Making Intersections Safer: A Toolbox of Engineering ... · MAKING INTERSECTIONS SAFER: A TOOLBOX OF ENGINEERING COUNTERMEASURES TO REDUCE RED-LIGHT RUNNING Improve Signal Visibility

An Informational Report

Federal Highway AdministrationFederal Highway Administration

Red-Light Running

Making Intersections Safer:A Toolbox of Engineering Countermeasures to Reduce

Red-Light Running

An Informational Report

Federal Highway Administration

Making Intersections Safer:A Toolbox of Engineering Countermeasures to Reduce

Red-Light Running

The Institute of Transportation Engineers (ITE) is an international educational and scientificassociation of transportation and traffic engineers and other professionals who are responsiblefor meeting mobility and safety needs. ITE facilitates the application of technology andscientific principles to research, planning, functional design, implementation, operation, policydevelopment and management for any mode of transportation by promoting professionaldevelopment of members, supporting and encouraging education, stimulating research,developing public awareness, and exchanging professional information; and by maintaining acentral point of reference and action.

Founded in 1930, ITE serves as a gateway to knowledge and advancement through meetings,seminars, and publications; and through our network of more than 15,000 members working insome 80 countries. ITE also has more than 70 local and regional chapters and more than 100student chapters that provide additional opportunities for information exchange, participationand networking.

Institute of Transportation Engineers1099 14th Street, NW, Suite 300 West

Washington, DC 20005-3438 USATelephone: +1 202-289-0222

Fax: +1 202-289-7722ITE on the Web: www.ite.org

© 2003 Institute of Transportation Engineers. All rights reserved.Publication No. IR-115

500/STP/CAA/PMR/0503

ISBN: 0-935403-76-0Printed in the United States of America

LIST OF FIGURES AND TABLES v

EXECUTIVE SUMMARY vii

ACKNOWLEDGEMENTS 1

CHAPTER 1Introduction

The Problem 3Alternative Solutions 3Objective of Report 4Report Organization 4

CHAPTER 2Understanding Red-Light Running

Red-Light Running Defined 5Red-Light Running Violation Frequency 6Safety Impacts of Red-Light Running 6Crash Types Related to Red-Light Running 9Driver Characteristics 10Intersection Characteristics 12Drivers’ Stop-Go Decision 14Causal Factors and Potential Countermeasures 15Summary 16

CHAPTER 3Engineering Countermeasures

Introduction 17Improve Signal Visibility 17Improve Signal Conspicuity 24Increase Likelihood of Stopping 28Address Intentional Violations 31Eliminate Need to Stop 36Summary 39

CHAPTER 4Red-Light Running Problem Identification andResolution Process

Introduction 41The Process 42Summary 49

CHAPTER 5Future Needs

Research and Development 51Improved Crash Data for Red-Light Running 53Improved Guidelines and/or Standards 53Improved Procedures and Programs 53

REFERENCES 55

table of contents

iiiiii

vv

CHAPTER 2Table 2–1Summary of Violation Data for Iowa Cities 6Figure 2–1Red-Light Running Crashes Injury Distribution 8Table 2–2Summary of Costs Resulting from Ran Traffic-Signal Crashes in Iowa (1996–1998) 8Figure 2–2Common Crash Types Associated with Red-Light Running 9Table 2–3Summary of Intersection Characteristics onLikelihood of Red-Light Running Crash 14Table 2–4Factors Affecting the Stop-Go Decision 14Table 2–5Driver Population Types 15Table 2–6Possible Causes and Appropriate Countermeasures for Red-Light Running 16

CHAPTER 3Figure 3–1Intersection with One Signal Head 18Table 3–1Minimum Sight Distance 18Figure 3–2Post-Mounted Signal Blocked by Vehicle 19Figure 3–3Post-Mounted Signal no Longer Blockedby Vehicle 19

Figure 3–4Intersection Approach with Two Signal Heads 20Figure 3–5 Approach Updated with One Overhead Signalfor Each Through Lane 20Figure 3–6Intersection with One Lane Approach withTwo Overhead Signals 20Figure 3–7More Than One Primary Overhead Signal Per Lane, Plus Pole-Mounted Secondary and Tertiary Signal Heads 21Figure 3–8Two Overhead Signals for a One Lane Approach and Supplemental Pole-Mounted Signal 21Figure 3–9Approach View of Supplemental Signal on Near Side at Intersection 21Figure 3–10Supplemental Signal for an Intersection in theMiddle of a Reverse Curve 22Figure 3–11Comparison of Signal Heads Using8-in. and 12-in. Lenses 22Figure 3–12Louvers Used on Traffic Signals on Two Closely Spaced Approaches 24Figure 3–13Alternative Arrangements of Two Red-Signal Sections 25Figure 3–14Field Use of Two Red-Signal Sections 25Figure 3–15Traffic Signals with Backplates 26

list of figures and tables

vi vi

MAKING INTERSECTIONS SAFER: A TOOLBOX OFENGINEERING COUNTERMEASURES TO REDUCE RED-LIGHT RUNNING

Figure 3–16Signal with High Intensity Yellow Retroreflective Tape on the Backplate 27Figure 3–17SIGNAL AHEAD Sign 28Figure 3–18Example of AWF and Sign Treatment 29Figure 3–19Example of AWF and Sign Treatment 29Figure 3–20Rumble Strips and Pavement Markings Used to Alert Drivers of Signal Ahead 30Figure 3–21LEFT-TURN SIGNAL Signs at an Intersection 30Figure 3–22Factors to Consider in Signal Removal 37Figure 3–23Roundabout 38Figure 3–24Summary of Engineering Countermeasures by Category 40

CHAPTER 4Figure 4–1Red-Light Running Problem Identification 42Figure 4–2Traffic Signal Field Review Checklist 44Figure 4–3Example of Signal View Restricted Utility Wires 45Figure 4–4Placement of Rat Box at an Intersection 48Figure 4–5Enforcement Light Installation 48

CHAPTER 5Figure 5–1Infrastructure-Based ITS System that Warns Potential Red-Light Violators Approaching Intersection 52

viivii

executive summaryFHWA AND ITE BRIDGE THE GAP

In 2000, the Federal Highway Administration (FHWA)and the Institute of Transportation Engineers (ITE)initiated preparation of an Informational Report entitled,Engineering Intersections to Reduce Red-Light Running.The principal focus of the effort was to examine theengineering features of an intersection that could reducered-light running. The intended purpose of the report wasto provide information that could be used to proactivelyensure that intersections were engineered to discouragered-light running. The report was to serve as aneducational tool for law enforcement agencies and otherswho may design red-light camera systems.

In order to develop the toolbox, ITE formed a panel ofexperts from federal, state and local governments, aswell as academia and the private sector, to shareknowledge and experiences in addressing red-lightrunning using engineering countermeasures. In addition,a process was established to collect information andsurvey practicing engineers to collect the broadestinformation possible on the topic. The end result was atoolbox that identifies engineering features at anintersection that should be considered to discourage red-light running. Making Intersections Safer: A Toolbox ofEngineering Countermeasures to Reduce Red-LightRunning addresses design and operational features thatmay need to be upgraded as necessary. It provides abackground of the characteristics of the red-light runningproblem; identifies how various engineering measurescan be implemented to address this problem; suggests aprocedure for selecting the appropriate engineeringmeasures and provides guidance on when enforcement,including red light cameras, may be appropriate.

THE GROUNDWORK FORIMPLEMENTING ENGINEERINGCOUNTERMEASURES

Research cited in the report suggests that “intentional”red-light runners are most affected by enforcementcountermeasures while “unintentional” red-light runnersare most affected by engineering countermeasures. Thereport also establishes the essential need for soundengineering at an intersection for the successfulimplementation of long-term and effective enforcementactivities, particularly automated enforcement. Thereport further concludes that education initiatives can bean effective complement for any approach or as a stand-alone program in its own right. Overall, red-light runningis recognized as a complex problem requiring a reasonedand balanced application of the three “E”s.

The engineering features presented in the report arecategorized according to the type of problem theyaddress. The expected benefits of variouscountermeasures in terms of reduced red-light runningviolations or crashes are presented where data areavailable. Other countermeasures are presented whenthere is substantial confidence in their effects based onthe field experience of practicing engineers.

COUNTERMEASURES WITHPROMISE

The problems contributing to red-light running that canbe addressed with engineering countermeasures includesignal visibility, the likelihood of stopping, eliminatingthe need to stop and signal conspicuity.


Improve Signal Visibility

One recent survey shows that motorists who violate thered traffic signal frequently claim, “I didn’t see thesignal.” In fact, 40 percent of red-light runners claimthey did not see the signal and another 12 percentapparently mistook the signal indication and claimedthey had a green-signal indication. For whateverreason—motorist inattention, poor vision, poor signalvisibility—the motorist did not see the signal, andspecifically, the red signal in time to come to a stopsafely. Signal heads placed in accordance with theMUTCD should ensure their visibility for all motorists.Yet, there are locations that are still not in compliancewith the MUTCD. At a minimum, stricter adherence tothe guidelines and standards presented in the MUTCDare needed to improve signal visibility. Thecountermeasures described in the report include theplacement and number of signal heads, the size of thesignal display and line of sight.

Improve Signal Conspicuity

In addition to improving the visibility of a traffic signal,various countermeasures can be applied to capture themotorist’s attention, i.e. making the signal moreconspicuous. Redundancy by providing two red-signaldisplays within each signal head can be effective inincreasing conspicuity. LED signal lenses are beneficialin that they are brighter, which is especially helpfulduring poor weather or bright sunlight. Backplatesimprove signal visibility by providing a blackbackground around the signals, thereby enhancing thecontrast. They are particularly useful for signals orientedin an east-west direction. Finally, strobe lights areconsidered because they attract the attention of themotorist and provide emphasis on the signal.

Increase Likelihood of Stopping

Intersections and intersection devices should be carefullyengineered so that the motorist is not enticed tointentionally enter the intersection on red. This mayinclude providing additional information to the motoristregarding the traffic signal. With the additionalinformation, the probability that a driver will stop for a

red signal may increase. Additionally, the intersectionsmust be designed so that a driver who tries to stop his/hervehicle can successfully do so before entering theintersection on red. An improvement in intersectionpavement condition may increase the likelihood ofstopping by making it easier for the driver to stop. Thecountermeasures detailed in the report include signal-ahead signs, advanced-warning flashers, rumble strips,left-turn signal sign and pavement surface condition.

Address Intentional Violations

The countermeasures presented in this section of thereport are mainly intended for those violators who “pushthe limits” of the signal phasing or try to “beat” theyellow signal. Previous surveys indicate that thecommon reasons drivers speed up and try to beat ayellow light include being in a rush and saving time.Although these drivers may not have intended to violatethe red signal, they did intentionally enter towards theend of the phase knowing that there was the potential thatthey would violate the signal. Often times, these driversdo miss the yellow and end up running the red. Thecountermeasures presented relate to signal timing. Thereare many different and specific signal countermeasuresthat can be implemented regarding signal timing. Therange in countermeasures includes signal optimization,modifications to signal-cycle length, yellow-changeinterval, all-red clearance interval and dilemma zoneprotection.

Eliminate Need to Stop

Eliminating the need to stop at an intersection canobviously eliminate the potential for red-light running.This can be done by removing the signal or redesigningthe traditional intersection. Other countermeasures inthis category that are described in the report includeunwarranted signals, roundabout intersection design andflash mode for signals.

PUTTING IT ALL TOGETHER

The solution to the red-light running problem ofteninvolves a combination of education, enforcement and


viii viii

engineering countermeasures. Though the principalfocus of the report is engineering countermeasures, thereport also provides information on how an agency canidentify the existence of a red-light running problemand then select the most appropriate countermeasure orcombination of countermeasures. The report details aprocess for determining if a red-light running problemexists and what types of countermeasures could beimplemented in a logical and systematic manner. Theprocess respects the fact that individual agencies mayalready have established procedures for conductingaudits and reviews of problem intersections, which mayaccomplish the same objective.

ixix

Executive Summary

11

Edward R. Stollof, AICPContracts Senior DirectorInstitute of Transportation Engineers

Patrick F.X. Hasson, E.I.T.National Technical Service Team Leader, Safety and Highway DesignFederal Highway Administration Resource CenterOlympia Fields, IL

This report was prepared for the Institute of Transportation Engineers (ITE) with funding assistance provided by theFederal Highway Administration (FHWA). This report was prepared by BMI of Vienna, VA. Hugh W. McGee, Ph.D.,P.E., a Fellow of ITE, is the principal author. Contributions to the report were made by BMI staff, including KimberlyEccles, James Clark, Leanne Prothe and Charles O’Connell. The author is indebted to those persons and agencieswho have provided information on the engineering measures they have used to provide safer intersections. The authorwould like to acknowledge the contributions of all reviewers who have made this a better report, most notablyEdward Stollof, ITE’s Contracts Senior Director and Patrick Hasson, Fred Ranck, Hari Kalla and Scott Wainwrightof FHWA.

acknowledgements

John K. Abraham, Ph.D., P.E., MITETraffic EngineerCity of Troy, Michigan Transportation Department

Joe Bared, Ph.D., PEResearch Highway EngineerOffice of Safety Research and DevelopmentFederal Highway AdministrationWashington, DC

Leanna BelluzTransport CanadaRoad Safety and Motor Vehicle RegulationOttawa, Ontario, Canada

James A. Bonneson, P.E.Associate Research EngineerTexas Transportation InstituteTexas A&M UniversityCollege Station, Texas

Patrick M. Cawley, P.E., PTOE™EngineerCH2M HillDetroit, Michigan

Michael S. GriffithActing Director, Office of Safety DesignFederal Highway AdministrationWashington, DC

PROJECT REVIEW COMMITTEE

PROJECT DIRECTORS


2

Glenn A. HansenLieutenantPolice DepartmentHoward County, Maryland

Hari Kalla, P.E.Transportation SpecialistOffice of Safety DesignFederal Highway AdministrationWashington, DC

Michael L. Kinney, E.I.T.ITS EngineerStreet SmartsFairfax, Virginia

Robert A. Kochevar, P.E.Director of Transportation Engineering/Traffic

Operations, and City Traffic EngineerCity and County of Denver, ColoradoPublic Works DepartmentDenver, Colorado

Chi Y. Lee, P. Eng., PTOETraffic EngineerCity of Red DeerRed Deer, Alberta, Canada

Ronald D. LippsAssistant Director, Office of Traffic and SafetyState Highway AdministrationMaryland Department of TransportationHanover, Maryland

