUTY IN THE TMC ............................................................................................ 49INITIAL TRAINING ................................................................................................................ 49SYSTEMS APPROACH TO TRAINING................................................................................. 51RECURRENT TRAINING ...................................................................................................... 52

9. PROCEDURES, DOCUMENTATION, AND JOB AIDS.................................................... 53

LEADING-EDGE TRENDS .................................................................................................... 60

10. OTHER DESIGN AND PRIVACY ISSUES ..................................................................... 63

ERGONOMICS IN CENTER DESIGN ................................................................................... 65PRIVACY ISSUES ................................................................................................................ 65

11. DESIGN CONSIDERATIONS BASED ON COMPARABLE SYSTEMS ANALYSIS...... 67

USER-CENTERED DESIGN ................................................................................................. 67STAFFING ............................................................................................................................ 67MANAGEMENT .................................................................................................................... 67AUTOMATION ...................................................................................................................... 68USER-SYSTEM INTERFACES ............................................................................................. 68ENVIRONMENTAL AND ERGONOMIC CONSIDERATIONS ................................................ 69

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

i v

LIST OF FIGURES

Figure No.

1 .2 .3 .4 .5 .6 .7 .8.9.10.11.12.13.14.15.16.17.18.19.20.21.22.23.24.25.26.

The comparable systems analysis method ................................................................... 5Continuum of operator roles in relation to automation ................................................. 30Example of a broad region zoomed-out with low granularity ....................................... 38Example of a region zoomed-in with medium granularity ............................................ 38Example of an intermediate region zoom-in display with medium granularity.. ............ 39Example of a local area zoom-in with high granularity ................................................ 39Example of an intersection zoom-in with high granularity ............................................ 40Example of a static wall map ...................................................................................... 42Example of a dynamic wall map ................................................................................. 43Example of a computer-drawn, plastic tile matrix wall map ......................................... 44Example of a projection television system (left center, large screen) .......................... 45Example of a large matrix of CCTV monitors .............................................................. 45Floor plan of a Cal Poly TOC simulation room ............................................................ 51Example of a procedural checklist developed by TMC operators ................................ 54Example of a flow chart for a detailed response plan .................................................. 55Example of a traffic incident checklist ......................................................................... 56Example of an incident report form............................................................................. 57Camera-related job aids: Location-camera number lookup ........................................ 58Camera-related job aids: Camera-control instructions ............................................... 59Camera-related job aids: Camera pattern-CCTV monitor layout ................................ 59Illustration of a documentation carousel ..................................................................... 60Example of labeling on the control panel .................................................................... 61Example of labeling below a control display ............................................................... 62Example of a control room developed using a user-centered design process ............. 64Example of a monitor located above the operator’s seated line of regard ................... 64Example of glare on a VDT ........................................................................................ 65

LIST OF TABLES

Table No.1. Phase I site visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....... 62. Phase II site visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63. TMC operational requirements ............................................................................... 19

LIST OF ACRONYMS AND ABBREVIATIONS

ACMIAMCASCATMSATSACAVCSCADCaltransCBCCTVCHPCMSCMZCRTCVODOTETAK, Inc.ETMSFAAFCCFHWAFMCFTMSGDSSHARHOVIDLSIDOTIDSIFRITMCIVHSLEDMCSSMTOODOMSPMSPTCCRASSAACSATTACCTMCTOCTravTekTSCUTCSVDTVMSWATC

Air Combat Maneuvering InstrumentationAir Mobility CommandAllied Services CorporationAdvanced Traffic Management SystemLos Angeles Automated Surveillance and Control SystemAutomated Vehicle Control SystemComputer Aided DispatchCalifornia Department of TransportationCitizen BandClosed Circuit TelevisionCalifornia Highway PatrolChangeable Message SignCongestion Management ZoneCathode Ray TubeCommercial Vehicle OperationsDepartment of TransportationProduces a Digital Mapping SystemEnhanced Traffic Management SystemFederal Aviation AdministrationFederal Communications CommissionFederal Highway AdministrationFreeway Management CenterFreeway Traffic Management SystemGlobal Decision Support SystemHighway Advisory RadioHigh Occupancy VehicleIncident Detection and Location SystemIllinois Department of TransportationInformation Dissemination SystemInstrument Flight RulesIllinois Traffic Management CenterIntelligent Vehicle Highway SystemLight Emitting DiodeMotorway Control and Signaling SystemMinistry of Transportation, OntarioOrigin - DestinationOperational Management SystemPerformance Measurement SystemPublic Transit Control CenterResponse Advisory SystemSimulator for Air-to-Air CombatSystems Approach to TrainingTanker Airlift Command CenterTraffic Management CenterTraffic Operations CenterTravel TechnologyTraffic Systems CenterUrban Traffic Control CenterVideo Display TerminalVariable Message SignWide Area Traffic Control

vi

Section 1 l

BACKGROUND AND SUMMARY

The traffic management center (TMC) is theprimary ganglion in the network of trafficsensors, traffic signals and controllers, voicecommunications, electronic signs, and otherresources that are designed to make traffic runsmoothly on our roadways. Older TMC’s arealready common in most major cities where theirfunctions usually include controlling andmaintaining traffic signals and performing otherrelatively low-technology functions. For themost part, these older centers are largelyindependent of other traffic-related services andhave few communication and coordinationrequirements.

The IVHS era is bringing advanced technologyto traffic management. Many TMC’s areexpanding their roles and upgrading theircapabilities. New, advanced technology TMC’sare being designed from scratch. It has beenestimated that by the turn of the century, over200 cities will have adopted some element ofIVHS technology.

IVHS TMC’s incorporate larger numbers ofmore capable sensors to provide more detailedand precise data. Then, given the large volumeof newly available data, they must employ datafusion and information processing equipment totranslate the new information into a form thatcan enhance operators’ situation awareness,i.e., their knowledge of the current status of allrelevant elements of the roadway environmentand trends toward future states. Situationawareness is a psychological construct thatcombines such processes as perception,cognition, memory and recognition, andinformation extrapolation and interpolation.

IVHS TMC’s will employ automation of routinedecisions and actions to help manage theoperators’ workload. The degree of automationwill be adaptable, providing operators andmanagers the opportunity to selectively engageor disengage automated systems according tomomentary workload levels and roadwaysituations.

Finally, IVHS TMC’s will provide new channelswith which to communicate information tovehicles and drivers. This may include directtransmission of data to in-vehicle displaysystems, presentation of brief messages onvariable message signs, and presentation ofmore complex information on short-rangehighway advisory radio transmitters.

TRAFFIC MANAGEMENT CENTER(TMC) OBJECTIVESThe overall mission of the TMC is to coordinateand facilitate the safe movement of persons andgoods, with minimal delay, throughout theroadway system. In order to perform thismission, the TMC must pursue the followingsystem objectives.“’

1. Maximize the available capacity of area-wide roadway systems.

The first objective of the TMC is to create themaximum effective capacity for existingroadways. By distributing the traffic loadspatially and temporally and managing queuesas they begin to form, the TMC seeks tominimize congestion and delay, and increaseeffective throughput. For special events, thismay include managing parking lotingress/egress, as well as inbound andoutbound traffic.

2. Minimize the impact of roadway incidents(accidents, stalls, debris).

Roadway incidents, particularly during highdemand times, have a significant impact ontravel times and create a threat to public safety.Reducing the effects of these incidents mayinvolve efforts on two fronts: reducing thelikelihood of incidents occurring and minimizingthe delays associated with incidents that dooccur. This includes planning and proceduredevelopment, as well as real-time trafficmanagement.

3. Assist in the provision of emergencyservices.

Active involvement of the TMC may facilitatethe provision of emergency services. Interactionwith emergency service providers (police, fire,medical, hazardous materials management)may include incident detection and verification,incident notification, coordination of responseswhere multiple services are needed, andmodification of system parameters to improvespeed or accuracy of emergency response.TMC support for emergency services is notlimited to roadway incidents. it may be neededin any situation in which the roadway is used byemergency responders (e.g., creatingambulance paths by clearing highways betweena disaster site and trauma hospitals as wasdone in Amsterdam after the 1992 El Al crashinto an apartment building).

4. Contribute to the regulation of demand.

The presence of maintenance activities, specialevents, or major incidents may create conditionsin which the overall demand exceeds theavailable capacity, no matter how efficiently thetraffic flow is managed. This objective requiresmotivating drivers to reschedule trips, reroutetrips, or take alternative modes of transportation.Regulation of demand may be short or longterm, reactive or proactive. It involvescooperative efforts with the media and othertransportation agencies.

5. Create and maintain public confidence inthe TMC.

In order to fulfill objectives 1 through 4, the TMCmust be perceived by the public as reliable andcompetent. The TMC must be perceived asproviding accurate and useful information to thepublic. Public relations efforts may beconducted in order to reinforce positive publicperception of the TMC.

It is important to note that these are theobjectives, as viewed by a panel of “IVHSvisionaries,” of an ideal IVHS-class TMCproviding an unusually broad range of services.The traffic management philosophies and needsof individual jurisdictions will govern their specificTMC implementations. One of the mostimportant observations made during the study ofexisting centers was the broad range of designand operating philosophies. While all centersshared some design and system aspects

(largely because of the limited number ofcontractors and vendors), no two had the sametraffic management needs or the samephilosophy of design and operation.

DESIGNING THE TMCThe design of complex systems like controlcenters can be based on any or all of a largenumber of factors. These factors include:

Published design standards andguidelines.

Accepted professional practices.

Designer’s styles and preferences.

Designer’s past experience.

Experience of other designers.

Lessons learned from past designs.

Detailed analyses of requirements.

Eagerness to try a specific newtechnology.

One of the best foundations for design of arelatively new class of systems is the body oflessons learned from the design,implementation, and operation of similarsystems. In fact, tort lawyers notwithstanding,analysis of past engineering errors and theirpotential consequences may be considered acrucial element in the evolution of engineeringdesign.(2) Some of the most important humanfactors studies have centered on identifying anddefining lessons learned from accidents andnear-accidents that can be related to theengineering design of the system.

The nascent IVHS-class TMC is a complexcommand, control, and communications (C3)center that is similar, in many respects, tooperations control centers used by the military toplan, direct, and monitor battles. It is similar tomodern process control centers in the chemicaland nuclear industries. It is also similar todispatch and control centers in other modes oftransportation. Because of this similarity,important design lessons for the IVHS TMC maybe derived from exploring design processes andoperational successes and failures of thesecenters.

2

It is to be expected, however, that the IVHSTMC will have most in common with the mostmodern and progressive of existing TMC’s. Themost valid view of the IVHS TMC of the year2000 timeframe can probably be derived byexploring near- and medium-term expansion andupgrade plans of these progressive existingTMC’s.

While increasing degrees of automation arebecoming feasible in the TMC, near-termdesigns will share most functions betweenhuman operators and automated subsystems.During some subfunctions, the operator willneed only to monitor the machine and, if itbecomes defective, disconnect and repair it. Inother functions, a more equal sharing ofworkload will be required. To ensure that systemoperators can effectively use the IVHS trafficmanagement system, that system must bedesigned to meet the validated requirements ofthe operators.

Recommendations and guidelines forrequirements-driven, user-centered design ofthe IVHS-class TMC are being developed.Visits to approximately two dozen TMC’s andother operation control centers in the UnitedStates, Canada, and Europe have provided aknowledge base of technology applications,operator functions and job descriptions, andlessons learned. A detailed series of analyses

has defined a set of TMC objectives, thefunctions necessary to meet those objectives,opportunities for full or partial automation of thefunctions, and operator performancerequirements and tasks within the partiallyautomated TMC.

This report summarizes the functions that mightbe incorporated into a generic IVHS-class TMC.It discusses how the functions are performed ina sample of existing high-technology TMC’s andhow they might evolve with near-termtechnology advancements. It provides humanfactors-related guidance for this evolution.Finally, it addresses a series of specific, crucialissues, including:

• Automation and user-computer roledefinition.

l Operatortraining.

qualifications , selection, and

• Methods for graceful evolution ofexisting centers.

• Design processes for displays andworkstations.

l Written documentation, procedures, andjob performance aids.

l Current application of human factorsinformation in the design process.

Section 2.THE COMPARABLE SYSTEMS ANALYSIS METHODOLOGY

The purpose of the comparable systems operation control centers, and their subsystems,analysis was to explore operation control that are similar to those that will be used in thecenters that are similar, in many respects, to the future in traffic management.generic, ideal TMC (ITMC). The threeobjectives were: (1) to define an appropriate set

The comparable systems analysis method,

of dimensions with which existing operationshown in figure 1, summarizes the methodology

control centers can be compared andused for the comparable systems analysis. The

contrasted, (2) to define the state-of-the-art inapproach consisted of two distinct phases, each

operation control center design with specialending ‘in a technical report. The initial phasedocumented features of a number of advanced

emphases on operator situation awareness andautomation, and (3) to explore lessons that have

control centers, assessed their comparabilitywith a notional ITMC, and recommended sites

been learned in the design and operation of for further detailed study.(3)

Figure 1. The comparable systems analysis method.

5

The second phase, documented in this report,involved a detailed study of the sites selected bythe comparability analysis and a number of othersites that were strongly recommended aspartially representing advanced TMC’s. Thesecond phase investigated the selected facilitiesin more detail, particularly in human factorsdesign features. Three of the sites examined inPhase I were included in Phase Il.

PHASE I. IDENTIFYINGCOMPARABLE SYSTEMSDuring Phase I of this task, a preliminary vision

of the design and configuration items of atypical, near-term IVHS-class TMC wasdescribed. Based on reviews of the literatureand discussions with traffic managementexperts, the following list of characteristics of theIVHS TMC was developed:

l Evolutionary, rather than revolutionary,departure from newer existing TMC’s.

l Borrows and adapts existingtechnologies now being introduced toadvanced command and controlsystems in military and civilianenvironments.

• Provides an environment of hybridautomation with tasks shared byoperator and machine and in which taskallocations may be revised in real time.The computer suite will have only limited

control authority without operatorassent.

Most of the sites chosen for the initial visits wererelatively progressive TMC’s that had adopted atleast one element of IVHS technology. Inaddition, a set of civilian and military commandcenters that shared some aspects of TMCdesign or functionality were visited. Siteschosen for the initial set of visits that were notTMC’s met the following criteria:

• Controlled a flow or process.

• Sensed and fused large quantities ofdata.

• Provided fused data through graphicaluser interfaces.

l Used (or could use) decision supportsystems.

During Phase I, half-day visits were conductedat the sites listed in table 1. During these visits,center management and operations personnelwere interviewed, providing an overview of thecenter design, setting, functions, dailyoperations, and emergency operations. Visitswere documented with still and videophotography (where allowed). The primarypurpose of Phase I site visits was to documentthe architecture, features, and operator tasks ofthe visited centers. Identification of humanfactors lessons was of secondary importance.In many centers, however, human factors issuescould not go unnoticed and, in such cases, theywere documented in the notes of the team.

Table 1. Phase I site visits.

6

In a parallel task, a different research team wasdeveloping the functional design for a future,IVHS-class TMC. The features of this notionalITMC were described to approximately the samelevel of detail as the documentation developedduring team visits to the real-world TMC’s.

Using a comparability assessment method, thesites most similar to the ideal TMC wereidentified and recommended for more detailedstudy.(5)During visits and interviews, a numberof additional sites were strongly recommendedby the managers and other experts with whomwe had met. Several of these were alsoselected for detailed study.

PHASE Il. IDENTIFYING LESSONSLEARNEDDuring Phase II, visits averaging 2 to 3 dayswere made to the sites listed in table 2. Duringthese visits, structured interviews wereconducted with managers and operators of thecontrol centers using a pre-prepared booklet ofquestions. These questions were based, in part,on the comparability dimensions identified duringPhase I. The questions covered such areas asthe following:

l Observation and analysis of a timeline ofactivity.

• Demographic information.

l Minimum qualifications for a TMC manager.

• Minimum qualifications for a TMC operator.

l Center layout (including recent or plannedchanges).

l Human factors of job handprocedures manuals.

books and

l Human factors aspects of control room,layout, and furniture.

l Human factors of operator displays.

l Human factors of operator controls.

l Human factors of system documentation.

l Allocation of manager and operator activitiesbetween tasks.

• Interjurisdictional coordination.

l Data archiving.

l Applications of automation.

l Human factors aspects of communicationsystems.

l Maintenance responsibilities.

l Center staffing.

• Large-screen displays.

l Console video monitors.

l Center design process, including contractorsand methods.

l Center environmental factors.

l Operators’ musculoskeletal complaints.

Still photographs were taken at each site.Copies of training documentation, procedures,standard forms, and checklists were alsoobtained, and their role in the operators’ taskswas documented. Where possible, operatorswere observed during normal operations andnotes were taken to describe their tasks, inputsand responses, roadway event history, and anysignificant successful or clumsy applications ofdisplay, control, or automation technology.

Human factors lessons learned in the designand operation of each site were thensummarized for inclusion in this document. Thefollowing sections provide the results andobservations of these site visits. Section 3provides overviews of all sites visited duringboth Phase I and Phase II; sections 4 through10 summarize the observations and lessonslearned; section 11 suggests provisional designconsiderations that are based on the lessonslearned.

l Sources of input information.

l Center configuration items (including recentor planned changes).

7

Table 2. Phase II site visits.

Section 3.SYSTEM OVERVIEWS FOR SITE VISITS

The overview for each command center visitincludes:

• Brief Description of the center, coveringthe command center’s mission and itsprimary functional areas of control.

• Unique Features of the center,emphasizing aspects of the setting.

l Human Factors Issues and Problems ofthe center, considering center mission,functional areas of control, and uniqueaspects of its setting.

These centers are presented in alphabetic orderbelow.

AIR COMBAT MANEUVERINGINSTRUMENTATION (ACM/‘)Brief Description: The function of this facility isto provide training, performance assessment,and debriefing for Air Force fighter pilots.During practice engagements, huge amounts ofdata concerning position, attitude, maneuvering,and switchology are collected. In order to makethese numbers meaningful, they are convertedinto various dynamic graphical displays that aremeaningful to the pilots, instructors, andreferees who will use them. The control roommust provide real-time displays of air combatoccurring many miles away in order to allow thereferee to enforce safety rules, score theeffectiveness of shots, and direct the aircraft.For research data collection, the system mustprovide the capability to allow the researcher toenter research data, initiate data collection andanalysis, enter comments into the data base,and terminate data collection. For debriefing,the data must be fused to support interactivedata visualization and the user must be able toflexibly visualize and understand the data.

Unique Features: A sensor pod is attached tothe aircraft that will be flying on the ACMI.Instruments on the pod detect aircraft attitude,maneuvering, and control activity, and radio thedata to the ground stations. Radar stations

track the aircraft positions and transmit the databy microwave. These data are fused at theACMI ground station to produce the displays.On the simulator for air-to-air combat (SAAC),system states are captured directly from thesimulation computers by the “strap-on”performance measurement computer system.Both training and research real-time datacollection, as well as additional off-line analysis,are supported within this system.

Human Factors Issues and Problems: Real-time monitoring and display of air-to-air combatrequires highly coordinated operator control andmonitoring. Situation awareness by trainers is asignificant factor in properly conducting trainingexercises, which are heavily dependent on thefusion of complex sensor data and voicecommunication. Debriefing of training exercisesaddresses important issues in evaluation ofcomplex performance data and providesflexibility in the reporting format. For example,the performance measurement system (PMS)supports both group and individual debriefingswith the ability to change point-of-view in thesortie. Maintaining past performance data, andpresenting fused data as informative datavisualizations, are important human factorsissues. A highly intuitive user interface requiresa minimum of training and paper documentation;it allows the user to devote maximum attentionto the displays. Many of these issues arerelevant for the current simulator development.

AIR MOBILITY COMMAND (AMC)TANKER AIRLIFT COMMANDCENTER (TACC)Brief Description: The mission of the TACC is tomaximize the efficiency of the AMC in its role ofmoving military cargo and personnel around theglobe. In addition, the TACC is a “ModelCommand Center” in which new approaches todesign and function are first implemented anddemonstrated before they are adopted AirForce-wide. Airlift requirements originate fromother DoD centers and are communicated toAMC Headquarters where they are planned and

9

scheduled. Mission requirements normally areplanned and scheduled by Current Operations,but some short-term requirements are handledby the operations center. The operations centerhas responsibility for tracking and directing allAMC aircraft that are in the air including crews,cargo loads, cargo spare capacity, plannedroutes and schedules, diplomatic clearances,weather, and relevant intelligence information.

Unique Features: The Global Decision SupportSystem (GDSS) is a replicated data base of allcurrent and planned AMC flights. Several keygeographic locations are equipped to usereplicates of the data base in the event that theprimary GDSS center is off-line. Various database utilities allow the system to be queriedfrom many different directions in order tosupport planning functions and enrouteoperations. A graphic display shows all AMCflights in progress (positions based on flightplans rather than real-time tracking). The TACCis staffed 24 h/d.

Human Factors Issues and Problems: Theglobal responsibility for military airlift of cargoand personnel presents unique problems incoordination between remote site facilities andthe TACC. Maintaining a global data base ofcurrent flight-related events involves manyissues of data validity and its security. Forexample, lack of real-time tracking impacts datavalidity and real-time intervention. The supportsystems for current and planned flights mayprovide useful information on operator decisionmaking and decision support systems potentiallyrelevant to Commercial Vehicle Operations(CVO) human factors requirements.

ALLIED SERVICES CORPORATIONDISPATCH CENTERBrief Description: Allied Services Corporation(ASC) is a major east coast trucking companyspecializing in the movement of newautomobiles from terminals (factories, sea ports,rail heads) to dealers. The Central DispatchCenter is responsible for planning, routing, andincident management for a fleet of over 1000trucks. The primary function is to minimize costand delay in the delivery of cargo, within thelimits of government regulation and union workrules, through effective central planning andcommunication. ASC fleet efficiency is a

complex combination of performancerequirements, including overall volume (bothnumbers of vehicles hauled and destinations),priorities of cargo, expected locations ofavailable trucks (controlled by planned routesand schedules), and trip load [both full andbackhauling (empty)].

Unique Features: Operators are responsible fordirecting drivers to their next terminal to pick upa load of automobiles. They receive from theterminals a daily list of automobile quantities andtheir destinations. Using this list, a map, and thedata base of driver assignments, drivers/trucksare assigned to pick up automobiles at theterminals. Assignments are based on routingefficiency and union rules. There are noautomated sensors; operators rely on telephoneconversations and route assignments to locateand track vehicles.

Human Factors Issues and Problems: Lackingautomation, operators use complex rules thatdictate the assignment of loads, routing, andincident management. Numerous opportunitiesexist for decision support systems andautomation in route planning and optimization.Such decision support systems may be usefulfor considering CVO human factorsrequirements.

ANAHEIM TRAFFIC MANAGEMENTCENTERBrief Description: The primary function of theAnaheim Traffic Management Center is topromote a smooth and delay-free entry andexodus of automobile traffic associated withspecial events such as conventions, footballgames, and concerts. As a secondary function,the TMC monitors and controls rush-hour traffic.System performance requirements includeplanning for special events, incident detectionand verification, traffic control, and coordinationwith other jurisdictions.

Unique Features: Custom graphics have beendeveloped for the display of multiple levels ofdetailed views from large geographic regionsdown to the intersection level of detailed view.A concave workstation design focuses attentionon the large-screen display and eight closed-circuit television (CCTV) monitors. Threeadditional 356-mm (14-in) graphics monitors arelocated in the control room with one mounted at

10

standing eye level on a movable arm near themidpoint of the concave workstation.

Human Factors Issues and Problems: TheAnaheim TMC represents innovative controlroom design concepts, computer graphics, andcomputer systems to support automation.Issues of integrating police department supportinto the centers’ activities have been addressed.Although multiple levels of views weredeveloped depicting roadway networks, onlyintersection and regional level views, such asaround the convention center or baseballstadium, are considered useful to operators.The convex workstation, as intended for multipleoperators, focuses attention on the large-screendisplay and eight CCTV monitors. Frequently,however, only a single operator is on duty. As aconsequence, the design, as intended formultiple operators, is no longer the best designsolution for a single operator since the convexsurface places other workstation displays andcontrols partially out of the field of view andreach.

AMSTERDAM PUBLIC TRANSITCONTROL CENTERBrief Description: The primary responsibility forthe Public Transit Control Center (PTCC) ispriority-based control of the trolleys and buses inthe Amsterdam area. The centralized facility,using the Operational Management System(OMS), controls trolleys and buses throughinterconnects to both the traffic signal systemand the mass transit traffic management center.

Unique Features: At numerous points on theirroutes, the vehicles pass over sensors in theroad, triggering an information exchange aboutthe vehicle identity, its schedule, and variationfrom the schedule. If the vehicle arrives lateover a sensor, it receives traffic signal priority(green lights) until it is back on schedule.Furthermore, information about the status ofeach vehicle is transmitted to the control centerthat contains numerous innovative graphicdisplays that allow operators to remain aware ofthe schedule performance of the system at alltimes. In addition, operators remain in radiocontact with each vehicle driver in order tohandle major and minor emergencies, dispatchmaintenance, or dispatch replacement vehicles.A complicated algorithm is used to schedule

vehicles during late evening hours so that atconnection points, the two connecting vehiclesarrive at the transfer point simultaneously.

Human Factors Issues and Problems: ThePTCC recently moved into a new facility. Thisnew facility is unique in that an ergonomicconsulting firm had a major design impact.Detailed task analyses were conducted on theoperators’ activities and ergonomists developeda series of scale mockups of the room andfurniture that was iteratively redesigned basedon operator comments. The resulting designwas observed to be highly functional.

AMSTERDAM SIGNAL CONTROLCENTERBrief Description: The Amsterdam SignalControl Center controls all of the traffic signals inAmsterdam. Within this responsibility are threeprimary functions. These include reducingcongestion by changing signal timing, detectingmaintenance problems with signal systems anddispatching service crews, and maintaining radiocontact with emergency vehicles and creatinggreen waves at their request. The operationscenter is in the process of moving from the oldcenter to a newly completed center.

Unique Features: In the Netherlands,particularly in Amsterdam, pedestrians andcyclists are given a higher priority than in NorthAmerica. Extensive bicycle-only trails existthroughout the Netherlands, includingAmsterdam, and in many cases there areseparate lanes and signals for cyclists, and mostpedestrian crossings at intersections aresignaled. Compliance to signaling is apparentlya significant problem. In Amsterdam, red lightsare ignored by large numbers of pedestrians andcyclists (some 60 to 80 percent). Detection ofcyclists at crossings creates a demand forgreen, but, given this non-compliance, in manycases the cyclist has already reached the otherside of the road when the light turns green(which translates into lost time in the signaltiming phases). A complex priority-basedcontrol system for trams and buses is used inAmsterdam. It is intended to minimize delays byoverriding the current intersection timing in favorof the transponder-equipped trams and buses.This is discussed in detail below, under theAmsterdam Public Transit Control Center.

11

Human Factors Issues and Problems: Like allHolland centers we visited, there weresignificant glare problems on the CRT’s. Thiscan be attributed to Dutch laws and buildingpractices that require the use of large amountsof glass to conduct sunlight into the workplace.System integration impediments were also notedin workstation design and operator interfaceusability. Two system vendors failed to providenecessary integration of their products, whichhas required two separate user interfaces ondifferent hardware platforms. Failure to identifysystem requirements has resulted in the loss ofseveral key design features that were present inthe old control room design, but were not carriedforward into the new design specification. Radioand traffic operators sat next to each other inthe old control room, allowing a “party line” toexist between them that enabled mutualresponsiveness to changing needs. In the newcenter, these operators are located at too greata distance to hear each other’s communicationtraffic. Placement of video monitors in the oldcenter was below the map display at aconvenient eye level. In the new center, thesemonitors are rack-mounted next to the mapdisplay at a less convenient location (with newlyintroduced glare problems).

AUTOMATED TRAFFICSURVEILLANCE AND CONTROLSYSTEMBrief Description.= The primary function of theLos Angeles Automated Traffic Surveillance andControl (ATSAC) system is to promote a smoothand delay-free flow of automobile traffic on themajor arterial streets of Los Angeles. Systemperformance requirements at ATSAC includeincident detection using loop detectors andincident detection algorithms, incidentverification using CCTV systems, and control ofstreet traffic using canned timing plans andadaptive control.

Unique Features: The control room layout wasrather primitive, with computers and displaysstacked roughly according to geography,encircling the operators. ATSAC is equippedwith the Urban Traffic Control System (UTCS),enhanced to implement signal timing plans.These provide automatic adaptive control in the“critical intersection” and “traffic-responsive”modes. Occasionally, “manual override” mode

is used to allow the operator to respond tounusual traffic conditions, such as sports eventsor holiday shopping, by creating an customtiming plan. ATSAC interfaces well with CaltransDistrict 7 and emergency-response personnelvia radio, computer-aided dispatch (CAD), andtelephone.

Human Factors Issues and Problems: ATSAC’sevolutionary control room design allows flexibilityand that allows responsiveness to changingneeds and requirements by operators. Thecurrent design does not consider glare or theadvantages of ergonomically designedworkstations and seating. Self-labeling ofworkstation displays and controls and theirarrangement are possible indications ofconfusion resolved by the additional labeling.Regional geographic system control is weaklysupported by the efficient arrangement ofcomponents and may not support the efficientcontrol of ATSAC by a single operator. SinceATSAC can handle multiple traffic incidentsconcurrently, voice communication and resourcesharing among multiple operators should beinvestigated.

CALTRANS DISTRICT 7 TRAFFICOPERATIONS CENTERBrief Description: The primary function of theCaltrans Traffic Operations Center (TOC) is topromote a smooth and delay-free movement oftraffic on freeways in the Los Angelesmetropolitan area, including coordinating initialresponse to freeway incidents. The systemperformance requirements at the Caltrans TOCinclude 24-h traffic monitoring and incidentdetection; incident verification; traffic flow controlusing ramp metering and variable messagesigns; and communication via radio, CAD, andtelephone.

Unique Features: A large wall map of theregion’s freeways contains color-coded light-emitting diodes (LED’s) representing each loopdetector array. Lights turn yellow or red toindicate levels of congestion. This map is saidto be heavily used. It interfaces well with localmunicipality TMC’s (e.g., ATSAC and Anaheim)and with other Caltrans districts. The presenceof a California Highway Patrol (CHP) liaisonenhances the efficiency of emergency response.Operators are formally trained in a continuing

12

education course at California PolytechnicInstitute and State University. CHP officers arecross-trained to serve as operators whenrequired.

Human Factors Issues and Problems: With alarge mixed staff operating in an absence ofautomation in such a busy center, Caltransprovides a wealth of potential human factorsissues in communication needs of operators foreffective interjurisdictional and interagencycommunication. The impact of automation onoperator communication activities is beingconsidered as designers develop plans for theirnext generation TOC. Only fixed light and colorchanges can be varied on the big-board display;fusion of any other related data to the big-boarddisplay is not possible. Maintenance and updateof the big-board display with a static map islabor-intensive, especially for changes in mapdetail (which require an outside contractor).

CHICAGO IDOT COMMUNICATIONSCENTERBrief Description: The primary function of theIllinois Department of Transportation (IDOT)Communications Center (ComCenter) is theassignment of emergency and maintenancevehicles and specialized crews to areasthroughout District 1 to maintain efficient flow oftraffic on the state-maintained roadways andInterstates in the Chicago metropolitan area. Asa communication hub, incident-related cellulartelephone *999 calls from motorists are receivedby a separately maintained exchange,summarized, and forwarded to the ComCenter.Communication links are also maintained withthe Traffic Systems Center. The TrafficSystems Center provides travel time for majorroad segments in the freeway system, andcongestion information disseminated by acomputer link and through a CCTV summarizedcongestion map of the Chicago area. When theADVANCE traffic information center becomesoperational, it will also provide its data to theComCenter as well.

Unique Features.= The ComCenter is staffed 24h/d with the highest level of staffing during themorning (6:00 a.m. until 9:00 a.m.) andafternoon/evening (3:00 p.m. until 7:00 p.m.)rush hours. The ComCenter has an unusuallybroad range of responsibilities, including

computer-based remote control over stormwater pump stations, highway lighting, ahighway advisory radio network (both AM andFM stations), changeable message signs, andthe reversible lane controls of the KennedyExpressway over radio and telephone links. Toassist in handling the large number of telephonecalls received, the ComCenter has anautomated telephone answering system routingmost calls coming into the ComCenter. Therealso exists the option of directly contacting anon-duty ComSpecialist.

Human Factors Issues and Problems:interjurisdictional issues for 300 municipalitiescan result in the four corners of an intersectionbeing under different jurisdictional control,complicating maintenance and incident-relatedinterventions. There are problems in controlroom environmental factors, such as loudness ofadjacent operator conversations (especially inhigh workload-related radio-telephonecommunications for maintenance dispatch), andambient lighting levels. Dissemination of routineHighway Advisory Radio (HAR) travelerinformation is complicated by providing bothcongestion and travel time information tomotorists. These two forms of travel informationare broadcast on separate automated radiostations using two different technologies(synthetic voice versus sampled speech fromlive operator) and hardware systems. Voicegender and speaker coding guidelines need tobe established for use with the synthetic voicesystem. Frequent gender and speaker changespotentially distract from the message content.

CHICAGO IDOT TRAFFICINFORMATION (ADVANCE) CENTERBrief Description: The ADVANCE program willprovide traffic information to approximately5,000 equipped vehicles in the metropolitanChicago area. The control room being designedis currently conceived to be a single consolewith one or two workstations.

Unique Features: With approximately 5,000vehicles in the ADVANCE evaluation, this will bethe largest test site to date of IVHS smartvehicles.

Human Factors Issues and Problems: As afeasibility study site, the designer’s highestpriority is to provide traffic information and

13

control to the vehicles, but not to examinecontrol room design issues. Potentially, thisapproach may prove to impact the intendedfunction of the center by poor implementation ofperson-machine interface.

CHICAGO IDOT TRAFFIC SYSTEMSCENTER (TSC)Brief Description: The primary function of theIDOT Traffic Systems Center is freeway traffic-surveillance and control for the metropolitanChicago area. Freeway and ramp monitoringare included in the summarized traffic flow datapresented in a graphic congestion display of themetropolitan Chicago area. This data iscommunicated to a variety of other agenciesand universities, including cable TV. A text-based version of travel time over major roadsegments is also communicated to agenciesover a computer link. A series of changeablemessage signs (CMS's) are controlled from theTSC that inform motorists ofcongestion/incident-related problems.