John A. McGill, P.Eng., PTOEPresidentSynectics Transportation Consultants Inc.St. Catherine, Ontario, Canada

Thomas K. Mericle, P.E., PTOECity Transportation EngineerCity of San Buena Ventura Engineering DepartmentVentura, California

Joseph S. Milazzo, II, P.E.Senior Research AssociateInstitute for Transportation Research and Education North Carolina State UniversityRaleigh, North Carolina

Dave A. MorenaSafety and Traffic EngineerFederal Highway AdministrationLansing, Michigan

John A. Nepomuceno, P.Eng.Associate Research AdministratorState Farm Insurance CompaniesBloomington, Illinois

Richard A. RettingSenior Transportation EngineerInsurance Institute for Highway SafetyArlington, Virginia

Fred N. Ranck, P.E., PTOESafety and Geometric Design EngineerFederal Highway Administration Resource CenterOlympia Fields, Illinois

Robert Blake Williams, P.E., PTOEMiami-Dade County Public WorksTraffic Signals and Signs DivisionMiami, Florida

W. Martin Bretherton, Jr., P.E., FITEChief Engineer, Gwinnett County, GeorgiaDepartment of TransportationLawrenceville, Georgia

Raymond H. Burke, P.E., FITEAssistant Director for Public WorksCity of North Las Vegas, NevadaTransportation Services DivisionNorth Las Vegas, Nevada

David A. Noyce, P.E.Assistant ProfessorUniversity of Wisconsin—MadisonMadison, Wisconsin

Weston S. Pringle, Jr., P.E., PTOEWeston Pringle & AssociatesLaguna Niguel, California

POLICY REVIEW COMMITTEE

PROJECT REVIEW COMMITTEE (continued)

33

Introduction1chapter

THE PROBLEM

One of the primary causes of crashes at signalizedintersections involves a vehicle entering an intersectionwhen the red signal is displayed. This type of collisionoccurs frequently. According to preliminary estimatesby the Federal Highway Administration (FHWA) for2001, the most recent year for which statistics areavailable, there were nearly 218,000 red-light runningcrashes at intersections (1). These crashes resulted in asmany as 181,000 injuries and 880 fatalities, and aneconomic loss estimated at $14 billion per year. Clearly,red-light running, which is reported to be on the rise aswith other aggressive driving behaviors such asspeeding, tailgating and not stopping or even slowing atstop-controlled intersections, has become a nationalsafety problem.

Red-light running is also a complex problem. There isno simple or single reason to explain why drivers runred lights. There is a tendency to cite driver error—either intentional or unintentional disregard of thetraffic signal. As will be presented in the report, red-light runners are more likely to be younger than 30-years old, have a record of moving violations, aredriving without a valid license and/or have consumedalcohol. There are elements of driver psychology andsociology behind the violations and any driver may besusceptible to committing a violation. There is alsoevidence that drivers may be induced into running redlights because of improper signal design or operation.These elements make red-light running difficult topredict and a difficult problem to solve.

ALTERNATIVE SOLUTIONS

As with many safety problems, the solution to the red-light running problem requires a combination ofcountermeasures involving the three “E”sstakeholders—education, enforcement and engineering.Educational solutions start with instructing newlylicensed drivers on the traffic laws and the rulesregarding yellow- and red-signal displays. Theycontinue with public information campaigns, such astelevision and radio public service announcements, thatalert the public of the red-light running problem and itscrash severity consequences.

Since every crash involving a red-light runner involvesa traffic violation, it is only natural that traffic lawenforcement be one of the countermeasures to consider.Enforcement includes both selective police patrols, andmore recently in some jurisdictions, automated-enforcement cameras. Traditionally, police enforcementinvolves targeted enforcement of red-light violations atintersections with a high number of violations and/orcrashes. However, this type of enforcement is laborintensive and therefore costly, and it can be hazardous,providing only short-lived effectiveness.

In some jurisdictions across the country, automated-enforcement systems, which use vehicle sensors andcameras to automatically identify a red-light runner andsubsequently issue a citation, are being used to reducethese violations. Based on a recent synthesis ofliterature related to the safety impacts of automated-enforcement programs, these systems do reduce theincidence of red-light violations and can improve

intersection safety, not only at the intersections wherethey are installed but at others within their influencearea (2). While neither thoroughly conclusive norconsistent for all intersections, these systems tend toreduce angle crashes (those that most often result fromred-light running violators) to a larger extent than theincrease in rear-end crashes that may be experienced.Overall intersection safety improvement is realizedbecause angle crashes are usually more severe thanrear-end crashes, resulting in injury and/or fatality.Nonetheless, these systems have come under scrutinyand criticism for a number of reasons related to privacyand fairness. With regard to the latter, they “catch” alltypes of red-light runners, some who violate the signalintentionally, but others who enter on redunintentionally. This may be attributed, in part, todeficiencies related to the design and/or operation of theintersection.

Numerous reports and anecdotal evidence from aroundthe United States and the world, suggest that there are anumber of engineering features of intersections thatcontribute to red-light running. For example, yellow-change intervals can be set so low that they trapmotorists into running red lights. At intersections withlimited sight distance to the signals, it can be difficultfor a motorist to see the signals in enough time to avoidrunning the red light. Since engineering deficienciessuch as these can contribute to red-light running,correcting and implementing other engineeringcountermeasures minimize the extent of red-lightrunning and can sometimes obviate the use ofautomated-enforcement systems.

OBJECTIVE OF REPORT

Often enforcement measures, whether they be selectivepolice or automated systems, are initiated beforeconsideration is given to addressing the problemthrough engineering solutions. This “toolbox” willidentify what engineering features of an intersectionshould be considered to discourage red-light running. Itaddresses design and operational features that may needto be upgraded as necessary. It is intended to provide abackground of the characteristics of the red-lightrunning problem; identify how various engineeringmeasures can be implemented to solve this problem;

suggest a procedure for selecting the appropriateengineering measures; and provide guidance on whenautomated-enforcement systems may be appropriate.

The report is intended for several types of readers.Engineers trained in the design and operation ofsignalized intersections should already be cognizant ofthe engineering measures discussed. Still, they canbenefit from being reminded of good engineeringpractice with the provision of a single informationsource focused on this topic. Law enforcement officialsshould become more sensitized to the variousengineering features that affect red-light running and besupportive of their implementation prior to takingaggressive enforcement measures. Other officials whofeel that aggressive enforcement measures, (includingautomated systems) should be implemented on a largescale basis will be made aware that engineeringmeasures have the potential to reduce red-light running,which may address the resulting safety problem moreadequately and equitably.

REPORT ORGANIZATION

Beyond these introductory remarks, the reader will find in:

�� Chapter 2, a discussion of the red-light runningproblem—what it is, who the offenders are, thecharacteristics of red-light running and the crashand severity consequences.

�� Chapter 3, an identification and discussion ofvarious engineering measures that can beimplemented to reduce red-light running andpromote a safer intersection. The measures aredescribed, and if known, their safety effectivenessis presented, as well as other considerations fordeployment.

�� Chapter 4, a systematic program for identifying ared-light running problem and selectingappropriate engineering countermeasure(s) toreduce the occurrence of violations and relatedcrashes. It also provides guidance on when andwhere automated systems may be beneficial.

�� Chapter 5, a discussion of what future actions needto be taken to address the issue and provide the bestpossible guidance for minimizing red-light running.

4 4


55

Understanding Red-LightRunning

2chapter

This chapter is provided to better understand theproblem of red-light running and its characteristicsthrough review of past research. Investigation of thered-light running problem includes:

�� Definitions;

�� Understanding the frequency;

�� Understanding safety implications;

�� Discussing the relationship between a red-lightrunner or intersection characteristics and thelikelihood of red-light running;

�� Understanding the stop-go decision process; and

�� Relating causes of red-light running to engineering,education, or enforcement countermeasures.

RED-LIGHT RUNNING DEFINED

Simply stated, red-light running is entering, andproceeding through, a signalized intersection after thesignal has turned red. According to the Uniform VehicleCode (UVC) (3), a motorist “…facing a steady circularred signal shall stop at a clearly marked stop line, but ifnone, before entering the crosswalk on the near side ofthe intersection, or if none, then before entering theintersection and shall remain standing until anindication to proceed is shown …” (section §11-202).An intersection is defined in the UVC as “… the areaembraced within the prolongation or connection of the

lateral curb lines, or if none, then the lateral boundarylines of the roadways of two highways which join oneanother at, or approximately at right angles, or the areawithin which vehicles traveling upon differenthighways joining at any other angle may come inconflict” (section §1-132). From an enforcementperspective, if the motorist stops past the stop line orinto the crosswalk, or even slightly into the physicalintersection, a citation would likely not be given. Themotorist usually has to pass through the intersection tobe cited for running a red light.

The law as stated in the UVC is considered a permissiveyellow law, meaning that the driver can enter theintersection during the entire yellow interval and be in theintersection during the red indication as long as he/sheentered the intersection during the yellow interval. As of1992, permissive yellow rules were followed by at leasthalf of the states (4). However, in other states there aretwo types of restrictive yellow laws that apply, namely:

�� Vehicles can neither enter the intersection nor be inthe intersection on red; or

�� Vehicles must stop upon receiving the yellowindication, unless it is not possible to do so safely.

In those states where the yellow phase is consideredrestrictive, it is possible that an officer might stop thedriver to discuss the law and to take appropriate actionas required. Doing such, however, includes subjectivityof the officer.

The slight differences mentioned above will surfacewhen developing a plan to address red-light running.For instance, if an automated-enforcement plan isimplemented, then the area defining the intersectionwill affect the placement of the detector loops and thelaw regarding the yellow phase may play into thedecision of a grace period. If a public information andeducation campaign is conducted, then the public shouldbe educated regarding the use of the yellow phase.

RED-LIGHT RUNNINGVIOLATION FREQUENCY

Red-light running can be considered a “big” problemwith respect to the number of violations that occur. Atwo-hour traditional enforcement effort at a high-volume intersection in Raleigh, NC resulted in 36tickets, which is a rate of 18 violations per hour or anaverage of one violation about every 3.5 minutes (min.)(5). A study conducted over several months at a busyintersection (30,000 vehicles per day) in Arlington, VArevealed violation rates of one red-light runner every 12min. and during the morning peak hour, a higher rate ofone violation every 5 min. A lower volume intersection(14,000 vehicles per day), also in Arlington, had anaverage of 1.3 violations per hour and 3.4 in theevening peak hour (6).

The Center for Transportation Research and Education(CTRE) at Iowa State University completed a study inDecember 2000 that “examined the frequency andeffects of red-light running at intersections withinselected communities and estimated the overall scopeof this practice in the State of Iowa” (7). Table 2–1summarizes the occurrence of violations that werepresented in the final report for the study. Violationswere counted by videotaping the intersection and thenanalyzing the tape. As seen in the table, the violationrate for the various intersections covers a wide range ofvalues ranging from 0.45 violations per 1,000 enteringvehicles to 6.08 (discounting the outlier value of 38.50for Intersection 1 in Dubuque).

Table 2–1 Summary of Violation Data for Iowa Cities

SAFETY IMPACTS OF RED-LIGHTRUNNING

Safety is the biggest concern associated with red-lightrunning. Safety is often measured by the number and

6 6


City Intersection Violations ViolationsNumber (one per Hour per 1,000approach per Enteringintersection) Vehicles

Bettendorf Intersection 1 1.66 2.77

Intersection 2 0.50 1.85

Davenport Intersection 1 2.25 2.61


Dubuque Intersection 1 9.78 38.50



Fort Dodge Intersection 1 0.09 0.74

Iowa City Intersection 1 3.14 6.08

Sioux City Intersection 1 0.15 0.79



West Des Moines Intersection 1 0.70 1.74

Source: Reference 7

The Problem

�� Intersection violation rates as high as18 violations per hour

�� There was a 15 percent increase in thenumber of fatal red-light runningcrashes during the past 4 years

�� For a red-light running crash there is a47 percent injury rate, which is higherthan other crash types

�� Economic impact of red-light runningcrashes is estimated at $14 billionannually

severity of the crashes occurring. Numerous statisticshave been published quantifying the problem in aspecific city, state, or across the country. Some of thefacts regarding the safety impacts of red-light runningare presented below.

�� Highway Safety Information System (HSIS) (amulti-state crash database maintained by FHWA)data from four states show that red-light runningcrashes account for 16 to 20 percent of the totalcrashes at urban signalized intersections (8).

�� In a study of police-reported crashes for four urbanareas, “ran traffic control” was the single mostcommon type of crash accounting for 22 percent ofurban crashes and 27 percent of all injury crashes(9). Of these crashes, 24 percent are the result ofred-light running, meaning that about 5 percent ofurban crashes are the result of red-light runners (10).

�� Overall, 56 percent of Americans admit to running redlights. Yet 96 percent of drivers fear a red-light runnerwill hit them when they enter an intersection (11).

�� One in three people claim they personally knowsomeone who has been injured or killed in a red-light running crash—similar to the percentage ofpeople who know someone who was killed orinjured by a drunk driver (11).

�� According to a survey conducted for FHWA,approximately 21 percent of respondents said theyfelt that drunken driving incidents are decreasing,but only 6 percent felt that red-light runningincidents were decreasing (11).

These statistics show that red-light running hasspecifically impacted the safety of signalizedintersections and is considered a very dangerous act bythe motoring public.

National/Multi-State Data

Retting et al. (9) used two national databases to quantifythe occurrence of red-light running crashes, as well asto summarize the characteristics of red-light runners.The databases include the Fatal Analysis ReportingSystem (FARS), which includes virtually all U.S.police-reported crashes involving a fatality, and the

General Estimates System (GES), which is based on anationally representative probability sample of crasheswith a varying degree of injury and property damage.As reported between 1992 and 1996, FARS dataindicated that 3,753 crashes could be attributed to red-light running, resulting in 4,238 fatalities. Also, 97percent of the 3,753 fatal crashes from 1992 through1996 involved two or more vehicles. The remaining 3percent involved pedestrians or bicycles. The FARSstatistics quoted in Retting’s report were updated toshow trends in red-light running crashes after 1996. Adefinition of a red-light running crash that bestduplicates Retting’s results isolates such crashes asthose where a vehicle was proceeding straight throughthe intersection, a driver factor as failure to obey trafficcontrol and at a signalized non-interchange intersection. Retting’s research also used GES to look at the impactof red-light running crashes beyond fatal crashes. Theresults from the GES system indicate a total of 257,849red-light running crashes during 1996, which isapproximately 4 percent of the estimated total numberof police-reported crashes. Additional conclusionsmade by Retting regarding these red-light runningcrashes are summarized below (9):

�� Red-light running crashes were more likely thanother crashes to produce some degree of injury (47percent versus 33 percent);

�� Red-light running crashes were more likely tooccur on urban roads than other fatal crashes; and

�� Red-light running crashes were somewhat morelikely to occur during the day.