Unique features: One of the largest trafficcenters, the TSC monitors 190 km (118 mi) ofroadway on eight Chicago area expressways.Since it has been in operation nearly 30 years,the TSC is an evolutionary blend of both old-outdated and state-of-the-art technology.

Human Factors Issues and Problems: Roadwaysensors and effectors are the driving technologyforce in the evolution of the TSC. The TSCpresents a difficult challenge in evolving from oldtechnology to new technology. Many humanfactors issues, such as noise level and systemintegration, are solved by application of currenttechnology. However, given the TSC’s efficientperformance, from a cost-benefit perspective itis difficult to justify massive upgrades.

FM AIR TRAFFIC CONTROLSYSTEM COMMAND CENTER(ATCSCC)Brief Description: The primary role of theFederal Aviation Administration’s (FAA’s) airtraffic control system command center(ATCSCC) is to minimize the number and lengthof air traffic control-related delays of enrouteaircraft operating under instrument flight rules(IFR). The air route traffic control center

(ARTCC) enters all IFR flight plans into a database. The ATCSCC tracks position and statusof all enroute aircraft and correlates thesefactors with flight plans/clearances for airports inthe specific region. It projects arrivals at airportsand predicts over-capacity periods. Using theEnhanced Traffic Management System (ETMS),the controller develops plans to reduce thenumber of aircraft approaching these airports atthe critical time, and then negotiates andimplements these plans.

Unique Features: A substantial amount ofautomated data fusion and analysis is performedby the ETMS on the raw aircraft position, flightplan, and weather data. These combined andfused data are used to identify congestionproblems, produce potential solutions, and carryout those solutions (with the permission of theoperator). Operators are assigned control overgeographic divisions. The geographic UnitedStates is divided into three regions, - east,central, and west. On-site weathermeteorological data are available from aNational Severe Weather Center staffed by twometeorologists. Except for the meteorologists,all staff members are experienced air trafficcontrollers on loan from air route traffic controlcenters.

Human Factors Issues and Problems: The highlevel of automation in fusion of data addressespotential issues in display of complex detaileddata from large geographic areas. Operatorworkload and a large and varied staff raisepotential problems in operator roles andnecessary training. Integration of weather datawith traffic control-related information addressesimportant issues in operator decision makingand decision support systems. The decisionaids and support systems appear to be well-designed from a human factors viewpoint.

FLORIDA FREEWAY MANAGEMENTSYSTEMBrief Description: The primary function of theFlorida Freeway Management Center (FMC) isto promote a smooth and delay-free flow oftraffic on the freeway system of metropolitanOrlando. Primary requirements are detection,verification, and clearing of incidents on thefreeway system.

14

Unique Features: No usable automation wasevident at this point. Some automatedcapabilities are planned for later implementation.The FMC is staffed only during morning andevening rush periods. Six quad-split televisionmonitors allow simultaneous viewing of all 18CCTV cameras on the 17.7-km (11-mi) freewaysystem. Florida Highway Patrols’communication center shares the control roomwith the Florida Freeway Management System.

Human Factors Issues and Problems: Thiscenter is hampered by poor user-interface andworkstation design, reflected in human factorsand ergonomics issues in system ease of useand monitoring of workload. For example,operators monitor six quad-split video monitorsmounted near the ceiling at an acute angle withindividual operators monitoring northbound andsouthbound lanes during peak traffic hours.Frequently used features of the system are notdesigned for accessibility through the user-interface design. Interjurisdictional problems,such as the overlapping areas of coverage andredundancies in function, are not efficientlymanaged between the Metro Orlando TMC andFMC.

MINNEAPOLIS DOT TRAFFICMANAGEMENT CENTERBrief Description: The primary function of theMinneapolis DOT Traffic Management Center(TMC) is to promote a smooth and delay-freeflow of traffic on the freeway system of themetropolitan Minneapolis-St. Paul area. TheTMC controls traffic using adaptive rampmetering. It monitors traffic in order to identifyareas of congestion and possible incidents.Units in the field and, when necessary,emergency response providers are alerted tosuch incidents. The Motorist Information Systemprovides the drivers with information aboutpotential congestion and delays.

Unique features: The Minneapolis TMC isunique in its reliance on closed-circuit television(CCTV) displays. With 96 monitors for 108cameras, it is almost possible for each camerato have a dedicated monitor. An automatedswitching system moves camera images amongthe monitors on the two video walls. Additionalbanks of 229-mm (9-in) monitors are beingadded for the traffic radio broadcaster’s use in

periodic comprehensive broadcasts. A 1829-mm (72-in) BARCO rear-projection monitor candisplay either graphical map data or CCTVimages. Because the number of CCTVsurveillance cameras exceeds the number ofCCTV channels and monitors, an automatedswitching process is in place to ensure that eachcamera receives equal priority. For the variablemessage sign (VMS) system, selection andpresentation of “canned messages” is partiallyautomated. Control of ramp metering is traffic-adaptive with manual override.

Human Factors Issues and Problems: Theheavy reliance on CCTV’s and the switching ofimages between monitors requires significantoperator monitoring in support of incidentmanagement. Given that the monitoring of thislarge number of CCTV’s well exceeds humanfactors recommendations, the use of thesemonitors by operators represents an importanthuman factors issue and deserves additionalexamination. Methods for data fusion andautomatic camera selection need to beexamined. It may be that operators are utilizingstrategies for reducing the complexity of thismonitoring task. Both automationenhancements and decision support should beconsidered for these operators. There is noTMC guideline for the team composition on ashift, in terms of operator roles and thenecessity of having a shift supervisor on duty.Currently, shifts work without a designatedsupervisor on duty and vary in the compositionof the experience and background of theoperators.

METRO ORLANDO TRAFFICMANAGEMENT CENTERBrief Description: The primary function of theOrlando Traffic Management Center (TMC) is topromote a smooth and delay-free flow of trafficon the streets of metropolitan Orlando. Asecondary function has been to serve as thecontrol center for the TravTek automobilenavigation system operational demonstration,which has been completed and is currentlyunder review. The TMC maintains signals in aneffective timing plan, and adjusts the timingplans as necessary. It identifies, verifies, andhelps clear incidents on Orlando’s streets. Italso helps to ensure efficient operation of theTravTek demonstration.

15

Unique Features: Metro Orlando, as the site for Unique Features.= Lane-based control requiresthe TravTek demonstration, has evaluated dense sensor arrays and numerous gantryprototype automation and communication needsnecessary to support the TravTek system,including fusion of data from information sourcesof varying reliability to provide real-time incidentinformation, and transfer of updated link traveltimes to the vehicles.

Human Factors Issues and Problems:Orlando’s state-of-the-art TravTek automobilenavigation operational test system provides asource of potential human factors issues on thisemerging technology. Key issues includeintegration of TravTek functions into generalTMC operational controls and displays, as wellas the integration of TMC communications withcar drivers. Large, static network maps used inthe center are labor-intensive to modify whenupdates are needed in the display of theinfrastructure. There is a significant need forcoordination with the Florida FreewayManagement Center, but inefficientcommunication practices sometimes hinder thiscoordination.

NETHERLANDS MOTORWAYCONTROL AND SIGNALING SYSTEMBrief Description: The primary function of theUtrecht and Amsterdam Freeway TrafficManagement Centers is to provide lane-controlmanagement to promote a smooth delay-freeflow of traffic. An additional function, with thesupport of the Motorway Control and SignalingSystem (MCSS), is to reduce the frequency andeffects of both primary and secondary accidents.Messages reflecting reduced speed limits,closed lanes, and required lane changes aredisplayed for each lane of the freeway onoverhead gantry-mounted signs. One significantcharacteristic of the European approach is anemphasis on incident prevention, rather thanincident detection, and management aspracticed in North America. European TMC’shave little involvement in incident managementwhere incidents are defined as things orevents on the roadway that cause congestion.Rather, European managers suggest that theyare more concerned with preventing theincidents from happening (by actively controllingcongested traffic) than in managing them afterthey happen.

mounted signs over each lane of the freeway.Located at approximately 1/2-km intervals is alane-control system consisting of an array ofdouble-loop detectors, a control box, and agantry sign. The loop detector array detectschanges in traffic speed that signal thebeginning of a congested situation. Outstationcontrollers located near the roadway sense setsof adjacent loop detector arrays, automaticallyformulate the appropriate pattern of messages,and request permission from a central computerto present the selected messages on thegantries. The central computer will eitherconsent to the changes or will modify theoutstation’s recommendation to reflect thesituation at other outstations or messagesmanually entered by the operator. In the eventthat control communication is lost betweenoutstation and central control, the outstationautomatically presents the messages on itsgantries. This form of “graceful degradation” isa designed feature of the control system. Also,because much of the analysis is done locally,only a modest amount of data needs to becommunicated between the outstation and thecentral computer.

Human Factors Issues and Problems: Controlof the lane signs is highly automated because ofthe complexities created by the rules that governthe proper signing of reduced speed limits,closed lanes, and required lane changes. Onthe initial installation IO years ago, a singleoperator was required to consent to eachchange in sign pattern. Sign changes occurredat the rate of up to 200/h, generating anunnecessarily high workload. Now, all signingdecisions are initially made by the automatedsystem based on the detector data, and cannotbe easily overridden by the operator.Complexities of automation in operatormonitoring and supervision have led to thedesign of alarms that signal the operator ofpotential needed interventions. However, theoperators are often responding to false alarms.Troubleshooting is assisted by selectiveactivation of CCTV monitors that have coverageover the trouble area causing the alarms.Providing the appropriate information foroperator-based troubleshooting is extremelydifficult to pre-plan and highly intelligentautomation approaches, such as adaptive

16

automation, continue to be impractical toimplement.

TORONTO 401 FREEWAY TRAFFICMANAGEMENT CENTERBrief Description: The primary purpose of theToronto Freeway Traffic Management Center(COMPASS), Ministry of Transportation ofOntario (MTO) is 24-h freeway trafficsurveillance and control for the 401 freewaysegment in the metropolitan Toronto, Ontarioarea. The design goal for this facility is to beable to run in as automated a manner aspossible. Currently, changeable message signs(CMS), response plans, and media informationdissemination are partially to fully automated.System evaluation has included examination ofsystem failure rates and response times ofoperator acknowledgment.

Unique Features: Twelve lanes of roadway, sixin each direction for 36 km (23 mi), are dividedinto a collector/distributor and express lanes ofroadway. This roadway is one of the mostheavily traveled and complex roadways in NorthAmerica. An extensive collection of incidentresponse plans has been assembled toautomate the response process. The categoriesof response plans range from the generalincident response process, vehicle breakdown,reportable accident, non-reportable accident,and hazardous goods spill, to abandonedvehicles. When a response plan is activated, adescriptive message is composed and added tothe list of notices currently being faxed to thesubscription base of the media and otheragencies, and is also sent to the appropriateCMS’s for roadway display. A separatecomprehensive CMS message design andtaxonomy has also been developed. This CMSmessage design ties a given CMS’s location to apre-defined portion of the roadway downstream

of the CMS, referred to as a congestionmanagement zone (CMZ), so that messages arebased on current traffic conditions within theCMZ.

Human Factors Issues and Problems: Animpressive level of automation has beenimplemented and is being considered in theCOMPASS center. The level of automation inthe issuance of messages on CMS’s is highlyprocedural in form, requiring a complexquestion-and-answer dialog to issue a messageand log an incident. Such an evolvingautomation environment presents difficultchallenges in developing effective humaninteractions with automated systems. Thedesign goal for such interactions should be toenable brevity in the required interactions, whileretaining enough operator control for the propersupervision of actions. Requiring graphicsdisplay and command entry through differentinterfaces presents human factors problems.The operators’ preference for monitoringcameras for incident detection, rather thanrelying on the capabilities of automatic incidentdetection using the computer graphics-baseddisplay system, may be a symptom of thisunusable system design.

Some management problems have been noteddue to the unique organizational structure. Thesystem design and engineering group who, infact, make most of the day-to-day proceduraland equipment decisions are in a differentbranch of the Transportation Ministry than theoperators and, therefore, have no directauthority over them. A hiring freeze preventedhiring of a TMC Supervisor, so the operators areextremely autonomous. The structure impedesthe introduction of revised procedures andmethods.

17

STATE-OF-THE-ART

OPERATIONAL REQUIREMENTSOn the basis of interviews with a group of IVHS“visionaries,” a set of operations that would beperformed by an ideal TMC of the next decadethat provides a broad range of trafficmanagement services was defined. That studyconcluded that the full-service IVHS-class trafficmanagement system must perform all of theoperations listed in table 3 for data sensing andinput, TMC performance, and TMC output.

One major purpose of the present task was tostudy and assess the state of the art in each ofthe operational requirements of the visionaryTMC in order to validate or calibrate the realismof that vision. Each of these operationalrequirements was found to be performed in oneor more control centers, although, in manycases, the technologies and procedures aremore primitive than those of the visionary TMC.In this section, each of the operationalrequirements is discussed in detail, including thevision, current technology and practice, andnear-future prospects.

MONITOR CURRENT TRAFFICCONDITIONSTraffic condition sensors can be classified intotwo main types on the basis of their data input tothe TMC, information calculated from (primarily)binary raw data and information in the form ofvisual images. Binary sensors are, in effect,switches that register “on” when a vehicle is intheir zone and “off’ when no vehicle is present.Visual sensors such as closed-circuit television(CCTV) under remote control from the TMCprovide a direct view of the roadway and arewidely used for confirmation of traffic conditions,verification of incident location, anddetermination of incident severity.

Roadway sensors report data such as trafficflow rates and speeds. The sensors must beable to measure traffic speed as well as trafficvolume. In key areas of the roadway system,the sensors must be capable of providing thesedata by lane and by vehicle size. Currently,

Section 4.OF IVHS TMC TECHNOLOGIES

these sensors consist of inductive wire loopsembedded in the pavement that detect thepassage of automobiles by registering variationsin the surrounding magnetic fields. By using apair of the loops a known distance apart, it ispossible to compute the speed of each individualvehicle. Generally, binary occupancy data fromloop detector arrays are fused for display on agraphical user interface. A measure of thetraffic “performance” at a given sensor locationcan be determined by speed and/or volume oftraffic. These performance data are thendisplayed as a series of histograms or, moretypically, as color-coded roadway segments on amap display.

Other traffic detection technologies, such asradar, ultrasound, and visual image processing,are being tested and may be introduced to theroadways in the future.

CCTV monitors are an important feature ofTMC’s in North America and somewhat less soin Europe. These sensors are not routinelyintended as a primary source of informationabout traffic flow, but are intended to be used assecondary sources to support traffic and incidentmanagement, and to check the status of signalsand signs. While older installations havegenerally used black and white CCTV becauseof its better performance under low illumination,color CClV is generally said to be moreeffective during daylight hours and is becomingthe standard in newer centers. In many centers,banks of dozens of monitors surround theoperators. Operators can take control ofindividual cameras using button boxes orjoysticks to pan, zoom, and focus. In the nearfuture, automated aiming and zooming on sitesof suspected incidents will be available, relievingthe operator of this manual control function.

A third data source; instrumented probevehicles, including rental vehicles and publictransport vehicles, report point-to-point traveltimes and incidents. Several demonstrations ofthis technology have been performed in theUnited States and in Europe. Typically, thelocation of each member of a fleet of probe

19

Table 3. TMC operational requirements.

vehicles is determined by a global positioningsystem (GPS) receiver and is transmitted to aTMC computer. The amount of time it takes avehicle to move between any two of a set ofdefined points can be used to calculate vehiclespeed and estimate the level of congestion.Selection of appropriate vehicles to serve asprobes is a major issue. For example,emergency vehicles are a poor choice becausethey typically travel significantly faster than thetraffic stream. Rental vehicles may not be thebest choice since they typically travel differentroutes than the urban commuter. Probe data isonly a valid measure when probes arerepresentative of the class of vehicles they aresupposedly sampling.

MONITOR CURRENT ROADWAY

System monitoring devices and processes areavailable that identify sensor and effectormaintenance needs such as burned-out bulbs intraffic signals and malfunctioning loop detectors.Such technologies exist today, but emergingtechnologies will increase the credibility of themalfunction reports. A major percentage ofcenters now depend on complaints from thepublic as a primary source of information onmaintenance needs.

Visibility sensors are also available to detect theonset of fog and to warn drivers and TMCoperators. One system, currently implementedin Europe, detects decreased visibility andautomatically illuminates a warning sign andreduces the speed limit to appropriate levels. Aprototype of a similar system has been installed

CONDITIONSIn the future TMC, road condition and weathercondition sensors will monitor precipitation andpredict its effects on traffic conditions. Thesesensors will also monitor the friction coefficientsof dry pavements, and provide early detection ofworn pavements that need resurfacing. Someof these sensors will be maintained at fixedlocations, such as on bridges that are likely tofreeze. Others may be mobile sensors, perhapsattached to maintenance vehicles. Althoughnot yet widely used in the United States,embedded road sensors now exist that canmeasure and transmit data on pavementtemperature, air temperature, wet pavement,and salt content of pavement moisture.

by the Tennessee Department of Transportationand another is being installed for testing in anarea of frequent fog near Adel, Georgia.

IDENTIFY LOCATION OF INCIDENTSIn the visionary TMC, an incident detection andlocation system (IDLS) will receive data fromroadway sensors and from external data links.Its purpose will be to detect and verify thepresence of incidents on the roadway system,and to determine the exact location of theincident. In some cases, the first indications ofan incident will be abnormalities in traffic flow inthe immediate area. In other cases, especiallywhen traffic volumes are light, the firstindications may come from other sources, suchas calls over cellular phones that subsequently

20

result in an entry into a dispatch data base.Incidents, as defined by the IDLS, will includenot only traffic problems, but also systemmalfunctions such as traffic signal outages.Upon detection of a probable incident anddetermination of its apparent location, incidentverification is attempted. The IDLS will provide apreliminary estimate of incident severity(perhaps simply yellow or red) based on thedata available.

Currently, initial indication of a freeway incident(accident, stall, debris spill, etc.) may come fromone of three general sources, - loop detectordata; closed-circuit television image; or a radioor telephone report from police, media, or thepublic.

On heavily instrumented highways, earlyindication of a problem can be seen in the datafrom loop detectors. An incident may firstappear as unexpected congestion that results ina decrease of speed and volume of traffic.Incident detection algorithms exercise the loopdetector data looking for patterns that correlatewith incidents, typically, high-occupancymeasures. Obviously, with the inherentvariability in “normal” traffic flow, it is difficult todistinguish between a “normal” variation and avariation due to an incident. Performance ofthese algorithms, therefore, is generally poor interms of the percentage of incidents detectedand the time required to detect them. Onecenter found that incident detection performancecould be improved by using custom-developedincident detection algorithms that:

• Utilized traffic volume (vehicles per unitof time) as well as loop occupancy(percent of time a loop is “on”).

• Screened the data for indications ofstuck-on loops that artificially inflateoccupancy data.

l Tuned down threshold settings(algorithm values at which an incidentis assumed) upstream of majorfreeway interchanges during peaktraffic.

l Adjusted thresholds for seasonalchanges.

The philosophy of some centers, especiallythose outside of the United States, is thatunexplained traffic congestion, itself, represents

a problem that must be managed. Therefore,the incident detection algorithms have no “falsealarms” because they always represent aproblem related to congestion that needs to beaddressed.

Some centers make heavy use of CCTV assetsto make initial detection of incidents. Asdiscussed above, managers often believe thatthe CCTV systems are overused to thedetriment of the incident detection systemsbased on loop detector data.

In addition to its organic sensor assets, manyTMC’s receive data from a number of externalsources. Voice communications are receivedfrom the traveling public via telephone (includingcellular phones and roadside emergencyphones) and citizens band (CB) radios. Theprimary use of these data sources is to obtainreports of incidents. One midwestern U.S. TMCplaced CB radio receivers at key locations nearthe roadway. These could be selectivelymonitored by telephone in the TMC.Unfortunately, the poor quality of thesetransmissions rendered this innovationimpractical for monitoring these local radiotransmissions. Voice communications alsocome from public safety authorities, includingradio channels used by police, fire, andambulance services, as well as telephones.These channels are used to obtain reports ofincidents, to verify incident location and severity,and to help coordinate incident response.

CLASSIFY TYPE AND SEVERITY OFINCIDENTSThe ideal TMC will obtain information regardingthe nature and severity of the incident in order toselect proper responses. The amount andduration of the incident’s effect on roadwaycapacity must be accurately estimated.

Current TMC’s obtain this information from foursources, - CCTV, reports from individuals at theincident site, “eye-in-the-sky” traffic helicopters,and commercial radio and television. The newsmedia receives its reports from much the sameresources. Several TMC’s currently track typeand severity of incidents as part of theirreporting process for incident management.Characterization of the type and severity hasalso been found to be useful information tocommunicate to maintenance units, as is

practiced in Amsterdam, to initiate the requestfor repair. Maintenance units are given type andseverity information, with classificationcategories indicating the degree of damage tothe infrastructure, such as failures in intersectioncontrollers, burned-out lights, cut communicationlines, and downed poles.

IDENTIFY LOCATION OFEMERGENCY RESOURCESIn the ideal TMC, computer data links aremaintained with public safety agencies,commercial railroad companies, commercialvehicle fleet operators, and roadwaymaintenance/construction providers. These willinclude dispatch data bases of police, fire, andambulance services. Any dispatch of a vehicleby these agencies will be entered into the database as part of the act of dispatching, and willbe immediately available to the ITMS.

Currently, TMC’s obtain the location ofemergency resources by radio, telephone, andoccasionally by computer-aided dispatch (CAD).Often, multiple radio channels must bemonitored simultaneously because each agencymaintains separate systems. Programmable,multi-channel, multi-position, integrated radioand telephone control systems are nowavailable and in use in some TMC’s that simplifythe monitoring of radio and telephone traffic.These systems address many of the issues ofcommunications integration necessary forimproving efficiency and reducing thecommunications workload. For example, thesesystems include a CRT-based summary ofoverall channel traffic status that can bedisplayed by caller’s name (if equipped withautomatic number identification). Among itsfeatures, it allows dedicated speaker selectionfor high-priority channels, individual channelvolume control, predetermined grouping ofchannels, and automatic transmittal to the lastcall received. In addition to vehicle identity,vehicle position is currently tracked in somecenters by knowledge of assigned routes andtimetables, and in the near future will be trackedby one or more methods of automatic vehiclelocation, such as Global Positioning System(GPS) satellite-based technology; deadreckoning systems; terrain-following technology,such as ETAK; terrestrial radio location pagersystems that use triangulation, such as Lo-jack

or Teletrac; or mobile radio modemcommunication technologies. Besidestechnology advances is the need for newapproaches for increasing transmissionbandwidth, since large metropolitan cities suchas Chicago have saturated current availablebands of the electromagnetic spectrum, such asmicrowave. The FCC has near-term plans toreassign the frequencies presently used toincrease the number of available channels.Moving frequencies closer together will affectboth transmitters and receivers because ofresulting changes in the bandwidth filters on thereceiver for channel reception and therequirements and restrictions in bandwidth of theactual transmitted signal. In fact, somesignaling strategies may also be renderedobsolete if they have been relying on signalenergy that is outside of their assignedbandwidth (but currently unused). It isanticipated that these changes will render someof the current equipment obsolete in tuning tothe new assigned frequencies (old equipment inthe worst case would pick up adjacentchannels). At one midwestern TMC they areanticipating this FCC change as a stimulus forcomplete replacement of presentcommunications equipment.

MONITOR CURRENT ANDPREDICTED WEATHERInclement weather, such as rain or snow storms,can have a greater impact on roadway capacitythan any other factor. In the ideal TMC,National Weather Service forecasts and datafrom other meteorological information serviceswill be used in anticipating or locating weatherconditions that will impact driving conditions.

A few TMC’s now maintain displays of wide-areaweather radar returns indicating strong stormcells and precipitation for use in weatherprediction. Typically, TMC personnel are nottrained to use this weather data to its fullpotential. Subscription weather services arenow commercially available that provide CRTweather displays with sufficient resolution andaccuracy to locate thunderstorm cell boundarieswith an accuracy of a few tens of meters. Thesecan be computer-overlaid on a local area map toindicate specific roads currently receivingprecipitation.

PREDICT DEMANDThe ideal TMC will have a capability to predictexcessive demand in time to identify andimplement strategies to reduce the demand toacceptable levels.

Currently, the experienced TMC operator is ableto predict and prepare for cyclical demand peaksand peaks in demand resulting from specialevents largely on the basis of experience. Atone center, a weekly meeting is held with TMCmanagement, traffic police, and representativesof all major traffic generators (stadium,coliseum, amusement parks, convention, andtrade show venues). During this meeting,special event schedules are exchanged,predictions of the level and volume of extratraffic are presented, and traffic and parkingmanagement plans are developed.

Little automation is currently used to support thisfunction. However, the Support SystemsContractor’s (SSC’s) Integrated ModelingEnvironment will have the capability to usehistorical data and traffic simulation models topredict demand.

MONITOR PUBLIC CONFIDENCETraffic management systems can be only aseffective as the public’s confidence in them.The ideal TMC will continuously track measuresof confidence using solicited (survey) commentsand unsolicited (e.g., complaint letter)comments.

Current practice in most TMC’s includes formalreporting and assignment of work details basedon complaint calls received at the TMC by thepublic. Frequently, TMC’s rely on publicreporting of problems with intersection timingand other signage problems because of the highcost for maintenance crews to continually reviewthe integrity of the infrastructure.

PREDICT TRAFFIC CONDITIONSIn the ideal TMC, a predictive traffic modelingsystem (PTMS) will use data from the roadwaysensors, origin-destination (OD) data fromvehicles currently in the roadway system, dataabout special events traffic, and data aboutincidents to predict traffic flow a few minutes intothe future (from 5 to 30 min). This is done sothat preemptive actions can be taken well before

an actual problem arises. The current rampmetering and traffic signal timing patterns in useby the Urban Traffic Control System (UTCS) willbe incorporated into the predictive model.

The PTMS will be able to predict problems intraffic flow by using a model of current trafficconditions, taking into account incidents,available OD data, and advisories currently ineffect. The model will contain expected baselinelevels of traffic. The baseline will be time- andday-dependent, and will be adjusted for weatherconditions. Many of the specific components ofthis baseline will be provided by historical datarecorded by the TMC. The PTMS will alsopredict the response of travelers to theimplementation of traffic controls by the ITMS.For instance, it will predict the percentage ofpeople that will take an alternate route when it issuggested on a variable message sign (VMS),and use this prediction in its forecast of futuretraffic conditions.

Support systems with functions similar to thePTMS are under development by the SupportSystems Contractor (SSC). These wouldinclude the Integrated Modeling Environment inconjunction with the real-time data provided bythe TMC Data Base Management System thatwill allow prediction of traffic conditions. Fornear real-time predictions, Wide-Area TrafficControl (WATC) will generate demand forecastsa few tens of minutes into the future.

Until such systems are implemented, thisfunction will be performed by the operator whomanually alters signal timing plans under UTCSbased on accepted procedures and pastexperience. Experienced and highly motivatedoperators learn to maintain an acute awarenessof the traffic situation and to predict areas ofcongestion well ahead of time. The PTMS-typesystem, when implemented, will be of the mostuse during periods of high levels of mentalworkload.

DETERMINE MAINTENANCE NEEDSThe ideal TMC will have a means for automatedidentification of the need for preventive andremedial maintenance of the elements of theinfrastructure, including roadway, structures,sensors, computers, and effectors.

Preventative maintenance in TMC’s is oftencontracted with commercial agencies and such

23

maintenance contracts are cost-effective andpractical in high-density areas, given that thenumber of signalized intersections that can bemaintained by one technician is in the range of25 to 35. In the majority of TMC’s, computersand much of the mainframe peripheralequipment are under expensive maintenancecontracts. The general trend in most TMC’s isto re-host systems to relatively inexpensivepersonal computer-based systems that can bemaintained by TMC staff. Such practices maybe eliminating preventative maintenanceapproaches in favor of system replacement atthe time of failure.

Remedial maintenance practices differ acrossTMC’s, but include three main methods: patrols,public complaints, and self-diagnosis. ManyTMC’s employ patrols of government employeeswho may either be sent out periodically orcontinually. For example, Los Angeles assignsemergency maintenance crews to routine signalinspection patrols when they are not respondingto trouble calls. Routine checks are donefrequently, with more involved inspectionsperformed less frequently. Public complaintsabout signal malfunctions can be a valuableresource in fault detection. In order for this tobe effective, however, the public must beeducated about the role they play and aboutwhere to call to report observed faults. Self-diagnosis routines in computerized controlsystems may automatically detect malfunctionslong before any public complaints are received.Computer control allows for confirmation ofpublic complaints and diagnosis of malfunctionsbefore repair crews are dispatched.

Currently, one European TMC has integratedthe process of incident detection/managementwith recording of outages and maintenanceproblems. As part of the normal reportingprocess for an incident, the operator classifiesthe severity of the incident and enumerates anycompromises to the integrity of the intersectiontraffic control hardware. This entry is part of adata base that automatically issues a repairrequest to the appropriate maintenance group.Other TMC’s have less automated approaches,but are establishing some form of computer-based communications with maintenancegroups.

SELECT TRAFFIC MANAGEMENTTACTICSIn the ideal TMC, a response advisory system(RAS) will help determine the appropriateresponse to current or predicted variations inroadway capacity or traffic flow. The RAS willprovide a simulation capability that allowsevaluation of various response alternatives,using a version of the PTMS to obtainpredictions. This simulation will use currenttraffic data and must be performed many timesfaster than real time.

In most current centers, traffic managementtactics are chosen based on operatorexperience. Some centers have developedcomprehensive procedures manuals thatdescribe appropriate tactics for varioussituations.

Two existing systems, COMPASS in Canadaand the Motorway Control and Signaling System(MCSS) in Europe, have largely automated thedecision process for selecting appropriate VMSmessage patterns and lane-control strategies,respectively. COMPASS requires the operator toinput data in case of an incident, while MCSScan operate without operator intervention.

In the Integrated Modeling Environment SupportSystem, under development by the SSC,various signalization and control models will beused to optimize control parameters on the sub-network or network level. The Wide-Area TrafficControl (WATC) support system will actuallyimplement these tactics with operator capabilityto influence choice of tactics.

SELECT INCIDENT MANAGEMENTTACTICSIn the ideal TMC, an RAS will help determinethe appropriate response to an incident.Responses to an incident can include doingnothing, informing motorists of a slight delay,rerouting some traffic, and, in extreme cases,closing a major freeway and rerouting all traffic.Incident management will also includecontacting and coordinating with emergencyresponse agencies in order to clear the incidentas quickly as possible.

In the case of the rerouting, the RAS willdetermine the optimal alternate route(s). TheRAS will learn from experience. The RAS will

24

also monitor the clearance of an incident andhelp determine when to return to normaloperation. For example, the RAS will determinewhen to clear an advisory message from a VMS.The RAS will also provide a simulation capabilitythat allows evaluation of various responsealternatives, using a version of the PTMS toobtain predictions. This simulation will usecurrent traffic data and must be performed manytimes faster than real time. Most current centershave a detailed set of procedures that must befollowed in case of an incident. Theseprocedures typically leave little latitude forindependent operator decisions.

The Motorway Control and Signaling System(MCSS) architecture, now being implemented inEurope, provides the closest approach toautomation of incident responses on freeways.Outstations consisting of loop detectors,computer, and gantry sign are locatedapproximately every 500 m. Each outstation isconnected to the upstream and downstreamstation as well as to a central computer. Whena variation in traffic flow meets some threshold,the outstation determines a response strategythat may include stopping traffic, closing a lane,diverting traffic into other lanes, and changingthe speed limit on a lane-by-lane basis. Theappropriate strategy may include sign changeson upstream gantries. The formulated plan iscommunicated to the central computer where itmay be modified according to information knownto the central computer, but not the outstation(e.g., a manually entered operator response).The outstation then carries out the responseplan at its location and orders appropriateupstream and downstream signing changes. Allof this can be done in approximately 0.4 swithout an operator in the loop. Typically, MCSSoperators have little coordination withemergency response agencies.

Another highly automated system, COMPASS,uses an extensive network of loop detectors andVMS to influence vehicle movement. Prioritieshave been established among several classes ofsigns, including incident management,congestion management, navigation, and publicservice.

SELECT DEMAND REGULATIONTACTICSIn the ideal TMC, the RAS will also selectstrategies for demand regulation. This isenvisioned to include strategies for overalldemand reduction as well as specific demandregulation tactics (e.g., for special events). Thisis a decision-making function in which computermodels may suggest a menu of demandregulation options and test them against currentor projected traffic levels. Given this decisionaiding, the operator will then choose theappropriate strategy. In current TMC’s, theoption generation and selection is strictly afunction of the manager’s or operator’sknowledge, experience, training, and judgment.

SELECT PUBLIC RELATIONSTACTICSIn Chicago, the commitment to highlyresponsive inclement weather road surfaceclearing and treatment has changed the public’sexpectations on responsiveness. As aconsequence, the TMC has utilized weatherforecasting and anticipated road conditions topre-position assets, such as road saltingequipment sitting on the highway in anticipationof the impending storm. Public serviceannouncements and videotapes produced byFHWA and the Canadian Ministry ofTransportation are providing detailed informationto the public on the operation of the TMC’s intheir area by providing a “users guide” toeffective use of VMS signage and other systemresources and appropriate motorist responses tocommon travel situations that can help reducecongestion and potential for incidents.

INFLUENCE ROUTE SELECTIONThe ideal TMC will provide outputs that influencedrivers’ route selection decisions, both pre- tripand enroute. Effective means of communicationwith drivers must be developed and used.

Variable message signs (VMS), also known aschangeable message signs (CMS), will be usedto distribute information relevant to the localtraffic. They may present information onexpected travel times and delays, incidentsahead, and, in some situations, suggestalternate routes. A VMS at a major railroad

25

crossing will provide an estimated time toclearance when traffic is stopped for a train.There are currently numerous theories on thebest use of the VMS assets.

The highway advisory radio (HAR) system,consisting of a network of low-poweredtransmitters, each with a limited coverageregion, will broadcast messages thatsupplement messages displayed on a VMS.This system will take advantage of theconventional radios in most vehicles. It allowslocalized advisories to be received by traffic inthe affected area. These recorded messagesmay also be heard over the telephone by dialingin; each transmitter is accessible separately.