For the purposes of this report, the statistics using GESwere updated using the database to reproduce the injurydistribution for 1999’s crash data. A red-light runningcrash was defined as one that took place at anintersection (either at an interchange or non-interchangelocation) controlled by a traffic signal and involvedeither (1) a driver who was charged with the violation“running a traffic signal or stop sign” or (2) the accidenttype was either a “changing traffic-way, turn into path,turn into opposite direction crash” or an “intersectingpath, straight path” crash. Using the GES weights, therewere an estimated 252,506 red-light running crashes in1999. The severity distribution for these crashes isshown in Figure 2–1.

77

Chapter 2: Understanding Red-Light Running

8 8


Figure 2–1. Red-Light Running Crashes Injury Distribution.

The data obtained from both the FARS and GESdatabases highlight the danger of red-light running.Although the number of fatal red-light running crashesseemed to peak in 1996 and has since decreased slowly,red-light running crashes still account for almost 40percent of fatal signalized intersection crashes.

Although the magnitude of the statistics presented hereis useful in portraying the size and injury severity ofred-light running crashes, caution should be taken withregard to the specific numbers reported. It is difficult tocreate a definition for a red-light running crash, with the

available database variables, that catches all true red-light running crashes and does not catch other crashtypes.

State/Local Data

The impacts of red-light running crashes are alsoreported on a state level and in smaller jurisdictions. Forexample, in December 2000, CTRE at Iowa StateUniversity completed a study of crashes and associatedcosts resulting from red-light running (7). Table 2–2shows the crash frequencies by severity type and costsfor a 3-year period in seven cities in Iowa, as well as forthe entire state.

In 1999, Oregon legislation approved a bill that allowedsix cities to conduct a 2-year demonstration project ofphoto red-light enforcement. The City of Beaverton wasthe first Oregon city to enact the program, with camerasactivated in January 2001. This action was prompted, inpart, by the fact that in the 3-year period of 1997 to

Table 2–2 Summary of Costs Resulting from Ran Traffic-Signal Crashes in Iowa (1996–1998)

Jurisdiction Fatalities Injuries* PDO** Total Crashes Total Costs

Dubuque 0 202 65 190 $ 3,115,509 Davenport 1 583 279 637 $ 11,752,603 Bettendorf 0 86 68 129 $ 1,691,487 Iowa City 0 150 125 235 $ 2,364,738 West Des Moines 0 126 70 154 $ 1,196,000 Fort Dodge 0 84 62 122 $ 1,198,732 Sioux City 1 322 146 335 $ 5,369,499 State of Iowa 12 5,881 3,435 7,138 $ 11,428,000

* Total injuries.** Number of Property Damage Only (PDO) crashes; some jurisdictions do not report all PDO crashes.Source: Reference 7

99

1999, crashes caused by red-light running had increasedby 20 percent over the prior 3-year period of 1994 to1996, and injury crashes related to red-light runningincreased by 82 percent for the same periods (12).

CRASH TYPES RELATED TORED-LIGHT RUNNING

While most red-light running crashes involve at leasttwo vehicles, crashes involving a single vehicle and analternative transportation mode (pedestrian or bicyclist)can occur. A single vehicle, hit fixed object crash couldoccur when either the running-the-red violator or theopposing legal driver takes evasive action to avoid theother and crashes into an object, e.g. a signal pole. Also,a running-the-red violator can hit a pedestrian orbicyclist who is legally in the intersection.

The two most prominent crash types involving multiplevehicles are the angle- and turning-crash types. Theangle crash is typically the offending motorist hitting orbeing hit by a vehicle legally in the intersection fromthe adjacent approach. A turning crash can occur whena left-turning vehicle collides with an on-comingvehicle from the opposite direction; either vehicle maybe the red-light violator.

Past research studies that have evaluated the effect ofcameras or other programs on red-light running focusedon three crash types. This includes the two mentionedabove (angle and turning) and also rear-end collisions.Rear-end collisions are not the result of red-lightrunning but rather the result of vehicles stopping for asignal at an intersection while others behind them donot. Some studies have noted a decrease in angle andturning crashes but an increase in rear-end crashes as aresult of concentrated enforcement of red-light running.Figure 2–2 displays the three crash types.

In developing a red-light camera effectiveness study,Persaud and Council (13) noted that the following crashtypes could be possible target crashes for a red-lightstudy:

1. Right-angle (side impact) crashes;

2. Left turn (two vehicles turning);

3. Left turn (one vehicle oncoming);

4. Rear end (straight ahead);

5. Rear end (while turning); and

6. Other crashes, specifically identified as red-lightrunning.


Figure 2–2. Common Crash Types Associated with Red-Light Running.

However, with these crash types, there are numeroussituations where a crash could have occurred that wasunrelated to red-light running.

When using an accident database to highlight red-lightrunning crashes, defining such a crash using databasevariables requires particular and detailed thought.However, a jurisdiction that is looking at a specificintersection and the problem of red-light running shouldfeel comfortable in investigating the angle, turn andrear-end crashes to monitor red-light running problems.

DRIVER CHARACTERISTICS

As mentioned in the Introduction, red-light running is acomplex problem. Causal factors range from driver- tointersection-related and there is also an element ofdriver psychology and sociology involved in the actionof violating a signal. The likelihood of committing aviolation varies from day to day, intersection tointersection and from person to person. A few studieshave been conducted to identify driver characteristics ofred-light runners. These studies have used a variety ofmethods including focus groups, field data collectionand observation and crash databases. These studiesprovide valuable information to address red-light running.

Red-Light Violators

The Department of Public Health (DPH) in SanFrancisco, CA has been very involved in the city’s red-light running program. The agency has developed a“Stop Red-Light Running” campaign that highlights theissue of red-light running to both the public and themedia through bumper stickers, billboards and pressconferences. The DPH also has conducted focus groupsto better understand the psychology of red-light runnersand hence, target campaign messages appropriately (14).

In 1998, the DPH conducted focus groups that dividedred-light runners into two groups, aggressive driversand distracted drivers. The information used in thisstudy identified the average red-light runner in SanFrancisco as a male older than 40 years of age. Thisinformation was used to better focus public educationefforts (14).

An additional set of focus groups was held in June2001, with plans to use the results in another mediacampaign. Three different groups were developed:

�� Group One—Violators who live outside of SanFrancisco but regularly drive on San Francisco’sstreets and have engaged in at least four trafficinfractions and have at least two tickets in the lastyear.

�� Group Two—Violators who live in San Franciscoand regularly drive on San Francisco’s streets andhave engaged in at least four traffic infractions andhave at least two tickets in the last year.

�� Group Three—Non-violators and pedestrians/bicyclistswho live in San Francisco and have engaged intraffic infractions no more than two times and havenot received any tickets in the last year.

Infractions, as defined in this study, include running ared light, speeding, running a stop sign and runningthrough a pedestrian crosswalk without stopping whensomeone was present (15). The findings indicate bothdifferences and similarities among the three groups. Forinstance, among those participants who lived in SanFrancisco, there was a difference regarding drivercourtesy between the violators and non-violators.Violators did not want to be “taken advantage of” whiledriving as opposed to the non-violators who had a morecourteous attitude. On the other hand, groups feltsimilar regarding red-light running. Participants spokeof running red lights because they felt the person behindthem was going to run it and they noted they were in ahurry and would do anything (including running redlights) to get to their destination more quickly.

Another study, involving field data collection, wasconducted to profile red-light violators (16). The studycompared characteristics of drivers that run red lightswith a group of drivers that had the opportunity to run ared light, but did not. Field observation, cameras anddriver records were used to record characteristics of 462violators and 911 compliers at one particularintersection in Arlington County, VA. Analysis of thefield data and a comparison between violators andcompliers indicate the following:

10 10


�� Violators were less likely to drive vehiclesmanufactured after 1991 and drive to large cars;

�� Violators were less likely than compliers to wearsafety belts;

�� Violators were generally younger, however, therewas no difference in gender distribution betweenthe compliers and violators; and

�� Violators had significantly more tickets for movingviolations and were three times more likely to havemultiple speeding convictions.

During the summer of 1999, the Social ScienceResearch Center (SSRC) of Old Dominion Universityconducted a telephone interview to learn more aboutred-light runners and driving behavior. The survey wassponsored by DaimlerChrysler Corporation, theAmerican Trauma Society and FHWA in order to gaindata for the “Stop Red-Light Running” Program. Thesurvey focused on what drivers reported to be their red-light running behaviors as opposed to their beliefs. Theresearchers acknowledge that self-reported data is onlya proxy for actual driver behaviors as respondents seekto present themselves as best possible and may stretchthe truth. However, the survey results reveal interestingtrends (17).

Overall, 55.8 percent of the respondents reportedrunning red lights. General characteristics of red-lightrunners include:

�� Younger drivers;

�� Persons without children;

�� Driving alone (the presence of passengerssignificantly decreased the likelihood of running ared light, especially child passengers);

�� Employed in jobs requiring less education orunemployed;

�� Rushing to work or school in the morning weekdayhours;

�� Driving more than two miles from home; and

�� More likely to have been ticketed for red-lightrunning.

In the same survey, drivers were also asked of theirresponse to the following scenario:

You are late for work, school, or an appointmentand have been stopped by several red lights in arow. You are approaching another intersection thathas had a yellow light for several seconds, but youknow it is about to turn red. Which of the followingwould you likely do?

A.Slow down and prepare to stop at the red light.B.Speed up to beat the red light.

Seventy-one percent of the respondents said they wouldslow down and stop while 29 percent indicated theywould speed up. Of those that would speed up, theywere asked why. The majority of the responses (69percent) were due to being in a rush and to save time.Only 12 percent reported being frustrated. Additionally,the main source of this frustration is discourteousdrivers and not congestion. This was highlighted as amajor finding to the survey “given the generalassumptions among safety experts that congestion is aleading and perhaps most important factor in predictingrisky driving actions such as red-light running oraggressive driving.”

The final survey questions dealt with the problem anddanger of red-light running. About 80 percent ofrespondents believe red-light running is a problem and99 percent believe it is dangerous. When asked: “Out ofevery 10 red-light runners, how many do sointentionally?” the mean response was more than five.However, respondents believe that less than two out of10 would be stopped or ticketed by police. The reportsummarizes that “drivers believe red-light running wasoften a choice with few legal consequences.”

1111


Old Dominion University Survey Results

�� More than 25 percent of peoplesurveyed would speed up to beat a redlight when in a hurry

�� Drivers believe red-light running isoften an intentional act with few legalconsequences

Bonneson (39) reports that in examining 10,018 signalcycles in 6 hours, 586 vehicles entered the intersectionafter the indication turned red. Of these, 84 were heavyvehicles. Overall, 0.86 percent of all heavy vehiclesviolated the red indication as compared to 0.38 percentof passenger vehicles running the red. Bonnesonconcludes that heavy vehicle drivers are twice as likelyto run red lights as are passenger drivers.

Red-Light Related Crashes

Retting et al. used crash databases to investigate theoccurrence of red-light running crashes nationwide andto investigate the characteristics of red-light runners(9). In order to compare driver characteristics, a subsetof the red-light crashes were selected. These crasheswere those involving two vehicles (as opposed toinvolvement of a pedestrian or bicyclist), both of whichwere proceeding straight through the intersection andonly one was charged with red-light running. Thecharacteristics of the violators and non-violators in thesame crashes were then compared. Some results of thecomparison are highlighted below.

From FARS:

�� Red-light violators were more likely than non-violators to be younger than age 30 and slightlymore likely to be male;

�� The driver who violated the red light was morelikely to be fatally injured than the non-violator (40percent vs. 34 percent) in the same crash;

�� Police were far more likely to report alcoholconsumption for the red-light violators (34 percentvs. 4 percent) than the non-violators;

�� In non-alcohol crashes, red-light running driversand their passengers were more likely to be killed(58 percent vs. 40 percent). However, when thered-light runner was affected by alcohol, the non-red-light running driver and passengers were morelikely to be fatally injured (55 percent vs. 41percent); and

�� Red-light runners were more likely to have beendriving with a suspended, revoked, or invaliddriver’s license and were also more likely to haveprior driving while intoxicated convictions and twoor more moving violations than the non-violators.

From GES:

�� Red-light runners were more likely to be youngerthan age 30 and slightly more likely to be male thannon-violators; and

�� Red-light drivers were slightly more likely to havebeen drinking alcohol than the non-violators (5percent vs. 1 percent). The difference was moredramatic for nighttime crashes—12 percent and 1percent for red-light runners and non-red-lightrunners, respectively.

INTERSECTIONCHARACTERISTICS

Bonneson et al. (18, 39) reviewed many past studiesregarding various intersection characteristics as theyrelate to red-light running. Three intersectioncharacteristics were highlighted as exposure factorsincluding flow rate, number of signal cycles and phasetermination by max-out. Field studies support the logicalconclusion that as more vehicles are exposed to thepotential of red-light running, the violation rate increases.The findings from that report are summarized below.

�� Flow rate or volume: Every vehicle approachingthe intersection at the onset of the yellow isexposed to the potential of red-light running. Adecision must be made to stop or proceed through

12 12


Summarizing the various studies,red-light runners are more likely to:

�� Be younger than 30 years old;

�� Have a previous record of movingviolations;

�� Be driving with a revoked, suspended,or invalid driver’s license; and

�� Be in a hurry.

the intersection. As the number of approachingvehicles increases, the number of red-light runnerswill likely increase.

�� Number of signal cycles: The more times theyellow phase is displayed, the more potential forred-light running. Hence, researchers should reportthat violation rates normalized by the number ofsignal cycles.

�� Phase termination by max-out: Actuated signalsystems operate using green extension time as longas the approach is occupied. However, the greenmay reach its maximum limit and “max-out”forcing the green phase to end regardless ofwhether the approach is occupied. Conversely, thesignal may “gap-out” because the approach hasbeen unoccupied for a set period of time. There isgreater potential for red-light running as thefrequency of max-out increases.