Smart vehicles will comprise a subset ofvehicles on the roadway system. These willinclude participating commercial fleet vehicles,government-owned vehicles, publictransportation vehicles, and some privatevehicles owned by individuals who participate ona voluntary basis. Vehicles with navigationsystems that use real-time traffic data will beable to report their origin-destination (OD) datato the TMC. Subsequently, reports of currentposition and planned route will be providedautomatically by the vehicles. In return, they willobtain real-time updates on traffic conditions.

Future automated vehicle control systems(AVCS) will. be partially controlled by the TMC,which will send signals that are received by in-vehicle systems. Many AVCS subsystems willrequire data from the Ideal Traffic ManagementSystem (ITMS) to achieve full operability. SomeAVCS subsystems will be deployed in the midterm and will require interfaces with the TMC.

In addition to its organic outlets, the TMC willsend data to other organizations that will, in turn,transmit information to the public. Commercialradio and TV traffic advisories will continue todisseminate traffic information in the near to midterm. The TMC will provide data for use inpreparing these advisories. Commercial radioand TV will also be used in major advertisingcampaigns to advise the public of majorconstruction or maintenance projects and oftraffic conditions associated with special events.The Traffic Channel cable TV outlet will relayTMC messages and data about current trafficconditions. It will also carry information frompublic transportation authorities. It will be shownat information kiosks at airports, truck stops,

hotels, and other places where travelers aremaking trip and route decisions, as well as inhomes and offices. A few installations of traffickiosks have been made, especially in areassupporting high levels of tourism. A trafficbulletin board service will be accessible bycomputer over telephone dial-up. This servicewill contain data on current traffic conditionsthroughout the metropolitan area. Trafficvolumes and rates will be reported by route, andincident locations posted upon verification.Third-party vendors will create software that willtake this data and create tailored trafficadvisories and route planning assistance forusers. Major users will include commercialvehicle operators, but many private individualswill also use the service. The service will beaccessible from homes and dispatch centers (foruse in trip planning), and also from smartvehicles equipped with intelligent systems thatuse the data for optimized navigation. Printmedia, especially local newspapers, will be usedto convey information about construction andmaintenance projects that will impact traffic,traffic management for special events in thearea, and general information on the operationsof the ITMS. The TMC will support preparationof print media features concerning such topicsas defensive driving and use of publictransportation systems.

An Information Dissemination System (IDS) willinterface with the communication links toresponse forces such as police, ambulance, firedepartment, and towing services, and to thedata services supplied to the mass media, theTraffic Channel, and the bulletin board service.The IDS will be capable of posting messages onVMS’s and creating voice messages forbroadcast on highway advisory radio.

INFLUENCE VEHICLE MOTIONIn the visionary TMC, a real-time adaptive trafficcontrol system will take data from the roadwaysensors and use it to continuously adjust trafficsignal timing and ramp metering plans tooptimize flow throughout an entire network. Thesystem will maintain optimal flow byautomatically responding to current andpredicted stochastic variations in traffic. A keyfeature, not present in existing systems, is thatcontrol will be integrated over surface street andfreeway traffic. The algorithms will be robust,

26

and will automatically accommodate unusualtraffic patterns, such as those associated withspecial events.

This vision is highly realistic. Traffic signals willremain a primary means for traffic control in thenear to mid term. Using control systems nowavailable or under development, the TMC willhave sophisticated control (both manual andautomatic) over the timing patterns for trafficsignals. Timing patterns may be adjusted tooptimize the local flow of traffic through theintersection controlled by that signal, and signalsin a given area can be coordinated to optimizeflow in accordance with current traffic conditions.While there is considerable discussion about theneed to integrate freeway and arterial signalsunder a single control system, this appears to beseveral years from reality.

The Motorway Control and Signaling System(MCSS) architecture in Europe provides theclosest approach to full automation of a systemfor influencing vehicle movement on freeways.

In Canada, COMPASS uses an extensive VMSnetwork to influence vehicle movement. Prior tosystem implementation, a team of expertsdivided the roadway into several hundred uniquezones. They then determined the appropriateVMS response for different types of incidents ineach of the unique zones. A large set of VMSmessages are preprogrammed. When anincident occurs, the operator enters informationabout the incident location and type, and theappropriate VMS response plan is automaticallycarried out.

Roadway traffic control for automobiles may beshared with that for other roadway-basedtransportation modes. In many European cities,for example, buses and light rail vehicles aregiven priority (e.g., green signals) when an on-board computer and transponder signals thatthey are behind schedule.

CONTROL ROADWAYACCESSThe ideal TMC will have the capability to controlthe access of vehicles to certain segments ofthe roadway system, or to the roadway systemas a whole. The control of roadway accessmust be capable of denying access to some orall vehicles (e.g., closing a segment to allvehicles except emergency vehicles) and ofcontrolling the rate at which vehicles gain

access to a segment. Vehicle occupancy can beused as a basis for controlling access (e.g.,HOV lanes) in accordance with the strategicregulation of demand.

Ramp metering, in some form, will continue tobe an important tool for controlling ingress ontolimited-access roadways in the near to mid term.A major problem with ramp metering systems,though, is a high degree of noncompliance. Infuture systems, ramp metering will becoordinated with nearby traffic signals on thearterial network to distribute the overall systemload.

Some TMC’s now have a capability to remotelyclose a gate on a ramp leading to a freeway. Inother jurisdictions, gates are present on theramps but they must be physically closed byfield personnel or police.

DISPATCH EMERGENCYANDMAINTENANCE RESOURCESThe ideal TMC will have the capability ofdispatching emergency and maintenanceresources that result in the timely provision ofappropriate services. These dispatches mustalert available service providers of the type ofservice required and location of the need on theroadway system.

Some, but not all, TMC’s have the responsibilityfor dispatching emergency services. Asmentioned above, TMC’s routinely dispatch byradiotelephone, telephone, and occasionally byCAD. In TMC’s without such dispatch services,the centers are in phone communication with thedispatcher or, rarely, by computer-aideddispatch (CAD) and may have a police liaison onduty during rush hours (at a minimum) to aid incoordinating with dispatchers. Such policeliaisons are also able to assist in coordinatingTMC activities directed at an incident scene andmake command decisions to close lanes andrequest additional emergency services such asfire and ambulance. In many centers, tensionsexist in the relations between police and theTMC staff, resulting in TMC’s providing incident-related information to police, but not receivingany confirmation of actions taken by the police.Because of such tensions, many centers havedeveloped their own emergency patrol services,which are dispatched by the TMC and travel onassigned routes and timetables. Dispatch of

27

emergency services will be greatly enhanced bythe emerging vehicle location technologies beingdeveloped.

INFLUENCE TRIP DECISIONS ANDINFLUENCE MODE SELECTIONThe ideal TMC must be capable of providingoutputs that influence trip decisions, includingwhether a trip should be made and when a tripshould be made. The influence of these outputsmust be strong enough to enable the ideal TMCto contribute to the strategic regulation ofdemand for the roadway system.

In many TMC’s, traffic channels carried by localcable companies inform the public of congestionand travel times. In some instances, it may bevaluable to persuade drivers to delay a plannedtrip for a few minutes or a few hours due tospecial events, congestion, or major incidents.Several TMC’s are also planning disseminationof traffic information kiosks, such as used inAnaheim. HAR and traffic reports are alsocarried on local radio channels, which mayinfluence trip planning and trip in-routeselections. Apparently, such stations arepopular with the public as evidenced inMinneapolis where the traffic radio personalityhas developed his popularity to the level of

having name recognition in the metropolitanarea.

PROVIDE PUBLIC INFORMATIONThe ideal TMC must be perceived as reliableand competent. The ideal TMC must be carefulto provide accurate information to the public.For example, motorists are less likely to heedadvisories recommending an alternate route ifthey do not trust the TMC. Through itscompetency in facilitating the safe movement ofpersons and goods with minimum delaythroughout the TMC’s area of influence, theideal TMC must also continually provide generalinformation to the public regarding itsoperations. Public relations campaigns areconducted, often in conjunction with the publictransportation system, in order to reinforcepublic perception of the ideal TMC.

All TMC’s recognize the importance of publicrelations and provide frequent tours of theirfacilities and routinely issue public serviceannouncements and videos of centercapabilities to provide public information. Manycenters have specific center design objectives toassist in the public tour of facilities and includesuch assets as big-screen displays, specificallyfor public relations use, to illustrate the TMCfunctional systems.

28

Section 5.USER-CENTERED AUTOMATION

Designing a complex human-machine systemlike the TMC from a user-centered standpointmust go far beyond assessing individual displayand control components and their arrangementin the workstation. Human factors inputs to thedesign process should have a significant impacton the high-level design philosophy. Inparticular, the rationale underlying the allocationof functions among operators and machinesmust be driven largely by human factorsconsiderations.

Early approaches to function allocation from thehuman factors perspective divided functionsbetween operator and machine according towhich could best perform the function. TheMABA-MABA (Men are better at:/Machines arebetter at:) approach served well for simplesystems that did not employ extensive feedbackloops between operator and system.

The results of this approach are far too simplisticfor today’s complex systems because there isno way to share the allocation of functions thatare performed jointly by human and machinewithout extending the analyses to a greaternumber of levels than is otherwise useful. Newapproaches to function allocation define theoperator’s type of interaction with the machinesystem. A long-term goal of some trafficmanagement systems is total automation; thegoal of others is to remain operator-intensive.Each philosophy has significant designimplications for the TMC.

As a crucial part of TMC configuration tradestudies, the role of the human operator in thesystem must be defined. Each of the selectedfunctions may be allocated to the humanoperator, an automated system, or somecombination of the two. The function allocationbecomes a basis for a detailed specification ofwhat the TMC hardware and software systemsare expected to do and what the human

operator is expected to do in interacting with thesystem.

The performance of a system function may bedescribed generically in terms of a control loopthat involves sensors, effectors, and a decision-making (loop-closing) agent that initiates controlactions based on sensed information about thetask environment. An operator role in a functioncan be defined by how loop closing (decisionmaking) is performed by humans versusmachines. Using the basic distinction betweenmanual and supervisory control (describedbelow), it is possible to define four genericoperator roles in a function, - Direct Performer,Manual Controller, Supervisory Controller,and Executive Controller.” Within each role,there are useful distinctions as described below.The basic distinction between manual controland supervisory control is as follows. In manualcontrol, machine components may be heavilyinvolved as sensors and effectors, but the loop-closing aspect of the function is solely theresponsibility of the human. In supervisorycontrol, a machine component is allowed toclose the loop under supervision of a humanoperator who may intervene and adjust oroverride the machine’s decisions.

Using this distinction, it is possible to define acontinuum of operator roles in relation toautomation of the performance of a function. Atone end of the continuum, a function is allocatedsolely to the human and, at the other end, solelyto the machine. In between, performance of thefunction is shared by human and machinecomponents. It is useful to divide the continuuminto four major regions; each region defines ageneric operator role in relation to automation.Because it is a continuum, within each regionthere are degrees of more or less automation.The continuum is illustrated in figure 2.

29

Figure 2. Continuum of operator roles in relation to automation.

VMS AUTOMATION PHILOSOPHIESThe area in TMC operation in which the mostautomation is found is the management ofnetworks of variable message signs. There areseveral VMS automation philosophies in use inthe TMC’s visited, representing the range ofhybrid automation approaches.

Manual input. At a few centers, VMS signs aremanually keyed into the system. In most cases,the computer screen presents a template thatthe message must fit and severe limits onvocabulary provide consistency in messagecontent. One center has installed a workingmodel of the actual VMS so that the manuallyproduced signs can be highly refined usingvariable letter spacing before they are actuallydisplayed. In one instance, the operator wasunable to produce a suitable sign for thesituation (police had requested a “strongerrecommendation” that traffic exit at a certainpoint) due to the vocabulary restrictions.

Manual input with computer assent. Insystems that have complex response plansinvolving VMS or lane-control signs, there aremany rules concerning the acceptable content ofadjacent signs. In lane-control systems, forexample, speeds may not be stepped down bymore than 20 km/h on adjacent signs.Frequently, operators manually request aninappropriate pattern. Decision aids are nowreaching the marketplace that reject these

inappropriate patterns, explain why they areillegal, and suggest acceptable alternatives.Operators may accept the suggested patterns,try a new pattern, or, at their own risk, insist oninstalling the illegal pattern.

Choose from selection of canned signs.Some other centers provide the operators with acomprehensive menu of sign messages thatmay be displayed. The operator picks the mostappropriate sign from the menu and inputs therequest that it be displayed.

Computer determines response plan,operator assents. In the more highlyautomated centers, the system becomes awareof congestion or incidents through sensor dataor operator input. A response plan, includingappropriate messages on a series of signs, ispresented to the operator for approval or editing.Only after approval by the operator is the set ofmessages displayed.

Computer determines and carries outresponse plan. In the most highly automatedcenters, the computer network becomes awareof slow or standstill traffic from sensor data.Using internal response logic, an appropriateresponse plan is determined by outstationcomputers, approved by a central computer, andimplemented without interaction with theoperators. The new status is then appended tothe operator’s display of computer decisions andthe operator may manually override the inserted

30

response plan if necessary.occur very rarely.

Such overrides

AUTOMATED RESPONSE PLANNINGThe COMPASS center on the 401 Freeway inCanada employs a comprehensively pre-planned set of responses (VMS messagepatterns) to potential incidents. The freewayunder control was divided into several hundredunique response zones based on distanceupstream or downstream from interchanges,distance upstream or downstream fromcrossovers between local and express lanes,roadway geometry, location of VMS’s, and othersimilar factors. Appropriate response planswere then developed for multiple incidentscenarios within each of the unique responsezones.

When no incidents are present, the VMS systemautomatically displays congestion managementinformation in response to loop detectormeasures of traffic volume. If no congestion ispresent, other information, including navigation(e.g., name and distance to next exit) or publicsafety (e.g., drunk driving admonitions), isdisplayed.

When the COMPASS operators detect anincident, they initiate an incident screen on theirCRT and, according to prompts, key in datadescribing the location and type of incident. Theautomated incident response system searchesits file of incident templates until it finds onematching the operator’s report. Thepredetermined sign patterns are displayed to theoperator who consents (usually) or makeseditorial changes (rarely). The automatedresponse, however, is limited to sign displays.The operator must notify and coordinate withincident-response resources by voice.

AUTOMATED LANE CONTROLMessages reflecting reduced speed limits,closed lanes, and required lane changes aredisplayed for each lane of the freeway onoverhead gantry-mounted signs. Lane-basedcontrol requires dense sensor arrays andnumerous gantry-mounted signs over each laneof the freeway. Approximately every 1/2 km,there is a lane control system consisting of anarray of double-loop detectors, a control box,and a gantry sign. The loop detector array

detects changes in traffic speed that signal thebeginning of a congested situation. Outstationcontrollers located near the roadway sensecollectives of the loop detector arrays andautomatically formulate the appropriate patternof messages and request permission from acentral computer to present the selectedmessages. The central computer will eitherconsent to the changes or will modify theoutstation’s recommendation to reflect thesituation at other outstations or messagesmanually entered by the operator. Various rulesgovern the allowed messages. Adjacent lanes,for example, cannot have speed limits that varyby more than 20 km/h. Speed limits must bestepped down by 20 km/h increments; forexample, a 50 km/h limit must be preceded by a70 km/h sign, which must be preceded by a 90km/h sign. Control of the lane signs is highlyautomated. On the initial installation 10 yearsago, the lone operator was required to consentto each change in sign pattern. Sign changesoccurred at a rate of up to 200 per hour,generating an unnecessarily high workload. Allsigning decisions based on loop detector dataare now made by the computer system andcannot easily be overridden by the operator. (Infact, the relatively less-experienced Amsterdamoperator believed that it was not possible tooverride the computer and she wouldn’t do iteven if she could.) The operator frequentlyorders changes in the signs, for example, toclose lanes for work zones. Given an operator-closed lane, the system makes a check on thelegality of the operator’s inputs and thenchanges neighboring signs to take the closureinto account. If the operator requests an “illegal”sign (e.g., diverting traffic into a median) thesystem will refuse and will provide anexplanation. The operator can then override thesystem, if desired, and continue the “illegal”procedure. In the event that controlcommunication is lost between outstation andcentral control, the outstation automaticallypresents the messages on its gantries. Thisform of “graceful degradation” is a designedfeature of the control system. Also, becausemuch of the analysis is done locally, only amodest amount of data needs to becommunicated between outstation and thecentral computer.

31

FLEXIBLE FUNCTION ALLOCATIONIn modern, computer-based systems, allocationsof function are not necessarily static, but may bealtered in response to operator learning andproficiency, system efficacy, and momentaryworkload.

Dynamic function allocation may be categorizedaccording to which actor (the operator or themachine) initiates the reallocation. In manycurrent systems, the operator may switchautomated systems on or off at will (e.g.,components of an aircraft autopilot). This maybe seen as one aspect of supervisory control.The inverse situation, in which the functionreallocation is done by the automated system(known as adaptive automation), has beenexplored philosophically and in controlledlaboratory studies, but it remains not much morethan an interesting user-system interfaceconcept.(5,6)

Dynamic reallocation of functions undersupervisory control is relatively common in manyautomated systems in TMC’s. Operators maycancel an automated timing plan and manuallyinsert a substitute. Operators may intervene ifthey notice that an automated response systemis implementing an inappropriate response plan.Operators may grant the automated systemgreater authority if their workload in approvingresponses becomes too high and they expectthe automated system to make no critical errors.

Dynamic reallocation of functions under adaptiveautomation has been explored most vigorouslyin the context of the fighter aircraft cockpit. Inthis environment, peaks in workload canmomentarily exceed the pilot’s capabilities; thepilot’s capabilities may change due tophysiological factors. It would be helpful to havea computer that can sense when the workload isabout to exceed the capabilities and reallocatethose functions to reduce the workload peak.

Morrison and Gluckman defined adaptiveautomation as “automation which is capable ofengaging and disengaging itself in response toeither (1) the occurrence of a critical event orevents, or (2) based on the performance of thehuman component in a person-machinesystem.”(5) The change may also be based on adynamic assessment of the human’sphysiological state.(6) Adaptive automation canpromote improved situation awareness,

increased task involvement, regulated workload,enhanced vigilance, and the maintenance ofbasic job skills. It can increase missioneffectiveness by better management ofresources in a complex task environment.

There are three basic strategies for adaptiveautomation, including (1) allocation, in whichentire tasks are reallocated between person andmachine; (2) partitioning, in which subtasks arereallocated from operator to machine while theoperator remains responsible for performing theoverall task; and (3) transformation, in which thetask structure is altered. Each may be morepractical than the others under certainconditions. In an allocation strategy, forexample, a support system might assumeresponsibility for selecting and inserting VMSmessages instead of recommending patternsand awaiting approval. In a partitioning strategy,a support system might allow operators tomanually control traffic signals in a problem areaof the city, while placing those in the rest of thecity under automated adaptive control. In atransformational strategy, an information fusionsupport system might automatically adjust thegranularity (level of detail) of informationpresented to correspond to its current criticality(e.g., automatically activate appropriate CCTVcameras and monitors in response to asuspected incident).

Adaptive aiding by a support system could beused to improve the quality of decision making.For an adaptive aiding system, specificationsneed to be developed in the following areas?

(1)

(2)

(3)

(4)

32

Motivation for specifying adaptiveaiding. This could include predictedoperator performance degradation,projected workload, criticality of thedecision, or ease of implementation.

Tasks to be aided. Either foreground(high criticality) tasks, background tasks(low criticality), or a combination ofthese could be aided.

Type of adaptive aiding. Can beallocation, partitioning, ortransformation.

Method of aid invocation. Could beunacceptable system performance,operator workload, the nature of thetask, or other factors.

(5) User-system communicationrequirements. Includes the amount ofinformation the operator needs aboutwhat the aid is doing, the aid’s outputproduct, and the aids functional process.

Adaptive automation currently exists in conceptmore than in reality. There have been somerelatively limited laboratory demonstrations, butthe technology to implement it has yet to berefined for use outside the laboratory.(8,9) Thesubstantial computing power and softwarerequirements required to implement an adaptiveautomation system may be prohibitively costlyfor the traffic management community.

Numerous crucial human factors questionsremain before a useful implementation of theconcept can be made. These include:

(1)

(2)

(3)

(4)

(5)

How to eliminate automation-inducedcomplacency.

What type of automation strategy is bestfor the TMC environment.

Measures of operator workload andperformance to trigger aiding.

Effects of automation on other,operator-performed tasks.

Interface designs for aiding systems.

AUTOMATION TO SUPPORT THEHUMANAn important aspect of user-centered design isthat system automation should be designed toassist and support the operator. The operatormust always remain “in the loop” and fully awareof: (1) what information the computer has, (2)what the automated system is doing, (3) why itis doing it, and (4) what it is going to do next.The operator must have, at a minimum, theability to switch off the automated system toprevent a future problem.(8)

Clumsy automation can increase the operator’sworkload and induce frustration. At one centerusing Urban Traffic Control System (UTCS)software, all signal timing plans are operated ontime-of-day schedules with automatic changes

of all intersections at 11:00 a.m. and 6:00 p.m.Operators can manually override the systemand customize the timing plans for individualintersections. Beginning at about 10:30 a.m.,the operator was observed to make extensivetiming plan alterations in response to congestionfrom a major special event. At 11:00 a.m.,without warning, the automated system erasedthe operator’s manually programmed timingplans and inserted the pre-programmed mid-daytiming plan. Later that same day, the operator’smanually programmed timing plans (preparedfor evening special events) were erased whenthe normal evening timing plans came on at 6:00p.m. The operator received no indication thatthe time-of-day changes were about to be madeand, in any event, had no way to prevent thechanges.

Many routine functions can be automated in theTMC. Successful implementation of theautomation appears to work well when agradual, iterative process is used. Initially, anoperator should be present in the control roomto review and approve all automated systemactions. A detailed log of automation failuresshould be kept. Early in the life of the MCSS,for example, automated system errors were verycommon. Frequently, maintenance equipmentmoving in a closed freeway lane was interpretedby the computer as wrong-way drivers and alltraffic was brought to a stop. The TMC operatorhad to take control and quickly restore order; anew version of the automation softwarecontained a fix that, in effect, ignores traffic inclosed lanes.

As such errors are eliminated from the software,it is possible to become increasingly lessdependent on the operator. Newer operatorsconsider the MCSS to be faultless and onefirmly stated that she was not capable ofoverriding system control functions (not true)and, in any event, would never do so for liabilityreasons. While total automation is the goal ofseveral centers and some are operating at nightwithout an operator, we have seen none that areconsidered ready to run automatically duringdaylight hours.

33

Section 6.SYSTEM EVOLUTION

During the next decade, dozens of traffic controlcenters and traffic management centers thatwere established in the 1970’s and early 1980’swill be upgrading their computer suites andsoftware systems, increasing levels ofautomation, and modernizing other equipmentand procedures. In some centers, this will be arevolutionary process in which most of the oldertechnology will be discarded. In many, however,the IVHS era will arrive one step at a time withthe introduction of, perhaps, a VMS system in 1year and a new support system a couple ofyears later. The third form of system evolutionis development of an IVHS-class trafficmanagement system for a small area initially,and continuously extending the boundaries ofthe controlled area and the functions of theoperators.

In many ways, modernizing an existing centercan be more challenging, from a human factorsdesign viewpoint, than building a new TMC fromscratch. The existing center will have a staff ofoperators who are well accustomed to theexisting equipment and procedures. There maybe great resistance to change, especially if thenew technologies seem exotic and threatening.It is crucial that the existing operators be fullyenrolled in the modernization by involving themheavily in a user-centered design process.

The user-centered design process can be usedeven more effectively during systematicupgrades to a TMC than during initial design.During planning for an upgrade, there is anavailable cadre of operators who haveexperience with the roadway network, TMCoperating philosophy, and the capabilities andconstraints of the existing system. Only one ofthe upgraded centers that were visited had useda formal user-centered design process involvingthe existing operators. Function and tasksanalyses were completed based on operatorinterviews and examination of documentation.Workstation mockups were constructed basedon the defined job requirements and wereiteratively refined using operatorrecommendations.

Other centers used less formal methods ofintegrating user recommendations. One centerasked operators to estimate the number ofCCTV monitors that they could effectively useand based the design on the answers. A small-scale mockup of a proposed center layout wasconstructed and operators were encouraged toexperiment with different console and equipmentlayouts using the model.

Another center that is highly dependent on VMSmessages for incident and congestionmanagement became justifiably concerned thatthe semantic content of messages beunderstood by the public. Focus groups weredeveloped from among local drivers to discusstheir perception of the meaning of candidateVMS messages. The final message philosophyand suite of messages was based on the resultsof these focus groups.

PLAN FOR EVOLUTIONAs new centers are designed, plans forexpansion and evolution should be incorporatedinto those designs. Several centers that werevisited need to expand their areas andresponsibilities, but have insufficient floor spaceto support the new operators and consoles suchan expansion would require.

At one new traffic signal center, on the otherhand, management has foreseen a majoradvantage in the possibility of co-locating withthe existing Freeway Traffic ManagementCenter in order to improve coordination ofoperations. Substantial coordination of incidentmanagement procedures are required becauseresources assigned to the signal center areresponsible for management of incidents on thefreeway as well. The new center was designedwith sufficient floor space to incorporate theFreeway Traffic Management System (FTMS) ifsuch a move should occur.

35

INTEGRATION OF THE OLD WITHTHE NEWOne of the key issues in system evolution is thatof system compatibility; the new systems mustbe easily integrated with the old. A new set ofVMS’s must be able to be controlled by thesame interface and procedures as the existingset. A new set of sensors must be able todisplay data on the same screen and in thesame format as the existing ones.

One signal control center was operated usingdetectors, controllers, and software developedby one vendor. When their system wasexpanded to several hundred moreintersections, a new vendor was selected. Onlyafter installation was complete was it found thatthere is no user interface compatibility and thecenter now has two separate signa-operatorconsoles with very different user interfaces.

It should be noted that introduction of high-endPC computers may provide a solution to someof these compatibility problems. It is possibleunder new PC operating systems to develop a

common user interface replacing the individualinterfaces of hardware that runs on differentcomputers and operating systems.

Another important factor in the upgraded systemdesign is the informal procedures and practicesthat have been developed by operators tooptimize performance. Where possible, thesystem upgrade should recognize and supportthese practices and procedures.

MORE IS BETTER?There is frequently a tendency in TMC design tothink that more of a good thing is necessarily abetter thing. A center in which 20 CCTVmonitors have been effectively used hasdeveloped a growth path toward 50 monitors.Discussion with the designer indicated theincreased workload due to the extra monitorswould require considerable analysis. Perhaps,more attention should now be paid to summarygraphic displays of situation information or otherform of data fusion introduced. More of thesame isn’t necessarily the proper approach.

36

Section 7.USER SYSTEM INTERFACES AND WORKSTATIONS

(Designing USI’s and workstations to improveindividual and team situation awareness)

Maintaining awareness of the traffic situation ona complex roadway network provides asignificant challenge to the TMC operator. Withthe large amount of information andcommunication technology that is available tothe IVHS-class TMC, it is possible to provide theTMC with all of the necessary data to paint adetailed picture of the dynamic traffic situation.The perceptual and cognitive limitations of thehuman operators, however, allow them to attendto a relatively small portion of this picture at agiven time.

The data available to the TMC can beaggregated at various levels of granularity tomeet the immediate needs of the operator. Forexample, the binary status of an individual loopdetector may be of interest during maintenancetroubleshooting; the volume and speed of trafficpassing that individual loop detector may be ofinterest during incident location activities; thevolume and speed of vehicles passing a seriesof detectors may be of interest during mosttraffic monitoring; and a signal from a supportsystem that no indication of anomalous trafficflows are seen in the network may be sufficientwhen the operator is busy with other duties.

For each different function the operator mustperform, the ideal is to provide all the necessaryinformation, but no more. New graphical userinterfaces to information processing systemsallow a significant amount of tailoring of datagranularity, display content, and displaytechniques. Use of these tools can promoteimproved operator situation awareness, but only

if the displays are developed to meet validateduser information needs.

For example, in one TMC, multiplelevels of zoom were defined for the surroundinggeographic region in which the TMC waslocated. Several geographic sources wereconsulted and the data were merged, then aseries of zooms and granularities were definedfrom large geographic region down to anintersection level. Five levels, typical of theavailable zooms are illustrated in figures 3through 7.

Operators with experience in using these zoomsfound that only two zooms, intermediate regionaldisplay (figure 5) and an intersection display(figure 7) were actually useful. In addition, thedistances between the objects displayed werenot made accurate with respect to truegeographic distances and did not later appear tobe problematic. What appears to be importantare the spatial relationships, not the actualdistances between displayed objects.

There may be significant individual differencesbetween operators in the amount of detail theywant to see on the screens. Inexperiencedoperators, for example, may want to have streetnames displayed, while experienced operatorshave these memorized. One form of dynamicscreen redesign, a “declutter” capability, canhelp solve this problem. Using declutter, theoperator can change the amount of detail on thescreen by toggling the display through, say,three pre-defined levels of detail.

37

Figure 7. Example of an intersection zoom-in with high granularity.

RAPID PROTOTYPING OF DISPLAYSDURING TMC DESIGNAs the large amounts of incoming data are fusedinto meaningful chunks, it is usually necessaryto provide summaries in graphical format tosupport the operators’ situation awareness. Theuse of maps, graphs, color coding, and blink-pattern coding can allow the operator toassimilate large amounts of data as well asproviding a warning when something needsimmediate attention. Numerous softwarepackages with which replicas of candidatedisplay screens can be programmed off-line anddisplayed for operator critique or performancetesting are now available. The screens can beeasily refined until they are acceptable and then,with many packages, display software isproduced for rehosting to the real TMC. Suchprototyping can eliminate display characteristicsthat might cause human performance difficultiesbefore they become part of the TMC. ln someTMC displays, 10 or more conditions had to bediscerned on the basis of color coding alone.One interface, for example, used a 12-colorscale to indicate system status. Differentelements of the system had different numbers ofdescriptive levels. One subsystem might have

three levels and use the top three colors;another might use all 12 available colors. As aconsequence, the color that meant “worstpossible” in the three level subsystem wasencoded as “pretty good” in the 12-level system.The cognitive workload of the operator wasincreased by this design and the likelihood ofinterpretation errors was increased. Usertesting of a screen prototype most likely wouldhave discovered and corrected the confusionthat this coding created.

THE CARDINAL RULE: HUMANSMAKE ERRORSThe primary rule to keep in mind during user-centered design process is that people makeerrors while doing their jobs. The design of jobsand systems, however, can have a powerfuleffect on the frequency and criticality of errors.It is relatively easy for an operator to make a“typo” while keyboarding system commands,and such errors should be guarded against.

At one center, the operator types <1> to invokean ordinary setup and <1> to invoke a fullsystem shutdown requiring several minutes torecover. An error trap in the software (“Do you

40

really want to shut the system down? Y N”)would eliminate the serious error this entrywould cause.

The design approach should be to minimize boththe frequency and criticality of operator errors.

HOW MANY OPERATORS?A major issue in control console and workstationdesign is the number and role of operators whowill run the center during different shifts andconditions. Numerous centers that were visitedwere-designed to be run using multipleworkstations, each with its own role. Yet, onevenings and weekends, the staff is reduced toa single operator who not only must be cross-trained on each of the job roles, but who mustmove between a series of workstations andretain awareness of a series of displays andcommunication systems.

One center solves this problem by having onemulti-function “supervisory” workstation that canbe used during the day by the supervisor tosupport any of the operators and can be used atnight to combine the functions of all workstationsinto one.

BIG BOARD DISPLAYSA dominant feature of nearly every center visitedwas some type of big board display. Typicallyused to present a broad overview system map,these displays are readily visible to all operatorsin the TMC. Three different kinds of big boarddisplays were seen: static wall maps, activewall maps, and projection television screens.

Most managers readily admitted that the primarypurpose of the big board was to provide adisplay of large proportions to impress centervisitors. At smaller centers with singleoperators, the big boards did seem to besuperfluous to center operation. At largercenters in which operators walk betweenworkstations to coordinate, some big boardinstallations were seen to provide a usefulcommunication tool. Dynamic big boards allowoperators to maintain partial situation awarenesswhen away from their workstations.

Of the three kinds of displays, static wall mapsappear to be the least useful (see figure 8).These can provide the geographic layout of theroadway system, and the location of ATMSresources, such as detector stations, signals,signs, and cameras. Most of these data arememorized by operators and they have no needto have it prominently presented before them.Dynamic wall maps are somewhat more usefulin traffic management because they providesome kinds of current traffic-related information(see figure 9). Tiny holes are drilled into thelarge panels at appropriate locations throughwhich tiny light bulbs or light-emitting diodes(LED’s) connected to sensors in the field areinserted from the back of the map. In trafficsignal-related centers, these maps provideinformation on signal status (illuminating, forexample, when the signal turns green for east-west traffic at that intersection). In freeway-oriented centers, the lights may represent trafficflow (illuminating when congestion is detected atthe location).

41

Figure 8. Example of a static wall map.

42

Figure 9. Example of a dynamic wall map.

One standard complaint about dynamic bigboard displays, since they usually consist oflarge painted maps, is that it is difficult andcostly to change them to match infrastructurechanges. In one ingenious solution to thisproblem (see figure 10), the map is computer-drawn on a matrix of plastic tiles, each about 2.4cm square. When a change in infrastructure ismade, the appropriate tile can be snapped out ofthe matrix and returned to the vendor forupdating.

The projection television provides bothadvantages and disadvantages as a big board

medium. One major advantage is its flexibility. Itcan be (and usually is) used to present a wide-area map with dynamic traffic flow or signalstatus information (see figure 11). It has theability, however, of displaying anything thatmight appear on any other TMC graphicsmonitor, including maps with various views andlevels of zooming, graphics of individualintersection status (if available in the software),live video from CCTV cameras, TV images ofofficials at other sites for teleconferencing, oreven combinations of these.

43

Figure 10. Example of a computer-drawn, plastic tile matrix wall map.

Primary disadvantages of projection televisionare in the image resolution and maintenancerequirements. Because of the pixel-basednature of the TV, it is not possible to presentvery fine print fonts. The image always appearsslightly blurry, even under the best conditions.Finally, sensitive optics are used in theprojection TV and the systems frequently requirerealignment and readjustment.