Bonneson cites other intersection characteristics anddriver behaviors that are considered contributoryfactors. These include the following:

�� Actuated control and coordination: This factor hasto do with driver expectancy. In actuated controlsystems, vehicles often travel in platoons throughseveral interconnected signals. Drivers expect thesignal to remain green until they pass through theintersection. As a result, drivers expect the yellowto be long enough for them to make it through theintersection so they can stay with the platoon. Thismay result in violations.

�� Approach grade: Drivers on downgrades are lesslikely to stop (at a given travel time from the stopline) than drivers on level or upgrade approaches.

�� Yellow-interval duration: Long yellow intervalscan violate driver expectancy, as drivers that stopare not “rewarded” with the red signal. In contrast,yellow intervals shorter than ITE-suggested values(19) have caught drivers off guard and resulted in ahigh number of red-light violations. (This was notdiscussed in the Bonneson report.)

�� Headway: Drivers that follow closely (headway ofless than 2 sec.) are more likely to run a red light.When a driver leaves a small headway, thefollowing car is “drawn” into the intersection as

more attention is given to the leading vehicle asopposed to the environment and traffic signal.

A similar study investigating intersectioncharacteristics was conducted using accident andintersection data for California, available in HSIS (8).The study investigated select intersectioncharacteristics and the relationship to red-light runningcrashes by developing mathematical models. The mainintersection variables of interest include the number ofcross-street lanes (surrogate for intersection width),average daily traffic (ADT) and traffic-control type.

Separate models were developed for the “mainline asentering street” and the “cross-street as entering street.”Therefore, a crash where the violating vehicle enteredthe intersection from the lower-volume road wasmodeled using “cross-street as entering street” and themainline road is considered the crossing street. Theanalysis used a total of 4,709 two-vehicle red-lightrunning crashes for a 4-year period. The findings aresummarized in the points below and in Table 2–3.

�� Effect of cross-street lanes: For the “cross-street asentering street” model, there is a 7 percent increasein red-light running crashes for each one-laneincrease on the mainline when controlling forsignal operation type, opposite street ADT and left-turn channelization. The results are different for the“mainline as entering street” where the number ofcross-street lanes had no effect on the number ofred-light running crashes.

�� Effect of ADT: The number of crashes could beaffected by both an increase of traffic on theentering street (more possibility of red-light

1313


Some Intersection Characteristics thatAffect Likelihood of Red-Light Running

�� Traffic volume

�� Actuated traffic control

�� Yellow change interval

�� Crossing-street width

�� Approach grade

runners) and an increase of traffic on the crossingstreet (the greater the possibility of hitting anothervehicle). For the “mainline as entering street,” anincrease in entering street ADT as well as anincrease in crossing street ADT both resulted in anincrease in red-light running crashes. For the “cross-street as entering street,” there was an increase incrashes with an increase in the entering streetvolume, but there was no increase in crashes with anincrease in the crossing-street (or mainline) volume.

�� Effect of traffic control: For both models, fullyactuated signals tend to have more crashes perapproaching street than approaches with semi-actuated or pre-timed signals. The models indicatea 35 to 39 percent greater number of red-lightrunning crashes at fully actuated signals ascompared to pre-timed signals.

Table 2–3Summary of Intersection Characteristics on

Likelihood of Red-Light Running Crash

Those variables shown to increase the likelihood of ared-light running crash are not negotiable designfeatures. For example, in order to reduce the likelihoodof a red-light running crash, one cannot reduce thenumber of lanes on the cross-street from four lanes totwo lanes, if the volume requires four lanes. The studyresults are helpful, however, in identifying intersectionsthat may have a high number of red-light runningcrashes and where careful engineering evaluations andenforcement may be necessary.

DRIVERS’ STOP-GO DECISION

Bonneson also discussed the factors that affect thedriver’s decision to stop or proceed through theintersection upon seeing the onset of the yellow (18).Based on earlier work by Van der Horst (20), there arethree main components of the decision process: driverbehavior (expectancy and knowledge of operation ofthe intersection), estimated consequences of notstopping and estimated consequences of stopping. Table2–4 summarizes these factors.

Table 2–4Factors Affecting the Stop-Go Decision

What if the driver makes his decision to proceedthrough the intersection based on the factors above, butends up running the red light? Bonneson divides red-light runners into two categories. The first is theintentional violator who, based on his/her judgment,knows they will violate the signal, yet he/she proceedsthrough the intersection. This type of driver is oftenfrustrated due to long signal delays and perceives littlerisk by proceeding through the intersection. The secondtype of driver is the unintentional driver who isincapable of stopping or who has been inattentive whileapproaching the intersection. This may occur as a resultof poor judgment by the driver or a deficiency in thedesign of the intersection. Bonneson further indicatesthat intentional red-light runners are most affected byenforcement countermeasures while unintentional red-light runners are most affected by engineeringcountermeasures.

14 14


Likelihood of Red-Light Running Crashes

Intersection Mainline as Cross-Street as Characteristic Entering Street Entering StreetIncreasing Number of No Effect Increasing Likelihood

Crossing-Street Lanes

Increasing ADT on Increasing Likelihood Increasing Likelihood

Entering Street

Increasing ADT on Increasing Likelihood No Effect

Crossing Street

Traffic Control Actuated Signal Actuated

Signal Type Indicates the Most Indicates the Most

Crashes Crashes

Components of the Factor Decision Process

Travel time Speed Actuated control Driver Behavior Coordination Approach Yellow interval

gradeHeadway

Estimated Consequences Threat of right-angle crash of Not Stopping Threat of citation Estimated Consequences Threat of rear-end crashof Stopping Expected delay

Source: Reference 18

1515

Milazzo et al. (10) constructs similar distinctions in red-light running driver types. Table 2–5 describes fourdifferent driver types. As pointed out in the report, “notethat any driver can assume any of these roles dependingon the situation, the driver’s current mindset and chanceitself.”

Such distinctions in driver types highlight the need fordifferent types of countermeasures. It is difficult todetermine the percentage of crashes as a result of aspecific red-light running driver type since each driveris capable of acting like any of the driver types abovedepending on the current situation. However,engineering, enforcement and educationcountermeasures are plausible solutions to address thedifferences.

CAUSAL FACTORS ANDPOTENTIAL COUNTERMEASURES

As part of a FHWA study (21) on the feasibility of usingadvanced technologies to prevent crashes atintersections, the researchers reviewed the policereports of 306 crashes that occurred at 31 signalizedintersections located in three states. Traffic-signalviolation was established as a contributing factor andthe reason for the violation was provided in 139 of thecrashes. The distribution of the reported predominantcauses is as follows:

�� 40 percent did not see the signal or its indication;

�� 25 percent tried to beat the yellow-signalindication;

�� 12 percent mistook the signal indication andreported they had a green-signal indication;

�� 8 percent intentionally violated the signal;

�� 6 percent were unable to bring their vehicle to astop in time due to vehicle defects orenvironmental conditions;

�� 4 percent followed another vehicle into theintersection and did not look at the signalindication;

�� 3 percent were confused by another signal at theintersection or at a closely spaced intersection; and

�� 2 percent were varied in their cause.

Care should be taken in interpreting this informationbecause it is self-reported, cannot be independentlyverified and is based on a small sample. If these causesare statements from the driver, it is safe to say that therewill be few who will not admit that they “intentionallyviolated the signal.” Nonetheless, from these examples,countermeasures can be identified that would addressone or more of the causes.

Countermeasures can be engineering, education, orenforcement actions. Different types of measures may


Table 2–5 Driver Population Types

Driver Type Characteristics

Reasonable and Prudent - Attentive and aware

- Does not intentionally act in a way to endanger himself or others along the roadway

Temporarily Inattentive - Temporarily distracted or inattentive

- Under adverse circumstances, an otherwise prudent driver may act in this manner

Reckless - Behavior displays a willfull disregard for the safety of himself and other drivers

- Aggressive driving behavior that could result in a crash

Mistaken Driver - Attentive driver who simply makes a mistake

- Unsuccessfully attempting to drive in a reasonable and prudent manner

be more appropriate to address the variety of causeslisted above. Table 2–6 correlates the causes discussedabove to the appropriate category of countermeasure. Acheck mark (�) signifies that the countermeasure typeis likely to address the cause, while a bullet mark (��)signifies that the countermeasure type could possiblyaddress the cause.

Table 2–6 Possible Causes and Appropriate

Countermeasures for Red-Light Running

SUMMARY

Based on what we know of the extent and nature of theproblem of red-light running, the three “E”s(engineering, enforcement and education) must beconsidered as a part of an effective solution. The three“E”s are sometimes considered separately; however, aneffective program uses all types of solutions to targetthe problem. Each “E” addresses different deficienciescontributing to red-light running, whether that of thedriver, vehicle, or intersection. However, as seen fromTable 2–6, engineering countermeasures would appearto be those that would address the majority of causes ofred-light running. The engineering countermeasuresthat aim to reduce red-light running or mitigate itsconsequences are the focus of this document and arediscussed in detail in the next chapter.

16 16


Possible Causes of Engineering Enforcement EducationRed-Light Running

1. Did not see signal ��2. Tried to beat yellow �� 3. Reported they had ��

green4. Intentional violation �� 5. Unable to stop ��

vehicle6. Followed another ��

vehicle 7. Confused by signal ��

�� – Likely countermeasure �� – Possible countermeasure

1717

INTRODUCTION

As Chapter 2 described, there are a number ofintentional and unintentional factors that cause driversto run red lights. With this information, severalengineering measures can now be developed thatreduce the occurrence of this behavior. From anengineering perspective, red-light running may bereduced if, in general, any one of these actions is taken:

�� Ensure that the traffic signal, and specifically thered display, is visible from a sufficient distance andcaptures the motorists’ attention (i.e., it isconspicuous);

�� Increase the likelihood of stopping for the redsignal, once seen;

�� Address intentional violations; and

�� Eliminate the need to stop.

If a traffic signal is the most appropriate choice oftraffic control for the intersection, it is important toensure that the motorist can see the traffic signal farenough away from the intersection so that he/she canstop safely upon viewing the yellow and red display.Then, upon viewing the yellow, and certainly the red,ensure that signal operations and conditions do notentice the motorist to intentionally or unintentionallyenter on red and ensure that a driver who tries to stophis/her vehicle can successfully do so before enteringthe intersection. Recognizing that there are some

motorists that will intentionally violate the red signal atcertain times and situations, those conditions thatencourage this behavior must be minimized. Engineersshould also examine whether or not the traffic signal isthe most appropriate choice of control for anintersection and if it can be replaced with another formof control or design that eliminates the signal andtherefore the problem.

This chapter identifies various engineering measuresthat can be grouped under these general solutions. Foreach, the measure is described, applicable designstandards or guidelines in the Manual on UniformTraffic Control Devices (MUTCD) (22) are provided,and where known, its effectiveness in reducing red-lightviolations and resulting crashes is presented. Otherconsiderations in implementation and use are notedwhere appropriate.

IMPROVE SIGNAL VISIBILITY

Motorists who violate the red traffic signal frequentlyclaim, “I did not see the signal.” As reported in Chapter2, 40 percent of those surveyed claim they did not seethe signal and another 12 percent apparently mistookthe signal indication and claimed there was a green-signal indication. While there is no doubt that many ofthese claims are false, there probably are situationswhere a more visible signal would not have beenviolated. For whatever reason—motorist inattention,

EngineeringCountermeasures

3chapter

poor vision, poor signal visibility—the motorist did notsee the signal, and specifically, the red signal in time tocome to a stop safely. The countermeasure for thisproblem is to ensure that the signal is visible from asufficient distance upstream.

Signal heads placed in accordance with the MUTCDshould be visible to all motorists approaching theintersection. Although the MUTCD requires aminimum of two signal faces be provided for the majormovement on an approach, locations such as that shownin Figure 3–1 are, unfortunately, not uncommon.Adherence to guidelines and standards presented in theMUTCD are needed to improve signal visibility.

Figure 3–1. Intersection with One Signal Head (Non-compliant with the MUTCD).

The MUTCD deals with signal visibility needs in anumber of ways. First, it requires (standard) that at leasttwo signals be provided for the major traffic movementSection 4D.15). Second, although it does not require aminimum visibility distance to the signal, it doesrequire that an advance-warning sign be used if theminimum sight distance prescribed in Table 4D–1 of theMUTCD (reproduced as Table 3–1 below) is not

satisfied. Third, there is a “cone of vision” requirementthat states that at least one traffic signal must be not lessthan 40 feet (ft.) beyond the stop line and not greaterthan 150 ft. from the stop line and within a 40-degreecone of vision centered on the center of the approachlanes. Finally, it provides standards for when 12-inch(in.) size signal heads are to be used instead of 8-in.heads.

Table 3–1Minimum Sight Distance

Placement and Number of Signal Heads

The placement (pole-mounted versus overhead) andnumber of signal heads have a profound effect on trafficsignal visibility. Numerous studies have beenconducted regarding the benefits associated with thelocation of the signal head and the number of heads perapproach. The following is a discussion on this issue.

Signals Placed Overhead

The MUTCD does not require that signals be placedoverhead rather than mounted on poles either on theroadside or in the median. Although pole-mountedsignals can serve a useful purpose (as discussed later inthis chapter) there are significant benefits to providingoverhead-signal head displays. Overhead-signaldisplays help to overcome the three most significantobstacles posed by pole-mounted signal heads, whichare: (1) they generally do not provide good conspicuity,(2) mounting locations may not provide a display withclear meaning and (3) motorists’ line-of-sight blockage

18 18



�� Placement and number of signal heads

�� Size of signal display

�� Line of sight85th Percentile Speed (mph) Minimum Sight Distance (ft.)

20 17525 21530 27035 32540 39045 46050 54055 62560 715

Source: Reference 22

to the signal head due to other vehicles, particularlytrucks, in the traffic stream. Figures 3–2 and 3–3illustrate the line-of-sight blockage often associatedwith pole-mounted signals.

Figure 3–2. Post-Mounted Signal Blocked by Vehicle.

Figure 3–3. Posted-Mounted Signal no Longer

Blocked by Vehicle.

Studies were conducted to develop guidelines forencouraging uniformity in the design of traffic-signalconfigurations and improving the performance ofsignals as critical traffic-control devices (23). Key to thiseffort were recommendations regarding traffic-signaldesign configurations, which included the recognition ofvisibility obstacles. An important finding of this studywas “that, in most cases, over-the-roadway signalswould be required to ensure adequate signal visibility.”Further studies have shown significant reduction inaccidents attributed to replacement of pole-mountedsignal heads with overhead-signal heads. For example,

in Iowa, the safety impacts of replacing pedestal-mounted signals with mast-arm-mounted signals at 33intersections, resulted in a 32 percent reduction in totalcrashes (24). In Kansas City, MO, replacement of post-mounted signals with mast-mounted signals at sixintersections contributed to a 63 percent reduction inthe number of right-angle accidents and a 25 percentreduction in the total number of collisions (25).Replacement of pole-mounted traffic signals withoverhead signals is likely to result in a decrease in totalcrashes.