A fourth kind of big board that was not seenduring any visits, but is being installed in somenot-yet-open TMC’s, is the video wall. A matrixof television monitors becomes the displayrather than a single monitor. This has the effectof multiplying the number of pixels available forthe display by the number of monitors. Thisapproach substantially reduces the resolutionproblems of the projection screen whilemaintaining display flexibility. The disadvantageis the dark lines that appear where the monitorsabut each other.

CCTV SYSTEMSCCTV displays usually are intended to supportloop detector data. Operators (especially in

North American centers) tend to ignore thefused sensor displays and concentrate theirmonitoring attention on the CCTV monitorswhen available (see figure 12). There areprobably several reasons for this preference:(1) CCTV provides a more interesting andcompelling display during periods of lowworkload and boredom, (2) CCTV provides amore natural and intuitive display, and (3) it ispotentially possible to detect traffic-flowproblems slightly earlier with the CCTV than withthe fused detector data.

Overdependence on CCTV, however, can havea major impact on operator performance duringperiods of heavy workload, and some centersare trying to force operators to use the graphicaltraffic-flow displays. In one center, the colormonitors are adjusted to provide only monotoneimages to make them less visually compelling.According to the philosophy of one majorvendor, the number of monitors in the TMCshould be limited to some minimum number andthey should not be turned on until there is anindication, in the detector data, of a problem.

44

In many facilities, locations of cameras aredetermined more by convenience than bystrategy. Existing vantage points, such as roofsof tall buildings, are prime candidates for acamera location. In these facilities, an extraincrement of workload is added to the task ofcamera control and monitoring becauseoperators must mentally determine the locationof the camera and its direction of regard in orderto know what street or roadway they areviewing. In the daytime, this is a minor problem;at night, when fewer landmarks are visible, itbecomes more difficult. Substantial training andexperience is required to learn this skill.

There are at least four solutions to this dilemma.First, on the graphical map display of the area ofinterest, the camera icon should be placed atthe proper location and the icon should rotate, inconcert with the camera, to indicate thecamera’s angles of regard. Second, thegeographic location of the camera should besuperimposed over the monitor image of thetraffic scene. A compass display may also besuperimposed to show the pointing angle of thecamera. Third, a consistent location forcameras should be used. One center, forexample, that manages an east-west freeway,installs all cameras on the south side ofinterchanges so that the operator is alwayslooking northward. (This also keeps sunlight outof the camera.) Finally, new cameras arereaching the market that have preset pointingangles that can be labeled on the controller. Tolook at Central Street to the west, the operatorpushes the appropriate button rather than usingan analog controller. These systems can alsobe programmed to be responsive to an incidentdetection algorithm to automatically turn to thelocation of a possible incident.

Another major issue with CCTV’s is theassignment of responsibility for control of thecameras. In many settings, cameras may becontrolled by more than one operator and evenby more than one center. Problems haveresulted when two or more operators weresimultaneously fighting for control of thecommon camera. It is difficult to establishguidelines for control but some schedule shouldbe set, based on the function and task analyses,to give priority to the operator with greatestneed.

VARIABLE MESSAGE SIGNSVariable Message Signs (VMS) can help tosupport the driver’s situation awareness byproviding information about the conditionsahead. Drivers may be influenced to alter theirroutes in response to major delays. They maybe influenced to decrease speed and becomemore alert in response to more minor congestionand delays.

TMC operators who control a network of signsneed to be aware of the message content ofthose signs to ensure that they are appropriateand up to date. There are frequent stories aboutoperators who forgot that a temporary messagewas left on a sign for hours longer than plannedbecause operators forgot to remove it. Thereare several potential approaches to thisproblem. First, signs containing temporarymessages may be indicated by special colorcoding on the map display of system resourcescommon to most TMC’s. Second, an aging timemight be applied to different messages and,after the appropriate time period, the operatorwould be cued as to the presence of themessage.

VIRTUAL DISPLAYSThe potential to present the operator of a TMCwith a virtual-reality display, including theroadway and cars, and potentially to present a“face-to-face” conversation with other operatorsat remote locations is a compelling vision.Operators could view a roadway scene of anaccident through special goggles that provide acombination of computer graphics and realimages. Operators might locate an emergencyvehicle by circling the area with a gesture anddispatch it by touching the designator on theCRT screen for the emergency patrol closest tothe scene. High-resolution views of an incidentmight be obtained from a CCTV cameramounted on a “robot” vehicle dispatched by theTMC operator and controlled by the operator’sgestures.

Although virtual reality has become a popularbuzzword with “frivolous” connotations, it israpidly becoming an interactive three-dimensional world of great relevance to manyscientific and industrial endeavors.‘“’ There is adesire to make a very natural interface betweenhumans and computers, where the realism

46

created by three-dimensional imagery, spatial long as building a real mockup (and with lesssounds, and even physical forces from motion to resolution).crude touch (possibly even smell), are availableto produce a compelling immersion of the userinto the synthetic environment. The basicbuilding blocks for a virtual reality include adisplay, a tracking device for interactivity, acomputer image generator, a three-dimensionaldata base, and the application software.However, virtual reality still needs much workand assessment before it can become acommon tool for industry. For example, to“build” a virtual mockup of a car may take just as

Experts in the field of virtual reality expect thatproblems in human tolerance of such anenvironment for extended periods (stress andfatigue are among the current problems) andselections of appropriate and specificapplications are needed to clarify the industry.Methods and devices must be standardized anddeveloped from a human factors perspectiveconsidering the specific design requirements forthe systems being designed. Virtual reality maynot be feasible today, but the gap is closingquickly.

47

Section 8.STAFFING, SELECTION, AND TRAINING

(Operator/maintainer initial and proficiency training;staffing and personnel selection)

WHO’S ON DUTY IN THE TMCOne crucial question for designers andoperators of the newly designed TMC involvesselection of operators. The qualifications of theTMC staff must be directly related to the designand function allocation philosophy. If theoperators’ functions are relatively repetitive,predictable, and non-critical, operators with alower level of qualifications may be acceptable.If unique problems are frequent, rapid reactionis needed, and/or criticality is high, operatorswith higher levels of training and experience areindicated.

The minimum qualification for a TMC operatorwould seem to be good verbal skills, a degreeof computer literacy, and good reasoning skillsaccording to the practices of several centershere and abroad. One center, as a result of aformal job study, defined their requirements as“good oral and written communication skills,good interpersonal skills, and a workingknowledge of the local freeway system.” A fewcenters employ part-time students asoperators, usually under the supervision of amanager or senior operator. Centers that areactive in incident management generally requiremore highly qualified operators than signalcontrol centers or those that see their role ascongestion management. At some centers,operators come from a technical staff of trafficengineers or computer scientists who haveother assigned technical duties or who may begiven special projects. One European centerwas totally staffed by police personnel; othershad police liaison officers on duty to handleinteractions with officers in the field. A smallnumber of centers require college engineeringor technical degrees.

INITIAL TRAININGIn the vast majority of centers, initial operatortraining is an on-the-job activity. Trainees are

provided with procedures manuals andequipment manuals.

Three formal training courses for TMCoperators were identified. The COMPASS TMChas a self-contained training room with its owntraining simulator for training COMPASSoperators. California Polytechnic Institute atSan Luis Obispo (Cal Poly) provides asimulator-based training course for Caltransoperators. Chicago Communications Centerhas a 3-month program of text and videotape.

At COMPASS, trainees undergo a 3-weektraining course consisting of 1 week of lectures,1 week of operational training in the simulator,and 1 week of supervised operator training onthe job.

(1)

Classroom topics are:

Introduction to Traffic Engineering,including inventories, capacity andtraffic volume calculation, travel timesand delays, and driver characteristics.

(2)

(3)

(4)

(5)

(6)

Freeway Traffic Operations, includingtraffic characteristics, problemsassociated with freeways, and accidentdata and reporting.

FTMS Theory and Purpose, includingsystem components, software, fieldhardware, overview of FTMS in NorthAmerica, MTO approach in Ontario,and other traffic engineering strategies.

Traffic Management, including incidentmanagement, congestion management,and special situations.

Overview of Highway 401 FTMS,including the system schematic, fieldsubsystems, communications, centralsubsystems, and potential for futuregrowth, including lane control.

FTMS Software, including control roomcomputers, FTMS softwareterminology, software systems, colorgraphics, and operator functions.

49

(7)

(8)

Dealing with Other Agencies, includingvisits to fire, ambulance, and metrotraffic communication centers.

Day-to-Day Operations, includingnormal operations, interagency incidentcoordination response, traffic responseplan guidelines, operational code, theoperator’s workstation, and operationalprocedures.

Cal Poly has designed a three-stage trainingprogram for Caltrans and CHP operators thatcenters around their Traffic Operations Center(TOC) simulator. The TOC simulator consistsof the simulation room containing four operatorworkstations and appropriate computers,monitors, displays, controls, andcommunication devices, as well as a controlroom from which various realistic trafficmanagement scenarios can be run. Thesimulation room can be reconfigured, asneeded, to resemble the home TOC of thetrainees. Figure 13 shows one floor plan of theCal Poly TOC simulator.

Training scenarios were designed to promotemeeting the following major objectives:

(1) Critical contacts will be made on atimely basis.

(2) incident response and trafficmanagement resources will beappropriately allocated.

(3) Inquiring agencies will be provided withthe most timely and accurateinformation.

(4) A professional image will be projected.

Initial scenarios included a mixture of incidentsthat are relatively common, including minorcollisions, an overturned truck with cattle on thehighway, and a zero-visibility fog bank.

The introductory-level program is a 3-daysession consisting of 30 percent classroom

lectures; 40 percent laboratory, workshops, anddemonstrations; and 30 percent simulatortraining exercises. Subjects include:

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

Comprehending CHP radio dialogue.

Using call signs and brevity codes.

Operating two-way radio consoles.

Interpreting traffic information.

Utilizing TOC software.

Operating TOC equipment.

Working with the incident commandsystem and other standard operatingprocedures.

Maintaining appropriate records anddocumentation.

Communicating with the media andother agencies.

The intermediate-level training course is forTOC operators and supervisors who have atleast 6 months of experience and is centeredon major incident and disaster response. Thisclass lasts 3 days, consisting of 30 percentlectures and 70 percent simulation training. Itcovers the following:

Incident command system.

Multi-vehicle accident investigation.

Hazardous spill procedures.

Earthquake emergency response.

Roadside fire response.

Multi-jurisdictional extended pursuits.

Release of information to the media.

The advanced-level training course is acustom-developed course addressing specificIVHS topics that are required by a specificjurisdiction.

50

Figure 13. Floor plan of a Cal Poly TOC simulation room.

The management at the IDOT CommunicationsCenter (ComCenter) at Schaumburg, IL, hasformulated a 3-month training program for theiroperators. The training regime is centeredaround a series of modules that incorporatebooklets with accompanying audio tapes. Thetraining approach utilizes distributed practiceover short training sessions and is limited incontent coverage to about seven items in agiven session. For example, operators mustmemorize the names of more than 100 roads inthe Chicago metropolitan area. In this trainingmodule, the operator learns seven road namesat a time, continuing with successive modulesuntil the entire Chicago road network is learned.Some content areas are limited tomemorization of essential details, such as thedozen of the over 100 radio codes that havebeen determined to be the most commonlyused by the operators. The less commonlyused codes are available for ready reference.These simple learning principles applied totraining operators have proven effective for theComCenter’s needs. In addition to the trainingmodules, in cooperation with FHWA, other

video-based modules are available that providean overview of the ComCenter’s functions andproper management of specific incidents, suchas those involving hazardous materials.Simulator time is also part of the trainingapproach for communication operators. Hands-on training is available at a fully functioningconsole. The console is located in the maincommunications room, but is separate from thefour main stations. The “training” console isused during regular operations to record HARmessages. This console also serves twoadditional purposes. First, it provides a placefor engineers and technicians to monitorsystem performance. Second, during majorincidents, managers and other staff memberscan use the console as an extension of theregular facilities.

SYSTEMS APPROACH TOTRAININGIn order to ensure that the training programproduces operators with the full set ofknowledge, skills, and experience necessary to

51

perform their job, a systematic approach todesigning and developing the training programmust be used. Such approaches, known assystems approach to training (SAT), have longbeen specified as the approach to developingmilitary training. Most approaches todeveloping civilian training are less effectivebecause they are not as rigorous.

A detailed analysis of the operator’s tasks andactivities should be developed as a part of theuser-centered TMC design. This analysis canprovide a foundation for the selection andtraining of the operators. Based on the taskanalysis, the training developer can prepare ahighly detailed list of things the operator mustknow and things they must be able to do.

Personnel with the appropriate combination ofknowledge and skills may be obtained throughselection, training, or, most likely, acombination of these. After defining the likelylevel of entry skills the new trainees will have,the training developer can, by subtraction,determine the necessary content of the trainingprogram.

Then, the training program is built around therequired knowledge and skills. Trainingmodules are specified for each of the requiredknowledge and skill elements. These modulesare then ordered in terms of prerequisiteknowledge. For example, the skill introubleshooting a failing controller (Module B)may depend on first knowing the differencesbetween normal and abnormal controller data(Module A). The instructional medium for eachmodule (e.g., lecture, self-study, videotape, orsimulation training) is selected according toexpected effectiveness and the medium ofadjoining modules. The instructional content ofthe modules is then developed and tested.Training development is an iterative process.The initial training package must be carefully

tested for efficacy, revised, and re-tested untilsatisfactory results are obtained.

Critical, complex, or infrequently performedtasks described by the task analyses should bedocumented in a concise, accessible, well-indexed procedures manual.

RECURRENT TRAININGNothing in the way of formalized recurrenttraining programs was seen at any of the sitesvisited. In general, when a new system isadded that changes the operator’s functions orprocedures, the vendor will supply sufficienttraining to allow some degree of proficiency.Most centers schedule detailed debriefings atthe termination of major incidents so thatperformance deficiencies can be recognizedand rectified as necessary.

As discussed above, Cal Poly is instituting aseries of operator training programs, includingsome on advanced topics for more experiencedoperators and supervisors, including majorincident and disaster management andspecialized topics based on the needs ofindividual TOC’s as they are upgraded to IVHSstatus.

The high degree of automation and computertechnology being introduced into trafficmanagement systems should make this anideal candidate for the use of embeddedtraining, employment of the operationalworkstation as a training device during periodsof low workload. Embedded training has beenwidely accepted in the military to allowrecurrent training in the field. Using themultimedia capability of future graphicalworkstations, it should be possible to presenttraining lectures, practice realistic scenarios,and receive diagnostic feedback onperformance without ever leaving the TMC.

52

Section 9.PROCEDURES, DOCUMENTATION, AND JOB AIDS

Nowhere is user-centered design moreimportant than in the “paperware” that is part ofthe TMC system. There were several distincttrends among the TMC’s visited.

Where design was imperfect, there is extensiveevidence of practical use of job aids and “homeimprovements.” Although it is possible thatsome of the design consulting firms whooriginally developed the facilities had consultedhuman factors guidelines, the resident designand operations staff were most often unawareof human factors and human-system interfaceconsiderations. They typically copy the existingpractice (established by the consultant orsystem vendor), and extend or modify anypractice by applying their own knowledge andjudgment to the remediation of proceduraldesign flaws and shortcomings of obsolescentsystems.

Cost-benefit considerations are typical of theTMC’s visited. All the TMC’s are severelylimited in funds for operational staffs andtraining, several had experienced recentfreezes or cutbacks in their workforce. Some ofthe TMC’s have the capability to automatemany of their job aids. There are concerns,however, that moving away from paper-intensive procedures toward a paperless officewould be a costly endeavor with littleimprovement in efficiency.

Nearly all centers visited had detailed, formallyprepared procedures manuals. Many of theformally prepared manuals emphasizedprinciples and procedures directed at the trafficor computer engineer. The most usable andreadable procedures manuals, though, wereprepared by the operators and emphasized theprocedures. Several operators reported thatthey had rewritten system documentation inconcise, checklist format so that they couldunderstand the procedures when they wereneeded. These manuals varied from the use oftext-based lists to inventive visually formatteddisplays of procedures. For example, at oneTMC, several operators with military trainingdeveloped standard operating procedures forsome of the more confusing and difficultoperations. Figure 14 illustrates the operator-developed procedures.

At this same TMC, the supervisor wasuncomfortable with the operator’s reliance onexclusively procedure-based instructionsbecause strict procedural lists did not addressnew and varied problems. Whenever theprocedural remedy failed for the operator, thesupervisor was called in to solve the problem(sometimes in the middle of the night).

For incident management, response plans fortypical incidents were both formalized in thecomputer system and provided to the operatorsin flow-chart form as illustrated in figure 15.

53

Figure 14. Example of a procedural checklist developed by TMC operators.

Traffic incidents are most commonly contain the approval chain and appropriateunexpected events. Effective incident-related distribution list for the completed forms.job aids are brief in presentation and containonly the essential details. At one TMC,operators and management have jointlydesigned some excellent lists and forms. Twoexamples are presented in figures 16 and 17.

Notice that these forms not only include theessential details for each incident type, but also

Good help systems are rare. None of thesoftware systems we examined had goodHELP systems. Those that did exist were oftenwritten with arcane traffic engineering orcomputer jargon that would be difficult tointerpret by the average operator with littleeducation beyond high school.

54

Figure 15. Example of a flow chart for a detailed response plan.

55

Figure 16. Example of a traffic incident checklist.

Figure 17. Example of an incident report form.