In some areas, there has been a conscious decision notto use any overhead signals for aesthetic reasons.Where this is the case, this recommendation could notbe implemented, but the engineer should then considerthe other measures discussed in this chapter.

Signal for Each Approach Lane

Even though the benefits of overhead signals arerecognized, the number and placement of the signalheads are crucial to meeting the motorist’s visibilityneeds. Currently, Section 4D.15 of the MUTCD onlyrequires that “a minimum of two signal faces shall beprovided for the major movement on the approach, evenif the major movement is a turning movement.” Underthis standard, it would be acceptable to have only twosignals on an approach with three or more throughlanes. In such a scenario, the signals would not beplaced over the center of each lane but at an equaldistance, splitting the width of the three lanes.However, when a signal is positioned such that it is overthe middle of the lane, it is in the center of themotorist’s cone of vision, thereby increasing itsvisibility. The additional signal head further increasesthe likelihood that a motorist will see the signal displayfor the approach. Figure 3–4 shows a three-laneapproach intersection with the required two-signalminimum. Figure 3–5 shows the same three-laneapproach with a signal placed over each through lane.Figure 3–6 shows the use of two overhead signaldisplays for a one-lane approach.

1919

Chapter 3: Engineering Countermeasures

Figure 3–4. Intersection Approach with Two Signal Heads(Minimum).

Figure 3–5. Approach Updated with One Overhead Signal

for Each Through Lane.

Figure 3–6. Intersection with One Lane Approach with TwoOverhead Signals.

In Winton-Salem, NC, an additional signal head wasinstalled on one or more approaches at 11 differentlocations. At six locations, the additional signal headwas mounted directly over a travel lane replacing adisplay where two signal heads served three lanes. Atfour locations, the additional head was mounted on theleft side of the road in an effort to make the signalvisible earlier to traffic rounding a right curve. At oneintersection, the additional head was mounted high on autility pole to make the signal visible above a verticalcurve. For all intersections combined, right-anglecrashes caused by motorists on the intersectionapproaches where the auxiliary heads were installeddeclined by more than 46 percent combined. Totalcrashes decreased significantly at five of the 11intersections (26). Care should be taken in reporting themagnitude of these findings since the study was asimple before-and-after study that did not account forother factors such as regression-to-mean that couldhave contributed to the crash reduction.

In British Columbia, the Insurance Corporation ofBritish Columbia (ICBC) has adopted a specific designconcept to combat the traffic signal visibility problems.The ICBC co-sponsored, with municipalities and theprovincial road authority, the installation of additionalprimary overhead signals per lane, as well as theexploration into installing pole-mounted secondary (leftside) and even tertiary (right side) signal heads atsignalized intersections throughout the province (seeFigure 3–7). The ICBC researched this topic thoroughlyand concluded that there is a significant safety benefitto this type of signal-head installation (27). Thisresearch supports similar findings of an early study(23), which concluded “mixed configurations(combining overhead and post-mounted heads) aregenerally better than either all-post or multipleoverhead configurations, except that the box spanperforms as well as the mixed configurations.”

Based on these findings, the preferred signalconfiguration to improve road user visibility needs is toprovide an overhead-signal display, centered over themiddle of each approach lane (two overhead signals fora one-lane approach) and, if necessary for reasons citedin this report, supplemental post-mounted signals.Figure 3–8 provides a photograph of this signalconfiguration. For very wide approaches (perhaps four

20 20


or more lanes) it may not be practical to provide anoverhead signal for each lane. Therefore, for very wideapproaches, placement of overhead signals directlyover each lane line of the approach may be sufficient toaddress the overhead visibility needs.

Figure 3–7. More Than One Primary Overhead Signal PerLane, Plus Pole-Mounted Secondary (Left Side) and

Tertiary Signal Heads (Right Side).

Figure 3–8. Two Overhead Signals for a One LaneApproach and Supplemental Pole Mounted Signal.

Supplemental Pole-Mounted Signal onNear-Side Approach

Traffic signals should be visible from a minimumdistance as prescribed by Table 4D–1 of the MUTCD(previously reproduced as Table 3–1 on page 18) and behorizontally placed at the intersection a maximum of

150 ft. from the stop line (if a 12-in. signal lens is used)as shown in MUTCD Figure 4D–2.

There are situations where these design criteria cannotbe met. The approach to the intersection may be on acurve, which restricts the sight distance, making itimpossible to meet the visibility distance criteriawithout drastic changes to the roadway. Additionally, atwide intersections the signals may have to be placedbeyond the 150-ft. limit. Where this is the case asupplemental signal should be provided on the near-side approach. Figure 3–9 shows how this wasaccomplished at an intersection in Vienna, VA. Asillustrated in the figure, the intersection approach iscurved. Without the supplemental signal, the driverswould not be able to see the upcoming signal due to thehorizontal curvature.

Figure 3–9. Approach View of Supplemental Signal on NearSide at Intersection.

Similar to the treatment in Vienna, a supplemental-polesignal, even using a double red signal, was provided foran intersection in Naperville, IL where the signalizedintersection is in the middle of a reverse curve. Inaddition to the supplemental signal, a BE PREPAREDTO STOP WHEN FLASHING sign and flasher wereused. Prior to the installation of the supplemental signaland advanced-warning devices, there was an average ofthree severe crashes and one fatal crash per year. Afterinstallation in 1996, there only has been one severecrash and no fatal crashes to this date (28). Thesupplemental signal, which has a double red display, isshown in Figure 3–10.

2121


Figure 3–10. Supplemental Signal for an Intersection in theMiddle of a Reverse Curve.

Size of Signal Displays

Increasing the size of the display improves signalvisibility. As specified in the MUTCD, there are twonominal diameter sizes for vehicular signal lenses, 8 in.and 12 in. Combinations of these sizes can be used in asingle signal head, although an 8-in. signal lens for acircular red signal cannot be used in combination witha 12-in. signal lens for a circular green signal indicationor a circular yellow signal indication. Obviously, asignal lens that is 50 percent larger than the minimum8-in. lens will be visible from a longer distance. Figure3–11 provides a visual comparison of these differentsized traffic signals.

Figure 3–11. Comparison of Signal Heads Using 8-in. and12-in. Lenses.

The MUTCD stipulates that 12-in. signal lens(standard) shall be used under the following conditions:

A. For signal indications for approaches where road users view both traffic control and lane-use control signal heads simultaneously;

B. If the nearest signal face is between 120 ft. and 150 ft. beyond the stop line, unless a supplemental near-side signal face is provided;

C. For signal faces located more than 150 ft. from the stop line;

D. For approaches to all signalized locations for which the minimum sight distance (specified in MUTCD) cannot be met; and

E. For arrow-signal indications.

Furthermore, it is recommended (guidance) in theMUTCD, that the 12-in. signal lens (for all signalindications) be used for the following conditions:

A. Approaches with 85th percentile approach speeds exceeding 40 mph;

B. Approaches where a traffic-control signal might be unexpected;

C. All approaches without curbs and gutters whereonly post-mounted signal heads are used; and

D. Locations where there is a significant percentage of elderly drivers.

Even if none of these two sets of conditions exist, using12-in. signal lenses should be considered for all signals,and especially those displaying red indications, toincrease signal visibility.

Some implementations have considered the impact of12-in. signal lenses on crash occurrence. Under theSystematic Safety Improvement Program, the City ofWinston-Salem, NC has identified, treated andevaluated countermeasures at crash locations since1986. One such improvement was to replace existing 8-in. lenses with 12-in. lenses on at least one approach at55 locations throughout the city. Using a simple before-and-after study, Winston-Salem reported a 47 percentdecline in right-angle crashes caused by motorists on theupgraded approach at all treated locations combined (26).

22 22


Winston-Salem is not the only city to use 12-in. signallenses. In fact, the policy for the City of Troy, MI nowrequires a 12-in. lens for all signals (red, yellow, green)leaving no 8-in. lenses in Troy (29). Similarly,Naperville, IL has a policy to use 12-in. lenses to ensurethat the signal indication can be well seen. Additionally,the British Columbia Ministry of Transportation andHighways has adopted 12-in. (300-mm) signal lensesfor the red, yellow and green as the new provincialstandard for overhead (primary) signal heads (30).

Line Of Sight

The line of sight between the signal display and thepoint of required visibility is critical to the motorist’sability to see the signal head. One study (23) researchedthe importance of line of sight to signal visibility. Someof the key findings were:

“An analysis of human factors principles affectingthe design of traffic-signal configurations revealedthat the driver’s perception-response tasks dependon his position on the approach. A conceptualmodel of these tasks was developed that identifiedand defined three distinct zones on the approach.Important aspects of signal configurationsincluded placing signal indications as close tothe line of sight as possible and, also, placing atleast one signal head in a consistent locationknown to and predictable by the driver.

Optical aspects of traffic-control signals were alsoinvestigated. The major variables affecting signalconfiguration design were found to be the distanceat which the signal first becomes visible and theoffset of the signal position from the line of sight.A comparison of required signal illumination at thedriver’s eye and luminance characteristics ofcommercially available traffic signals showed that,in most cases, over-the-roadway signals would berequired to ensure adequate signal visibility. Thiscomparison also led to development of specificrules for the use of oversized signal indications.”

Therefore, the signal head should be installed as closeas practical to the projection of the driver’s line of sight.Care must be taken to eliminate obstacles, which block

the motorist’s line of sight, such as utility cables/wires,structures, vegetation, or large vehicles in the trafficstream. In addition, the following are other measuresthat can be applied to enhance the motorist’s line ofsight for improved signal visibility.

Programmable Lens (Visibility-Limited)Signals

The optically programmed or visibility-limited signalslimit the field of view of a signal. This is similar to thepurpose of louvers; however, visibility-limited signalsallow greater definition and accuracy of the field ofview. For example, programmable lens are used tocontrol the motorist’s lateral or longitudinal field ofview. Lateral separation is useful for instances such asseparating left- or right-turn lanes or locations withadjacent parallel roadways like a frontage road. Anexample of longitudinal separation (also known asdistance separation) is for closely spaced intersections.Additionally, visibility-limited signals do not reduce theintensity of the visible light and also do not have theproblem of snow and ice build-up or bird nests assometimes incurred with louvers and visors.

The MUTCD speaks of visibility-limited signals mostlywith regard to left-turning traffic at an intersection. Twoexamples are presented below:

�� At a left turn operating under the protected mode,either a LEFT-TURN SIGNAL sign or a visibility-limited circular red signal must be used.

�� With protected/permissive phasing of the left turn,if the circular green and circular yellow signalindications in the left-turn signal face are visibility-limited from the adjacent through movement, theleft-turn signal face shall not be required tosimultaneously display the same color of circularsignal indication as the signal faces for the adjacentthrough movement.

Additionally, the MUTCD permits the use of visibility-limited signal faces in situations where the road usercould be misdirected, particularly when the road usersees the signal indications intended for otherapproaches before seeing the signal indications for theirown approach.

2323


There are a few concerns and extra precautionsnecessary when working with visibility-limited signals.Because the field of view is restricted and requiresspecific alignment, these signals require rigid mountinginstead of suspension on overhead wires. Additionally,there is some concern associated with glare and thelimitations of seeing the signal. These signals have alsobeen known to create driver confusion in a few specificinstances. In these instances, the signal initially appearslike there is no indication—a malfunction. However, asthe driver gets closer to the intersection and even passesthrough, they notice the signal does indeed show anindication. At that point it may be too late to stop as thevehicle is already in the middle of the intersection,waiting to make a left turn. Signal-visibility alignmentrequires attention both in design and in fieldmaintenance.

Visors

The addition of a visor to a traffic-signal head that is indirect sunlight can improve visibility of the signal byproviding additional contrast between the lens and thesignal head. There are different types of visorsincluding complete circle (or tunnel), partial (or cut-away) and angle visors. Cut-away visors are preferredas snow and water cannot accumulate at the bottom ofthe signal indications. Additionally, cut-away visorsreduce the problem of birds nesting in the visor.

The MUTCD requires that “in cases where irregularstreet design necessitates placing signal faces fordifferent street approaches with a comparatively smallangle between their respective signal lenses, each signallens shall, to the extent practical, be shielded or directedby signal visors, signal louvers, or other means so thatan approaching road user can see only the signallens(es) controlling the movements on the road user’sapproach.” Additionally, the inside of signal visorsshould have a dull black finish to minimize lightreflection. The MUTCD also recommends using signalvisors, which direct the light without reducing theintensity of the light, in lieu of signal louvers.

Louvers

Louvers are used to avoid confusion on intersectionapproaches where approaching motorists may be able to

see the signal indication for another approach, typicallydue to a skewed approach angle at the intersection. Thepurpose of a louver is to block the view of the signalfrom another approach. They are similar to angle visorsbut are better in limiting signal visibility to a narrowcone to the front of the signal. The problem withlouvers is that they reduce the amount of light emittedfrom the signal and require higher luminance to obtainthe same visibility as a signal without a louver.

As stated in the discussion of visors, louvers may beused to limit the view of the signal by approachingmotorists at intersections with a small angle betweensignal lenses. However, it is stated in the MUTCD thatsignal visors should be considered as an alternative tosignal louvers because of the reduction in light emittedcaused by the louvers. The MUTCD requires the entiresurface of louvers have a dull black finish to minimizelight reflection and to increase contrast between thesignal indication and its background. Figure 3–12shows louvers on traffic-signal heads at an intersectionwith closely spaced approaches.

Figure 3–12. Louvers Used on Traffic Signals on TwoClosely Spaced Approaches.

IMPROVE SIGNAL CONSPICUITY

In addition to improving the visibility of a traffic signal,various countermeasures can be applied to capture themotorist’s attention, i.e. to make the signal moreconspicuous. The following are measures, some ofwhich are found in the MUTCD, that should beconsidered in improving the signal conspicuity.

24 24


Redundancy

Providing two red-signal displays within each signalhead should increase the conspicuity of the red displayand further increase the likelihood that the driver willsee the signal. While “doubling-up” on the red signalsection is not normally needed, where there is a highincidence of red-light running, the engineer may wantto consider this option. It is permitted by the MUTCDto repeat a signal indication within the same signal face(section 4D.18). The proper alternative arrangements oftwo red-signal sections are illustrated in Figure 3–13,excerpted from the MUTCD. A photograph of thismeasure is shown in Figure 3–14.

Figure 3–13. Alternative Arrangements of

Two Red-Signal Sections.

Source: MUTCD Figure 4d-3.

An evaluation of this treatment applied at nine locationsin Winston-Salem, NC showed a 33 percent decrease inright-angle crashes caused by motorists on the upgradedapproaches following implementation (26).

Figure 3–14. Field Use of Two Red-Signal Sections.

Light Emitting Diode (LED) SignalLenses

An LED traffic signal module is made up of a lens andan array of individual LEDs that are tiny, purelyelectronic lights that display a single color. Each LED isabout the size of a pencil eraser. LEDs can be used toreplace an incandescent lamp and colored lens thatmake up a traditional traffic-signal optical unit.

LED units are used for three main reasons: they arevery energy efficient, are brighter than incandescentbulbs and have a longer life increasing the replacementinterval (31). For example, in producing the sameamount of light as a traffic signal with incandescentlamps, a LED traffic signal uses 90 percent less power.LEDs also emit light that is brighter because the LEDsfill the entire surface of the traffic bulb and also provideequal brightness across the entire surface. Literatureproviding data that would substantiate that LED signalsare brighter compared to incandescent bulbs could notbe found. However, it was observed by the author thatat two intersections on an arterial in South Carolina,where one of the two signal displays was replaced witha red LED signal, that the LED signals were noticeablybrighter and more conspicuous than the adjacent signalwith the incandescent bulb. Finally, LED traffic signalmodules have service lives of 6 to 10 years as comparedto incandescent bulbs that have a life expectancy ofonly 12 to 15 months.

2525



�� Redundancy

�� LEDs signal lenses

�� Backplates

�� Strobe lights

Section 4D.18 of the MUTCD supports the use of LEDtraffic-signal modules as a traffic-signal optical unit.The MUTCD states that traffic signals should conformto ITE standards (32) with regard to the intensity anddistribution of light for a signal indication. At this time,only the red and green lamps meet specifications fortraffic applications. Arrows and yellow LEDs that meetspecifications are not currently available.

As stated previously, one of the major benefits of LEDtraffic-signal modules is that LEDs are brighter, whichis especially helpful during poor weather or brightsunlight. However, there may be the potential for aglare problem at night because of the brightness.

The light output of LED traffic-signal modules is moredirectional than the output for traffic-signal opticalunits with incandescent lamps. As a result, signalindication visibility for some installations is limited to anarrow cone of vision below the horizontal axis of thesignal face. If signal heads with LED traffic-signalmodules are used, they should be mounted on mastarms. If they are installed on a span wire they arevulnerable to wind that can make the signal head tiltbackwards and forwards, making the signal appear to bein flash mode. This is commonly referred to as“blanking.” Even if they are mounted on mast arms,there is the possibility that the signal indication may notbe visible due to the approach grade to the intersection.“Blanking” can be avoided by tethering the signal to apole. If tethering is not a viable solution or if reducedvisibility is caused by the approach grade, the LEDapplication can be modified to an expanded pattern thatincreases the vertical visual cone of the LED byincreasing the LED count and modifying the lenses (33).

Laboratory research has found that with brighter lights,there are quicker reaction times and fewer missedsignals among test subjects (34). Although it wasdifficult to find a field study that confirmed the effect ofLEDs on intersection safety as measured by signalviolations, many cities are installing LEDs. Forexample, in 1998, the City of Scottsdale, AZ initiated aprogram to convert all of the city traffic signal’s red andgreen indications to LEDs. The city stated four waysthat LEDs improve safety at signalized intersections,including a reduction in signal indication outages,elimination of “phantom illumination” caused by

colored lenses, longer re-lamping cycle (which reducesthe time traffic is disrupted due to maintenance) and theability to operate on battery backup systems duringpower outages (35). Bonneson’s research (39) revealsthat the use of yellow LEDs may reduce red-lightrunning by 13 percent.

Backplates

Backplates, as shown in Figure 3–15, are commonlyused to improve the signal visibility by providing ablack background around the signals, therebyenhancing the contrast. They are particularly useful forsignals oriented in an east-west direction to counteractthe glare effect of the rising and setting sun or areas ofvisually complex backgrounds. Guidance for their usefor target value enhancement against a bright sky orbright or confusing background is provided in theMUTCD (Section 4D.17) for these very conditions. TheMUTCD (Section 4D.18) requires the front surface ofthe backplate to have a dull black finish “to minimizelight reflection and to increase contrast between thesignal indication and its background.” In manyjurisdictions, it is general practice to use backplates forall signal heads, not just those in the east-west direction.

Figure 3–15. Traffic Signals with Backplates.

At six locations of varying types in Winston-Salem, NCbackplates were added to signal displays on one ormore approaches to call attention to the signal display.Angle crashes caused by motorists on the approachwhere the backplates were installed declined by morethan 31 percent at all locations combined (26).

26 26


In British Columbia, Canada, an evaluation wasconducted of high-intensity yellow retroreflective tapeon the backplates of signals at six intersections on anarterial in Saanich (36). The authors hypothesized thatthe framed signal heads would be more visible tomotorists at night and the safety of the intersectionwould improve. Figure 3–16 shows a daytimephotograph of a signal head with the high intensity tapearound the backboard.

A comparison of crash frequency for a three-year periodafter installation, compared to one year before, showedthat the number of night crashes stayed the same thefirst year (14 crashes) but decreased significantly (fiveand three crashes) in the subsequent two years. Volumelevels actually increased in each of the four years (36).The use of retroreflective tape on the backplate iscontrary to the MUTCD standard requiring a dull blackfinish. Hence, its use in the United States would requireexperimentation approval from FHWA.

Figure 3–16. Signal with High Intensity YellowRetroreflective Tape on the Backplate.

One precaution to note when using backplates is theadditional load on the mast arm or cable caused by thebackplate. This is due to the additional weight andadded wind load. The additional load must beincorporated into the design of the signal-head mastarm and/or other signal parts.

Strobe Lights

Strobe lights have been used in rare occasions as asupplement to red signals to attract the attention of themotorist and provide emphasis on the signal. A strobelight, which flashes within the red signal display, canhave either of two shapes—halo or horizontal—and beeither of two mechanisms—xenon tube or light emittingdiode (LED). Typically, a strobe light will have a flashrate of once per second. They can be used with bothincandescent and LED signals.

One state’s guidelines (37) for the use of strobe lightsare as follows:

�� At a location with high approach speeds (> 45mph) and a documented accident problem;

�� On an approach to the first signal in a series oftraffic signals;

�� On long, flat unobstructed approaches that alterperception at high speeds;

�� At isolated intersections; and

�� To be used only after other standard traffic controldevices have failed.

There has not been a comprehensive study of thisdevice, and the present limited evaluation has shownmixed results. A 1994 study in Virginia of sixintersections with one and two strobe lights had bothincreases and decreases in rear-end and angleaccidents—accident types that should be affected bythis measure (38). The overall conclusion of theresearcher was that there is no clear benefit from usingstrobe lights and that other measures, cited in thisreport, should be tried.

While mentioned as a possible measure, care should betaken in using this device. Because conclusive evidencehas not shown a reduction in crashes, FHWA’s currentposition is that they will no longer be approved forexperimentation. Since this device is not approved byFHWA and not included within the MUTCD, an agencyusing this device may be subject to liability in the eventof litigation resulting from a crash.

2727


INCREASE LIKELIHOOD OFSTOPPING

Recalling the general solutions to red-light runningpresented in the introduction to this chapter, the secondsolution is to increase the likelihood of stopping for thered signal, once seen. Intersections and intersectiondevices should be carefully engineered so that themotorist is not enticed to intentionally enter theintersection on red. This may include providingadditional information to the motorist regarding thetraffic signal. With the additional information, theprobability that a driver will stop for a red signal mayincrease. Additionally, the intersections must bedesigned in such a way that a driver who tries to stophis/her vehicle can successfully do so before enteringthe intersection on red. For example, an improvement inintersection pavement condition may increase thelikelihood of stopping by making it easier for the driverto stop. These types of countermeasures deal with thefollowing reported causes of red-light running:

�� Driver reported they had green;

�� Followed another vehicle into the intersection onred;

�� Did not see signal; and

�� Confused by signal indication.

Most of the countermeasures discussed in this sectionare not innovative or required intersection elements, butrather are treatments used occasionally for specificreasons at targeted locations. Installation of thesecountermeasures requires a careful evaluation of thelocation and use of engineering judgment. The

evaluations of specific implementations discussedbelow provide useful information when addressing thesolution to a specific location.

Signal Ahead Sign

When the primary traffic-control device used is a trafficsignal, the appropriate sign is the SIGNAL AHEADsign (W3-3) shown in Figure 3–18.

The MUTCD requires an advance traffic controlwarning sign when “the primary traffic-control deviceis not visible from a sufficient distance to permit theroad user to respond to the device.” For a traffic signal,the visibility criterion is based on having a continuousview of at least two signal faces for a distance specifiedin Table 4D–1 of the MUTCD (See Table 3–1). TheMUTCD also permits the use of this device even whenthe visibility distance is satisfactory. In addition, theMUTCD allows for the use of a warning beacon withthe sign typically flashing yellow lights on either side oron top and bottom of the sign. The placement of thissign prior to the intersection is a function of theapproach speed. Table 2C–4 in the MUTCD providesthe recommended distances.

Figure 3–17. SIGNAL AHEAD Sign (W3-3).

28 28


Increase Likelihood of Stopping Through:

�� Signal ahead signs;

�� Advanced warning flashers;

�� Rumble strips;

�� Left-turn signal sign; and

�� Pavement condition.

Advance-Warning Flasher

The purpose of an advance-warning flasher (AWF) is toforewarn the driver when a traffic signal on his/herapproach is about to change to the yellow and then thered phase. In North America, there are three generaltypes of advanced warning devices and the decision ofwhich to use is based on engineering judgment. TheseAWFs include:

�� Prepare to stop when flashing—A warning sign,BE PREPARED TO STOP with two yellowflashers that begins to flash a few seconds beforethe onset of the yellow and continue to flashthroughout the red phase. A WHEN FLASHINGplaque is recommended in addition to the sign.

�� Flashing symbolic signal ahead—Similar toprevious type except the wording on the sign isreplaced by a schematic of a traffic signal. Theflashers operate as above.

�� Continuous flashing symbolic signal ahead—Thesign displays a schematic of a traffic-signal symbolbut in this case, the flashers operate continuously(i.e. they are not connected to the signalcontroller).

Examples of different field implementation of the signsare shown in Figure 3–19 and Figure 3–20.

Figure 3–18. Example of AWF and Sign.

Figure 3–19. Example of AWF and Sign Treatment.

The effectiveness of AWFs is measured by vehiclespeeds approaching the intersection, the number of red-light violations and its effect on accidents. A before-after study was conducted at one intersection inBloomington, MN (40). At the intersection, BEPREPARED TO STOP and WHEN FLASHING signswere pedestal mounted and accompanied by dual 8-in.yellow beacons. Data were collected immediately afterinstallation of the AWFs and again one year afterinstallation. During the before-period, the yellowinterval was 6 sec. and the all-red interval was 2 sec.After the installation of the AWFs, the all-red wasreduced to 1 sec. Data collected included the number ofred-light violations and speeds, vehicle type (car versustruck), time after the onset of the red interval when theviolation occurred and time of day of violation.

From these data parameters, the authors concluded theAWFs were effective in reducing the number of overallred light violators, the number of trucks violating thered light and the speed of the violating trucks. One yearafter installation, there was still a reduction in overall,car and truck red-light violations, as well as a slightdecrease in the average speed of violating trucks.However, this was an increase from the previous yearsafter-data indicating that the effectiveness haddecreased over time. Additionally, the study did notemploy a control or comparison group of intersections.Therefore, the changes observed could have been due tosomething other than the AWFs (for example,regression-to-mean).

2929


Another study utilized and analyzed data from BritishColumbia using two different methods (41). Modelswere used to develop expected accident rates at 106signalized intersections for total, severe and rear-endaccidents. Twenty-five of these intersections had AWFs.Although the results indicate that intersections withAWFs have a lower frequency of accidents, thedifference between those with AWFs and those withoutis not statistically significant. An additional before-and-after study was performed for the 25 intersectionsequipped with AWFs to estimate the accident reductionspecific to each location and its approach volumes. Acorrelation was found between the magnitude of theminor approach traffic volumes and the accidentreduction capacity of AWFs, showing that AWF benefitsexist at locations with moderate to high minor approachtraffic volumes (minor street AADT of 13,000 or greater).

Rumble Strips

Another warning device that has been used to alertdrivers to the presence of a signal are transverse rumblestrips. Rumble strips are a series of intermittent, narrow,transverse areas of rough-textured, slightly raised, ordepressed road surface (22). The rumble strips providean audible and a vibro-tactile warning to the driver.When coupled with the SIGNAL AHEAD warning signand also the pavement marking word message—SIGNAL AHEAD—the rumble strips can be effectivein alerting drivers of a signal with limited sightdistance. This treatment is illustrated in Figure 3–21.

There are no known studies reporting on how thistreatment can reduce red-light violations or theresulting crashes; hence their use should be restricted tospecial situations. If used, they should be limited tolower-speed facilities (less than 40 mph) and bereserved for locations where other treatments have notbeen effective.

Figure 3–20. Rumble Strips and Pavement Markings Usedto Alert Drivers of Signal Ahead.

Left-turn Signal Sign

When a motorist wanting to turn left approaches asignalized intersection using left-turn protected onlymode, he/she may be confused with the combination oftwo or more signals displaying a green ball for thethrough movement and a left-turn signal displaying ared ball or red arrow. To compensate for this, a sign thatclearly identifies the left-turn signal is to be used. TheLEFT TURN SIGNAL sign provides additionalinformation not given in the actual signal indication tothe driver by specifying the control device for differentintersection movements. This is illustrated in Figure3–22. Such information may eliminate driver confusionwhen approaching an intersection and prevent red-lightrunning for left-turning traffic.

Figure 3–21. LEFT-TURN SIGNAL Signs at anIntersection.