CCTV control requires complex procedural jobaids. Figures 18 through 20 illustrate threegeneral problem areas encountered in thecontrol of CCTV cameras: (1) selection ofcamera through non-intuitive mapping ofcamera location and identification number[figure 18]; (2) camera display on monitors that

have multiple modes (full screen or quadrantdisplay) [figure 19]; and (3) camera display on amatrix of monitors where an individual cameraimage may be shifted periodically (especiallywhen there are more cameras than displaymonitors) [figure 20].

57

Figure 18. Camera-related job aids: location-camera number lookup.

58

Figure 19. Camera-related job aids: camera-control instructions.

Figure 20. Camera-related job aids: camera pattern-CCW monitor layout.

59

LEADING-EDGE TRENDSEffective job aids can support teamwork andcan integrate center activities. Figure 21illustrates a documentation carousel at acommunications center of a TMC. Designedjointly by the staff, the carousel allows thesame forms and reference materials to beeasily shared among all the operators. Thecarousel can be turned from any operator’sworkstation, with compartments on the side forall of the routine forms used in the center.Under the Plexiglas top, the carousel containsthe detail maps and references for commonreferral. The currently active incident forms areplaced on the carousel for other operators toaccess and elaborate as events unfold.

A well-written, on-line help system can alsoeliminate many errors. However, such on-line-based systems were never observed. This is a

byproduct of the observation that no TMCvisited had conducted any systematic analysisof error sources. Each of the centers visitedwould benefit from more attention to thepotential for human error and measures thatmight be used to reduce that potential. Atmost, operators employed user aids (such asterse notes and stick-on post-its) to treat errorsources at the most obvious levels. Managerstended to see error in machine terms, such asa faulty system output, rather than as humanmistakes. Any system can be designed forgood on-line help to address human errorswith the assistance of human factorsguidelines. The growing trend in the use ofrelatively inexpensive personal computers inthe workstations of TMC’s provides readyaccess to adequate development tools for theeasy development of such help systems.

Figure 21. Illustration of a documentation carousel.

60

Workstations should be designed with spacesreserved for operator-designed job aids.Operators frequently have a need to postmiscellaneous pieces of information for theirlater use. This may be an informal reminder toperform a task, a list of commonly usedtelephone numbers, a list of communication 10-codes, or other such data. These notes maybe temporarily posted using pads ofcommercially available sticky-backed paper ormay be more permanently applied by taping onneatly typed notes. Operator-designed stick-on

job aids were extremely common at all TMC’sas illustrated in figures 22 and 23. The use ofsuch homemade labels should be viewed as anindication of weaknesses in TMC workstationand documentation design. TMC managementshould explore means of eliminating the needfor informal labels, including modification ofprocedures, control devices, computer screens,or documentation. Where a real need for theinformation is identified, durable labels in astandardized format should be considered.

Figure 22. Example of labeling on the control panel.

61

Figure 23. Example of labeling below a control display.

62

Section 10.OTHER DESIGN AND PRIVACY ISSUES

This report has demonstrated the importance ofthe human operator in efficient operation of theTMC. Centers and systems must be designedto aid the operator in performance of theallocated tasks. The movement towardautomation only emphasizes the challenge ofhuman factors design.

Of the centers studied, only one center waslargely designed by an ergonomics consultant.The full user-centered design process wassuccessfully employed. A few centers areusing local university consultants and someseek inputs from existing operators. In mostcenters, however, designs are simplyevolutions of the architects’ previous work andare likely to include many of the previousmistakes. None of the centers we visited wasfamiliar with any use of human factors designdocuments in their design.

The process used to design a complex systemlike a TMC can be classified as eitherrequirements-oriented design ortechnology-oriented design. The mostefficient method of designing a TMC is arequirements-oriented and user-centereddesign process. The requirements-orienteddesign process carefully analyzes theobjectives and performance requirements of theTMC. Once all of the requirements have beendocumented, a trade study of availabletechnologies is performed and the best, andmost integrated, approach to meeting thoserequirements is determined. In technology-oriented design, a particular technology suitethat is favored by the designer becomes thecenterpiece of the design process and theTMC’s objectives are built around it.

One mistake frequently made by TMCdesigners, or by agencies soliciting TMC designand development, is to define specificconfiguration items too early in the engineeringanalysis process. Technology-orienteddesign often neglects one or more of thecritical operational or functional requirements,resulting in sub-optimal achievement of theoverall system objectives. At best, technology-oriented design neglects to perform full tradeoff

studies of the available ways of meeting theobjectives, often ignoring new or not widelyknown approaches.

Vendors suggest that solicitations for TMCdevelopment should provide detailed functionalrequirements and allow potential contractors tosuggest how the requirements might be met.This approach would encourage vendors toprovide innovative system architectures anduser interfaces, keyed to each TMC’s uniquerequirements, rather than trying to make oldsolutions fit new requirements.

User-centered design brings in the operatorsvery early in the design process. Whilerequirements analyses for the configurationitems are being conducted, requirementsanalyses for the operator’s performance andtasks are also conducted. User-centered designis an iterative process, beginning during therequirements definition phase, in which usercomments are incorporated into the design ofthe user-system interface. The new interface istested by the users and, based on results andcomments, refinements are made. Theprocess continues until a stable design isreached.

User-centered design was effectively applied ina remodeled European control center. Adesign consultant performed requirements,function, and task analyses using interviewswith the operators, procedures manuals, andjob aids. Based on these analyses, threecandidate mockups of the new control roomwere prepared. Operators selected theirfavorite and then worked with the designer torefine the design. Based on our observations,a very functional center was the result (seefigure 24). The workstation design placed allrequired displays, controls, and communicationdevices on an arc with the operator at thecenter. The infrequently used CCTV monitorswere mounted high on a wall, within view of allfour workstations. Innovative, fully integrateddisplays, using extensive color coding, couldinform operators aware of system problemswith a brief glance.

63

ERGONOMICS IN CENTER DESIGNMany centers have relatively poor ergonomicdesign of workstations and systems. Thesecan contribute to operator stress. At onecenter, for example, there were consistentcomplaints of neck pain. Inspection revealedthat one of the primary graphics monitors waslocated substantially above the operators’seated line of regard (see figure 254, creatingstress on the operators’ neck muscles. Atmany of the centers, operators had to look overconsoles, monitors, and other equipment toview monitors on the wall. In many cases, abottom row of monitors was not fully visiblefrom a seated position.

Glare is one of the most common problemswith video display terminal (VDT) workstations.There are numerous ways of attacking thisproblem, including glare shields between VDT’sand light sources; covering of light sources,including windows; and depending on tasklighting. A major percentage of the centers didlittle to address the glare problem (see figure26).

Console work surfaces were generally at anappropriate height for average-sized operators,but there was no adjustability. One short-statured operator had to raise his chair seat tothe top of its adjustability in order to workeffectively and see over the console. This lefthis feet dangling, a common cause of legdiscomfort. A footrest should have beenavailable for this operator.

PRIVACY ISSUESThe sensors and information processingequipment available in the IVHS-era TMCprovide unprecedented opportunities forinvading the privacy of individual drivers andother citizens. It would be feasible, using GPSand cellular phone technology, to track anindividual driver from trip origin to tripdestination, including any stops for errands. Itwould be possible to calculate each driver’sspeed along the route. All this informationcould be recorded and archived for later use.

Figure 26. Example of glare on a VDT.

65

Such a data base of origin and destination (O-D)would be very useful to help predict congestion ifmany drivers were found to be converging on aparticular destination (perhaps for a specialevent). However, the data base could bemisused by police departments to automate theissuance of speeding citations.

Managers also expressed concern that an O-Ddata base would be subject to frequentsubpoena by attorneys wanting to show theroutes and stops of citizens involved in divorcecases or other litigation. For these reasons,system developers are hesitant to collect andutilize O-D data in central data processing andstorage.

The large batteries of CCTV cameras, with theirhigh magnification zoom capabilities, provideanother opportunity for privacy invasion.Several jurisdictions have already faced thedilemma of operators using the cameras for

non-highway-related activities. Metal masksmounted around the cameras now limit theCCTV field of view. This solution both serves asa deterrent to unauthorized use and providesvisible proof to the public that their privacy hasbeen restored.

A more serious dilemma is presented whenprivacy is invaded on the roadway itself. What,for example, is the proper TMC response whena crime is seen taking place on the highwayshoulder? Should a tape be made of theincident? Should police be called? Should aclose-up view of the incident be taped, includingvehicle license numbers and faces of assailantsand victims? Are such tapes considered to bepublic records available to the media? Do theappropriate answers depend on whether theoperator is also a police officer? These thornyissues are now being addressed by some of thepioneering TMC’s.

66

Section 11 lDESIGN CONSIDERATIONS BASED ON COMPARABLE SYSTEMS ANALYSIS

The lessons learned from the comparablesystems analysis lead to a number of tentativeconclusions about the human factors of TMCdesign and operation. Some of theseconclusions are based on existing humanfactors design standards, the importance ofwhich were underscored during site visits.Other conclusions are more specific to the TMCand were reached on the basis of successesand failures reported by TMC personnel.Finally, some conclusions were reached byinvestigators based on their training andexperience. These lessons are brieflysummarized here in the form of provisionaldesign considerations.

USER-CENTERED DESIGNFormal human factors inputs into TMC designcan enhance the efficiency of the center (section10)

User-centered design in the modernization of anolder center may be especially importantbecause it serves to enroll existing operators inthe design of the new center, thereby reducingany operator resistance (section 6).

The users (operators) should be involved indefining the content of specific TMC displays toensure that unneeded display screens are notincorporated (section 3).

Consoles should be designed to accommodatethe extreme ranges of staffing levels. If a singleoperator is on duty during the night shift, forexample, one single workstation combining all ofthat operator’s anticipated functions should beavailable (section 3).

Operator self-labeling of workstations providesclues to human factors design problems in theoriginal designs (section 9).

High-granularity data in a geographic informationsystem need to be fused to be of use to theoperator. The operator should have somecontrol of the degree of fusing that is performedbefore display (section 3).

Allocation of tasks between operator andcomputer should remain flexible until the impactof the allocation is tested. Some automationschemes, for example, may actually result inunexpected increases in operator workload(section 3).

Informal procedures and communicationmethods developed by operators to simplify theirjobs should be considered and, where possible,kept in place during TMC upgrade andremodeling (section 4).

Use of prototypes and control-room models thatcan be altered by operators appears to be auseful approach to user-centered redesign of acenter (section 4).

Early use of computer rapid prototypingpackages to design and redesign displayscreens can prevent many costly human factorsproblems from appearing late in theimplementation cycle (section 7).

Procedures manuals, written in the language ofthe operator (not the traffic engineer or softwaredeveloper) are an important element in TMCoperation (section 9).

STAFFINGCenters whose functions include significantinteraction with police officers in the field shouldconsider including a police liaison officer on theirstaff (section 3).

Minimum qualifications for a TMC operator aregood verbal skills, good reasoning skills, aworking knowledge of the roadway system, andsome degree of computer literacy (section 8).

Initial operator training is typically on-the-job(section 8).

MANAGEMENTWhere two or more TMC’s share geography andresources, staff members at each TMC need tobe fully aware of the capabilities, limitations, andprocedures of the other (section 3).

67

The effectiveness of the TMC in meeting itsgoals is strongly impacted by the compliancelevel of the public (section 3).

The TMC must continuously ensure that itmaintains the trust of the public (section 3).

When two or more software systems are meantto be used together, it is important to ensurethat they run on compatible platforms and havea common user interface. A significant penaltyin operator workload and error rate results fromthe use of redundant workstations andprocedures for closely related functions (section3)

IVHS technology provides unique hazards forinvasion of privacy. Center managers mustidentify and address these issues (section IO).

Centers should be designed with futureevolution in mind, including sufficient spare floorspace for potential expansions (section 6).

In centers for which special events represent themajor source of traffic variation, regularlyscheduled meetings of major traffic generatorsare helpful in developing and coordinating trafficand parking management plans (section 3).

In selecting message suites for VMSinstallations, focus groups of typical drivers maybe used to define the “meanings” of candidatewords and messages (section 5).

AUTOMATIONOlder practices of assigning tasks betweenhumans and automated systems using simplelists comparing capabilities are not effective withcomplex systems like TMC's. A hybridautomation function allocation approach isnecessary (section 5).

Allocations of function should not beunchanging; they should be changeable byeither operator initiative or computer initiative inresponse to changing situations, workload, oroperator performance (section 5).

Automated systems should keep the operator “inthe loop” and, at a minimum, report the nextmajor action (something that would takesubstantial time to undo) and allow the operatorto veto it (section 5).

lmpleme ntation of automation should be agradual, iterative process, proceeding to the

next step only when it approaches full reliabilityon the previous step (section 5).

USER-SYSTEM INTERFACESBig board displays are primarily of use in publicrelations in smaller centers. In larger centers,they may assist communication and coordinationbetween operators (section 7).

Static wall maps are the least useful type of bigboard display. Dynamic wall maps using LEDboards or computer-generated images canprovide more useful data (section 7).

Projection television should not be used whensmall, very precise text or symbols must be read(section 7).

Maintaining operator awareness of the directionin which the camera is pointed is difficult,especially at night when many landmarks arenot visible (section 7).

In some situations, two or more operators maybe competing for control of the same camera ormonitor. Methods of prioritization need to beestablished (section 7).

Operators must be made aware of VMS signswhose content is no longer appropriate (section7)

For automated voice messages, whethersampled or synthetic, the “speaker” shouldremain consistent in gender and other voicequalities to avoid distraction (section 3).

Use of gender change to draw attention to animportant voice message, or coding of messagetype by gender, however, may be a usefulapproach (section 3).

An effective “help” system in TMC softwarewould reduce error potential (section 9).

Vocabulary limitations on VMS sign content canrestrict the usefulness of the message boards(section 5).

Operators should not be prevented from takingan action that the computer considersunacceptable; neither should such actions bemade easy for them (section 5).

Error traps should be used in system commandsoftware to prevent catastrophic errors. (section7)

68

Operators, by preference, will monitor trafficusing color closed-circuit television systemsrather than well-designed support systems withgraphical user interfaces when both areavailable. This preference may result in highvisual workload (section 3).

In architectures in which support systems usesensor data to detect congestion or incidents, itis not necessary to activate the CCTV camerasand monitors until a problem is identified. Thisapproach, though, has a relatively high falsealarm rate and incidents are not detectable asearly as with CCTV’s alone (section 7).

Various levels of data fusion may be appropriatefor different tasks. The operator should have

control over the degree of data granularity(section 7).

ENVIRONMENTAL AND ERGONOMICCONSIDERATIONSCenters that require significant voicecommunication may be impaired by high noiselevels. Some means of noise control should beconsidered by such centers (section 3).

Glare on computer workstations should beavoided through appropriate placement of lightsources (section 7).

Line-of-sight problems (e.g., monitors thatcannot be easily seen behind consoles) shouldbe avoided (section IO).

69

1.

2.

3.

4.

5.

6.

7.

8.

9.

Folds, D. J.; Brooks, J. L.; Stocks, D. R.; Fain, W. B.; Courtney, T. K.; and Blankenship, S. M.(1993). Functional Definition of an Ideal Traffic Management System. Atlanta, GA: Georgia TechResearch Institute.

Petroski, H. (1985). To Engineer is Human. New York: St. Martins Press.

Kelly, M. J.; Gerth, J. M.; and West, P. D. (1994). Comparable Systems Analysis: Evaluation ofTen Command Centers as Potential Study Sites. Technical Report, FHWA-RD-93-158,Washington, DC: Federal Highway Administration.

Folds, D. J.; Fisk, A. D.; Williams, B. D.; Doss, J. E.; Mitta, D. A.; Fain, W. B.; Heller, A. C.; andStocks, D. R. (1993). Operator Roles and Automated Functions in an IVHS-Level AdvancedTraffic Management System. Atlanta, GA: Georgia Tech Research Institute.

Morrison, J. G. and Gluckman, J. P. (1991). Program Plan for the Adaptive Function Allocation forIntelligent Cockpits (AFAIC) Program. Warminster, PA: Naval Air Development Center.

Parasuraman, R.; Bahri, T.; Deaton, J. E.; Morrison, J. G.; and Barnes, M. (1990). Theory andDesign of Adaptive Automation in Aviation Systems. Warminster, PA: Naval Air DevelopmentCenter.

Andes, R. C., Jr. and Rouse, W. B. (1991). Specification of Adaptive Aiding Systems. SearchTechnology, Norcross, GA.

Ballas, J. A.; Heitmeyer, C. L.; and Perez, M. A. (1990). Interface Sty/es for Adaptive Automation.Washington, DC: Naval Research Laboratory.

Morrison, J. C.; Gluckman, J. P.; & Deaton, J. E. (1991). Adaptive Function Allocation forIntelligent Cockpits. Warminster, PA: Naval Air Development Center.

10. Woods, D. D. (1988). “The Effects of Automation on the Human’s Role: Experience from Non-Aviation Industries,” in Fight De&Automation: Promises and Realities, S. Norman and H. Orlady,(Eds.), Moffett Field, CA: NASA Ames Research Center, pp. 61-85.

11. Adams, J. A. (1993). “Virtual Reality Is For Real,” IEEE Spectrum, pp. 22-29.

71

Design and Operation of Advanced Control Centers€¦ · high-technology centers and how the...

Documents

Transcript of Design and Operation of Advanced Control Centers€¦ · high-technology centers and how the...