30 30


The MUTCD provides information regarding the use ofthe sign for different modes controlling the left-turns(protected, permissive, protected/permissive, variableleft-turn) and signal arrangements for an approach(shared versus separate). For instance, (See MUTCDSection 4D.06.C1) under protected/permissive left-turnphasing and separate signals for the left-turn andthrough movements that do not display the samecircular signal indications, a LEFT TURN SIGNALsign (R10-11) and a LEFT TURN YIELD ON GREEN(R10-12) sign is required. For protected left-turnphasing where the left-turn signal includes a circularred, left-turn yellow arrow and a left-turn green arrow,the circular red must not be seen by the through trafficor the signal must clearly be designated as the left-turnsignal. The circular red can be limited by using hoods,shields, louvers, positioning, or design. Alternatively, aLEFT TURN SIGNAL sign can be used.

Pavement Surface Condition

According to NHTSA, 2,627 fatal crashes and 215,000injury crashes in the year 2000 occurred during rainyweather conditions. This is approximately 7 percent ofall fatal crashes and 10.4 percent of injury crashesoccurring on wet pavement. Additionally, another 2.4percent of fatal crashes and 3 percent of injury crashesoccurred during snowy or sleeting weather conditions,likely on wet pavement (42).

As a vehicle approaches a signalized intersection andslows to stop for a red light, it may be unable to stopdue to poor pavement friction and as a result, proceedinto the intersection. A vehicle will skid during brakingand maneuvering when frictional demand exceeds thefriction force that can be developed at the tire-roadinterface. The friction force is greatly reduced by a wetpavement surface. A water film thickness of 0.05 mmreduces the tire pavement friction by 20 to 30 percent ofthe dry surface friction. Therefore, countermeasures toimprove the pavement condition should seek to increasethe friction force at the tire-road interface and alsoreduce water on the pavement surface (43).

The coefficient of friction is most influenced by speed;however, many additional factors affect skid resistance.This includes the age of the pavement, pavement

condition, traffic volume, road surface type and texture,aggregates and mix characteristics, tire conditions andpresence of surface water. Countermeasures to improveskid resistance include asphalt mixture (type andgradation of aggregate as well as asphalt content),pavement overlays and pavement grooving. Additionally,countermeasures such as SLIPPERY WHEN WETsigns and reducing the speed limit can also be used.

The MUTCD permits a SLIPPERY WHEN WET signto be used to warn of a possible slippery condition. Thesign is to be placed an appropriate distance prior to thecondition and at appropriate intervals along the affectedsection.

ADDRESS INTENTIONALVIOLATIONS

The third general solution presented at the beginning ofthis chapter is to remove the reasons for intentionalviolations. The countermeasures presented in thissection are mainly intended for those violators who“push the limits” of the signal phasing or try to beat theyellow signal. Previous surveys indicate that thecommon reasons drivers speed up and try to beat ayellow light include being in a rush and saving time.Although these drivers may not have intended to violatethe red signal, they did intentionally enter towards theend of the phase knowing that there was the potentialthat they would violate the signal. Often times, thesedrivers do miss the yellow and end up running the red.The countermeasures presented in this section all relateto signal timings.

3131


Address Intentional Violations Through:

�� Signal optimization;

�� Signal cycle length;

�� Yellow change interval;

�� All-red clearance interval; and

�� Dilemma zone protection.

Installing the optimum signal timings is important toensure respect for traffic signals. The MUTCDrecommends signal timings be reviewed and updated ona regular basis (every 2 years) to ensure that it satisfiescurrent traffic demands. There are many different andspecific signal countermeasures that can beimplemented regarding signal timing. The range incountermeasures includes changes to the signal system(such as progression) as well as changes to the signal-cycle length and individual signal phases (such as theyellow interval). Some of these countermeasures arediscussed in the following sections beginning withsystem level changes and narrowing to changes inspecific signal phases.

Signal Optimization

Poor signal timings are not only inefficient, but maycause a driver to become frustrated and respondinappropriately to the signal. The traffic demands ateach intersection must be carefully accounted for whenthe phase sequence and timings are developed. Oncethese timings are developed, the relationship of thesignal to other signals must be considered.

Interconnected signal systems provide coordinationbetween adjacent signals and are proven to reducestops, reduce delays, decrease accidents, increaseaverage travel speeds and decrease emissions. Anefficient signal system is also one of the most cost-effective methods for increasing the capacity of a road.With reduced stops, the opportunity to run red lights isalso reduced. In addition, if drivers are given the bestsignal coordination practical, they may not be ascompelled to beat or run a red signal.

Signal-Cycle Length

Proper timing of signal-cycle lengths can reduce driverfrustration that might result from unjustified short orlong cycle lengths. Timing of the various signal phasesis based on the characteristics of the intersection and theindividual approaches. Signal timing includes thegreen, yellow and red phase for each approach as wellas the overall signal-cycle length.

Although there are federal and state standards thatbound signal timing, there are also local or regionalpractices of signal timing. There are philosophies andconsiderations that support both shorter and longercycle lengths for reducing signal violations. The effectsof cycle length vary on traffic and driver. Drivers andtraffic engineers may perceive shorter cycle lengths asmore efficient as vehicles have shorter periods wherethey have to remain stopped. A driver that knows thewait is not excessive may be less inclined to beat theyellow or run the red. Conversely, under higher trafficvolumes, the short cycle length may not be sufficient toclear all queues and drivers may find themselves waitingthrough two or more cycles. This may cause an increase indriver anxiety resulting in an increase in driversattempting to beat the yellow and violate the red signal.

With longer cycle lengths, drivers strive to get throughthe signalized intersection or suffer the perceived longdelay associated with sitting for the red signal.However, many traffic engineers use longer cyclelengths to move significant volumes on the mainline ofarterial roadways. By providing a sustainableprogression along a corridor, the saturated roadway canmove higher volumes and reduce queue lengths.Delays associated with numerous start-up times are alsodiminished if progression is maintained.

When comparing cycle lengths, it should be noted thatwith longer cycle lengths, there are actually fewernumbers of times per hour when drivers are confrontedwith the yellow and red signal intervals. For example,when comparing a cycle length of 1 min. to 2 min., inan hours time in the 1-min. cycle, there will be twice asmany opportunities for drivers to be confronted with thechanging signal from green to red. Consequently, thelonger cycle length does reduce the number ofopportunities for traffic-signal violations.

After consideration of the pros and cons, one of the besttools to utilize in determining signal-cycle length iscomputer simulation and optimization. The computergenerated optimized cycle length combined with thetraffic engineers’ knowledge and experience shouldresult in the most efficient traffic-signal timingpractical. As part of signal-timing strategies, the need toaddress specific times of day should be included. Forexample, typical timing plans would include multiple

32 32


plans to accommodate the morning or afternoon peakperiods, midday, late night, weekends, etc.

Yellow-Change Interval

The MUTCD (22) requires that a yellow-signalindication be displayed immediately following everycircular green or green-arrow signal indication. It isused to warn vehicle traffic that the green-signalindication is being terminated and that a red indicationwill be exhibited immediately thereafter.

A properly timed yellow interval is essential to reducesignal violations. An improperly timed yellow intervalmay cause vehicles to violate the signal. If the yellowinterval is not long enough for the conditions at theintersection, the motorist may violate the signal.Motorists have some expectancy of what the yellowinterval should be and base their decisions to proceed orstop based on their past experiences. In order to reducesignal violations, the engineer should ensure that theyellow interval is adequate for the conditions at theintersection and the expectations of the motorists.

In many jurisdictions, the yellow-change interval isfollowed by an all-red interval. During this all-red“clearance” interval, the red-signal indication isdisplayed to all traffic. The yellow interval and all-redinterval are often referred to collectively as the changeperiod or change interval. The all-red interval isaddressed separately in a subsequent section.

There is currently no nationally recognizedrecommended practice for determining the changeinterval length, although numerous publicationsprovide guidance including the MUTCD (22), TrafficEngineering Handbook (44), and the Manual of TrafficSignal Design (45). The MUTCD provides guidancethat a yellow-change interval should have a duration ofapproximately 3 to 6 sec., with the longer intervalsreserved for use on approaches with higher speeds.

In the current edition of ITE’s Traffic EngineeringHandbook (44), a standard kinematic equation is providedas a method to calculate the change interval length. Theequation for calculating the change period, CP, is as follows:

[1]

The principal factors that are taken into account in thedevelopment of the change period are:

�� Perception-reaction time of the motorist, t,typically 1 sec.;

�� Speed of the approaching vehicle, V, expressed inft./sec.;

�� Comfortable deceleration rate of the vehicle, a,typically 10 ft./sec.2;

�� Width of the intersection, W;

�� Length of vehicle, L, typically 20 ft.; and

�� Grade of the intersection approach, g, in percentdivided by 100 (downhill is negative).

The equation allows time for the motorist to see theyellow signal indication and decide whether to stop orto enter the intersection. This time is the motorist’sperception-reaction time, generally 1 sec. It thenprovides time for motorists further away from the signalto decelerate comfortably and motorists closer to thesignal to continue through to the far side of theintersection. These times are dependent on thecharacteristics of the traffic and the roadwayenvironment. If there is a grade on the approach to theintersection, the equation adjusts the time needed forthe vehicle to decelerate.

If available, the 85th percentile speed should be used asthe approach speed in this equation. In the absence of85th percentile speed, some jurisdictions use postedspeed as the approach speed. In most cases, using the85th percentile speed will produce intervals that aremore conservative (i.e., longer). In no case should theapproach speed used in the calculation be less than theposted speed limit.

The deceleration rate of 10 ft./sec.2 suggested by ITE isbased on a comfortable deceleration rate that has beensupported by research. The 2001 American Associationof State Highway and Transportation Official’s A Policyon Geometric Design of Highways and Streets,otherwise known as the “Green Book,” (46)

3333


recommends 11.2 ft./sec.2 for determining stopping-sight distance. They note that this is a comfortabledeceleration for most drivers. The deceleration ratesuggested by ITE is a more conservative decelerationrate for purposes of calculating the yellow interval andwill result in longer intervals.

Effectiveness of Decreasing Violations

Various studies have evaluated the relationship betweenthe length of the change interval and the occurrence ofsignal violations. Retting and Green (47) examined red-signal violations in New York where the yellow or all-red intervals were shorter than a practice proposed byITE in 1985 (48) that is similar in calculation toEquation 1. They conducted a before-and-after studywith a control group at 20 approaches. For the after-period, the researchers retimed the yellow interval atfour sites, the all-red interval at five sites, and both theall-red and the yellow at four sites. Seven sites wereused as the control group. The yellow retimingincreased the yellow change interval by 0.5 to 1.6 sec.,depending on the intersection. The all-red retimingincreased the red-clearance interval by 0.8 to 3.6 sec.The researchers recorded the number of cycles withred-signal violations and the number of cycles with lateexits at the intersections. Red-signal violation cycleswere defined as cycles where at least one of the vehicleson the approach entered the intersection on red. Late-exit cycles were defined as cycles where at least onevehicle from the approach was still in the intersection atthe release of conflicting traffic.

Logistic regression was used to analyze the data. Theresearchers concluded that increasing the length of theyellow signal towards the ITE proposed practicesignificantly decreased the chance of red-signalviolations. They also found that late exits decreased asthe all-red interval increased. It appeared that sites withshorter yellow signals had more late exits. Increasingthe yellow to ITE-suggested values was as effective asincreasing the all-red clearance interval at decreasingthe occurrence of late exits.

Wortman et al. (49) conducted a similar before-and-after study at two intersections in Arizona. In the after-period, the yellow interval was extended from 3 sec. to4 sec. The researchers observed a statistically

significant reduction in the percentage of vehiclesentering during the red-signal indication. These resultsshould be viewed cautiously, however, since thetreatment sites only included two intersections andsince there was no indication of comparison or controlsites.

R. A. van der Horst (50) found that increases in theyellow interval decreased the amount of red-signalviolations. He conducted a behavioral before-and-afterstudy at 23 urban and rural intersections in theNetherlands. One year after the yellow intervals werelengthened by 1 sec., the number of red-signalviolations at the intersections lowered by one half.Bonneson’s research indicates (39) that yellowincreases in the range of 0.5 to 1.5 sec., that do not yielddurations above 5.5 sec. can potentially reduce red-lightrunning by about 50 percent.

Drawbacks of Lengthened Yellow Intervals

Although lengthening the yellow interval may reducesignal violations, an interval that is too long coulddecrease the capacity of the intersection and increasethe delay to motorists and pedestrians. Present thoughtis that longer intervals will cause drivers to enter theintersection later and it will breed disrespect for thetraffic signal. The tendency for motorists to adjust to thelonger interval and enter the intersection later isreferred to as habituation.

The before-and-after study by Retting and Greene (47)evaluated the presence of habituation to the longeryellow interval by using a second after-period. Thesame after-period measurements (cycles with signalviolation and late exits) were collected in a secondafter-period approximately six months after the firstafter-period. They were compared to the first after-period. The authors concluded that habituation to thelonger yellow did exist although it may have been onlylargely present at one site for signal violations. Nosignificant habituation was observed for late exits. Inthe before-and-after study at the two intersections inArizona, Wortman et al. (49) compared plots of the timeof entry of vehicles into the intersection. Theresearchers observed an increase in the number ofdrivers entering towards the end of the interval,possibly due to the lengthened yellow interval.

34 34


Additional research is needed to further understand theeffect of lengthening the yellow interval on driverbehavior.

The goal of traffic engineers has been to find anoptimum interval for the yellow change and all-red (ifused) while recognizing that there are traffic androadway variables that must be considered. The Manualof Traffic Signal Design (45) cautions that changeintervals greater than 6 sec. should be examinedcritically before being implemented. They cite loss inefficiency and capacity at the intersection and atendency for local drivers to use more of the changeinterval when they know that it is longer than normal.

All-Red Clearance Interval

An all-red interval is that portion of a traffic signalcycle where all approaches have a red-signal display. Ifused, the all-red interval follows the yellow-changeinterval and precedes the next conflicting greeninterval. The purpose of the all-red interval is to allowtime for vehicles that entered the intersection during theyellow-change interval to clear the intersection beforethe traffic-signal display for the conflicting approachesturns to green.

In many states, it is legal to enter the intersection duringany portion of the yellow interval. Hence, if a vehicleenters the intersection at the end of the yellow intervaland if an all-red interval is not provided, the vehicle willbe in the intersection while a conflicting approachreceives the green signal. Hence, there is a potential fora crash, even when no one entered the intersectionillegally.

It should be pointed out that providing an all-redinterval (or the length of the all-red interval) does NOTaffect the decision or the act of the motorist in runningthe red light. Because use of the red-clearance intervalhas been shown to increase the safety of an intersection,it is mentioned as a countermeasure in this toolboxbecause it can impact the safety of a signalizedintersection.

As stipulated in the MUTCD, the all-red clearanceinterval is optional. The duration of the all-red interval

shall be predetermined, which means the length of theinterval should be calculated based on knownintersection conditions and the length of time of theinterval should not vary on a cycle-by-cycle basis. TheMUTCD also stipulates that the duration of the all-redinterval should not exceed 6 sec. There are noguidelines in the MUTCD for when the all-red intervalshould or may be used. For most agencies, the decisionto use an all-red interval is tied to the determination ofthe yellow-change interval. In the latest version ofITE’s Traffic Engineering Handbook (44) it issuggested that when the calculated change interval isgreater than 5 sec., an all-red interval provides theadditional time beyond 5 sec. Many agencies allocatethe third term of Equation [1], shown previously, as theall-red interval.

While the use of an all-red interval is optional, surveyresults support that most jurisdictions use it at amajority of their intersections. In response to a surveyconducted by The Urban Transportation Monitor (51),of the 76 city traffic engineers that responded,approximately 80 percent indicated that they use all-redat all signals and 20 percent indicated that they use it atsome signals. The survey results indicate that only 35percent apply the same standard interval length forapplications. Those standard intervals ranged from 0.5to 2 sec.

Studies have shown that providing a red interval doeshave a positive effect on the safety of the intersection.Four studies, summarized in Chapter 5 of the Synthesisof Safety Research Related to Traffic Control andRoadway Elements—Volume 1 (52), investigated theeffect of adding an all-red interval on intersectioncrashes. All studies were performed in the 1970s invarious states and cities and all of the study resultsindicated more than a 40 percent reduction in right-angle accidents at the study locations. These results,however, should be viewed cautiously because thestudy summaries did not indicate that measures weretaken to control for trends and regression to the meanbias.

A positive safety benefit was reported in a more recentstudy by Datta et al. (53). Several improvements weremade to three intersections in Detroit including the useof an all-red interval (1.5 to 2.0 seconds). Based on the

3535


results of a statistical analysis of a multi-year before-and-after study of crashes, they concluded that therewas a significant reduction in right-angle crashes andinjuries after implementation of all-red intervals. Thisreported reduction in crashes could not be attributedsolely to the all-red interval since other improvementswere made. However, it does support that reductions incrashes can be realized from a combination ofimprovements identified in this toolbox including theuse of an all-red interval.

The drawback to using an all-red is that it takes awayfrom the green available for other movements andhence reduces the capacity of the intersection. Theamount of reduction is dependent upon the number ofphases and cycle length. For example, for a simple two-phase signal with a 120-sec. cycle length timing plan,the reduction in capacity is only 2.5 percent (comparedto the same signal without an all-red) from the additionof a 1.5 sec. all-red after each phase. The reduction issmall, however in congested corridors, its use wouldexacerbate delays.

Dilemma-Zone Protection

The “dilemma zone” has been defined recently to be thearea in which it may be difficult for a driver to decidewhether to stop or proceed through an intersection atthe onset of the yellow-signal indication (54). It is alsoreferred to as the “option zone” or the “zone ofindecision” (55). One potential countermeasure toreduce red-light running is to reduce the likelihood thata vehicle will be in the dilemma zone at the onset of theyellow interval. This can be accomplished by placingvehicle detectors at the dilemma zone. They detect if acar is at the dilemma zone immediately before the onsetof the yellow interval. If a vehicle is there, the greeninterval can be extended so that the vehicle can travelthrough the dilemma zone and prevent the onset of theyellow while in the dilemma zone. When combinedwith a speed detector, the countermeasure is even morebeneficial. This countermeasure is referred to asdilemma-zone protection or green extension systems.

Zegeer and Deen conducted a before-and-afterevaluation of green extension systems at threeintersections in Kentucky to determine their effect on

crashes (56). The duration of the before-period was 8.5years and the duration of the after-period was 3.7 years.There were 70 accidents in the before-period and 14accidents in the after-period. The authors found a 54percent reduction in accidents per year at the three sitescombined. No comparison or control groups were used.McCoy and Pesti conducted an evaluation of dilemmazone protection in Nebraska (55) as part of anevaluation of active advance-warning signs (discussedin a previous section). Dilemma-zone protection usingconventional detectors was compared to dilemma-zoneprotection with active advance-warning signs in across-sectional evaluation. Overall, the two methodsperformed similarly when red-signal violations werethe measure of effectiveness.

ELIMINATE NEED TO STOP

The final group of solutions to red-light running asdescribed in the introduction of this chapter involveseliminating the need to stop. This can be done byremoving the signal or redesigning the traditionalintersection. An intersection should be designedfollowing standards and guidelines found inAASHTO’s A Policy on Geometric Design of Highwaysand Streets (46). Other design guidelines can be foundin two ITE publications: The Traffic Safety Toolbox: APrimer on Traffic Safety (54) and Traffic EngineeringHandbook (44).

Unwarranted Signals

If there is a high incidence of red-light runningviolations, this may be because the traffic signal isperceived as not being necessary and does notcommand the respect of the motoring public. Thedecision to install a traffic signal is based on the trafficvolume of the intersecting streets, pedestrian traffic and

36 36


Eliminate Need to Stop Through:

�� Unwarranted signal removal;

�� Roundabout intersection design; and

�� Flash mode.

the flow of traffic through a network. Warrants areprovided in the MUTCD that define the minimumconditions at which installing a traffic signal may bejustified. However, sometimes signals are installed forreasons that dissipate over time. For instance, trafficvolume may decrease due to changing land-use patternsor the creation of alternative routes.

There have been studies of the impact of removingtraffic signals and converting the intersection to STOPsign control. Kay et al. (57) found that at 26intersections converted to multi-way stop control, therewas a decrease in the average annual accidentfrequency of more than one accident per year. Wheresignals were replaced by two-way stop control, theyfound that on average there was an increase in right-angle crashes, but it was offset in the number ofcollisions and injuries by a reduction in rear-endcrashes.

One of the primary factors that caused an increase inoverall crashes was the presence of inadequate corner-sight distance. They also found that one-wayintersections with low volumes experience an overallcrash reduction. This was the same finding in a study of199 low-volume Philadelphia intersections, where itwas determined that traffic-signal removal resulted in a24 percent crash reduction (58).

Kay et al. (57) concluded that replacing unjustifiedsignals with two-way stop control has the followingbeneficial impacts:

�� Total delay per vehicle is reduced byapproximately 10 sec.;

�� Idling delay per vehicle is reduced byapproximately 5 to 6 sec.;

�� Stops are reduced from approximately 50 percentof total intersection traffic to about 20 to 25 percentor even less if side road volumes are low in relationto total intersection volume; and

�� Excess fuel consumption due to intersection stopsand delays is reduced by approximately 0.002gallons per vehicle.

The removal of a traffic signal should be based on anengineering study. Factors to be considered in such a

study are enumerated in Figure 3–22 from ITE’s TrafficControl Devices Handbook (59). Specific guidance onsignal removal can be found in Kay et al.’s report (57).Once it is established that a signal can be removed,Section 4B.02 of the MUTCD suggests a five-stepprocess for removal of the signal.

Figure 3–22. Factors to Consider in Signal RemovalSource: Traffic Control Devices Handbook (59).

3737


�� Intersection operations that merit signal removal

�� Projected traffic volumes relative to signalwarrants

�� Total stopped time delay with and without thesignal

�� Projected collision problems with and without thesignal

�� Treatment of any sight-distance problems

�� Impact on signal system progression

�� Savings in operational costs with removal suchas electrical consumption, maintenance costsand repair parts

�� Improvements in environmental issues such asair quality, fuel standards and noise pollution

�� Analysis and recommendation of intersectiontraffic control to replace signal control

�� Recent geometric and/or sight restrictionimprovements in the vicinity of the intersection

Roundabout Intersection Design

A “modern” roundabout can be described as acirculatory intersection that features channelizedapproaches, yields control for vehicles entering thecircle and geometric curvatures that promote lowerspeeds within the roundabout. Other features include acentral island off-limits to pedestrians and raised“splitter” islands on approach legs that divide enteringand exiting traffic, as well as provide a refuge forpedestrians. Figure 3–23 shows a roundabout inColorado.

Figure 3–23. Roundabout.

With respect to the topic of this report, the roundaboutreplaces the traffic signal and obviously eliminates thered-light running problem. Assuming the roundabout isoperationally more efficient (which may not be the casefor many intersection conditions), the issue is whetheror not it is a safer intersection considering all crashes.

Currently there are no recommended criteria orguidelines for when a roundabout should be considered.Roundabouts are acknowledged in the latest edition ofAASHTO’s A Policy on Geometric Design of Highwaysand Streets (46), with sparse design criteria, and there isguidance for appropriate pavement marking andsignage in the MUTCD (see Section 3B.24). However,the most comprehensive guidance from planning todesign is found in a FHWA document calledRoundabouts: An Informational Guide (60).

Although use of roundabouts is limited in the UnitedStates, they are commonly used internationally. Many

international studies have found that roundaboutsgreatly reduce the number of accidents and severity ofaccidents at converted intersections. Other benefits ofroundabouts include: reduction in vehicle emissions,noise, fuel consumption and traffic delays, as well aseliminating the need for maintenance and electricalcosts of operating a signalized intersection. The centerisland also provides a good location for landscaping andarchitectural treatments for improving the aestheticquality of the intersection. Particular sites appropriatefor roundabouts include locations with heavy delay onthe minor road, an intersection with heavy left-turningtraffic, an intersection with more than four legs orunusual geometry and intersections where U-turns aredesirable (61).

There are certain drawbacks associated withroundabouts. The biggest drawback is to pedestrians, asthey are limited to cross only on the approach legs andhave no exclusive right of way (i.e. no pedestrianphase). Roundabouts are not appropriate everywhere asthey do require a fair amount of right-of-way (outerdiameter of approximately 100 ft.) and have a limitedvehicle capacity. Public support of roundabouts in theUnited States is also a concern.

Conversion of a signalized intersection to a roundabouteliminates red-light running because the signal has beenremoved. The real test of safety effectiveness is in termsof total crashes. A conversion from traffic-signal controlto roundabouts reduces the total number of injurycrashes by 30 to 40 percent (62). Another study statesthat left-turn accidents are eliminated and angleaccidents are reduced by 80 percent (61). Suchreductions can be attributed to the reduction in thenumber of conflict points and induced slower speedsthrough the intersection.

Flashing Mode

During periods of low traffic volumes no one shouldhave to wait needlessly at a traffic signal. Today’smodern traffic-signal control technologies are fullytraffic actuated/adaptive systems that incorporateadvanced loop or video detection methods. If workingproperly, even minor street traffic may not necessarilyhave to face a stopped condition. Yet, there are locations

38 38


that are not instrumented to take advantage of theseadvanced technologies. Therefore, during low volumeconditions at an intersection, it may be appropriate forsignalized traffic-control devices to operate on flashingmode. The Traffic Control Devices Handbook (59) liststhe following benefits associated with flashing modeoperation:

�� Reduce stops and delay to major-street traffic;

�� Reduce delay to cross-street traffic;

�� Reduction in fuel consumption due to reduceddelay; and

�� Reduction in electrical consumption by the trafficcontrol signal equipment.

By using the flashing mode, the need to stop (and/orwait) at an intersection is greatly reduced.

When in flashing mode, the MUTCD recommends aflashing yellow on the major street approaches and aflashing red on the minor street approaches. A lesscommon arrangement (although common in California)is to use flashing red on all approaches to theintersection. The MUTCD also provides furtherrequirements regarding the application and operation oftraffic-control signals including the transition fromsteady to flashing mode and the need for flashing modecapability for emergency situations (see MUTCDSection 4D.11 and 4D.12).

The Traffic Control Devices Handbook cautions that theaccident pattern at the intersection should be monitoredto determine if the flashing mode has caused anincrease in accidents. The indicators mentioned includethe following three points:

�� A short-term rate of 3.0 right-angle accidents inone year during the flashing operation;

�� A long-term rate of 2.0 right-angle accidents permillion vehicles entering during the flashingoperation if the rate is based on three to fiveobserved right-angle accidents; and

�� A long-term rate of 1.6 right-angle accidents permillion vehicles entering during the flashingoperation if the rate is based on six or more right-angle accidents.

These conditions were developed as a result of a FHWAstudy investigating different effects of flashing signaltraffic control on intersection operation and safetyconcluded in 1980. After studying data from across thecountry, the study concludes that flashing yellow/redoperation significantly increases the hazard of night-time driving. The major exception is an intersectionwhere the ratio of major street volume to minor streetvolume is greater than three, or where the major-streettwo-way volume is less than 200 vehicles per hourduring flashing operation (63).

The potential hazardousness of using flashingoperations was recently confirmed by Polanis (64) inWinston-Salem, NC. At 19 intersections where low-volume flashing operations were removed (i.e.,returned to normal signal control), right-angle crashesdeclined at every intersection (of which 16 hadstatistical significance), and for all the intersectionscombined there was a 78 percent reduction. Totalaccidents decreased by 33 percent, but for fourlocations an increase was observed. A follow-up afteranalysis showed that the right-angle crash reductionwas sustained over a longer period. Polanis uses thisfinding to conclude that the use of flashing operationduring low volume periods “…is a strategy to reducedelay that need not be abandoned, but its use requirescareful application and additional monitoring.”

SUMMARY

There are a number of factors or reasons that causedrivers to run red lights. There are also a number ofcountermeasures that can address these factors anddiscourage red-light running. The engineeringcountermeasures discussed in this chapter andsummarized in Figure 3–24 address signalvisibility/conspicuity, increasing the likelihood ofstopping, removing the reasons for intentionalviolations and eliminating the need to stop. Most ofthese actions are low cost countermeasures. However,specific unit costs were not provided here since thesecosts can vary considerably among jurisdictions.

It is very difficult to prioritize countermeasures basedon a relative estimate of cost effectiveness or crashreduction potential for a number of reasons.

3939


Information provided from past studies thatinvestigated the effectiveness of measures is limited.Additionally, the results of such studies are site specific.Moreover, the best countermeasure cannot bedetermined strictly from the effectiveness potential butmust be appropriate for the specific intersection. Forexample, although modifying the yellow interval hasbeen shown to reduce violations, the most appropriatecountermeasure for a section with horizontal alignmentproblems is to provide warning of the upcoming signal.

Selecting the best countermeasure to use depends onindividual site characteristics. The countermeasuremost suited to the specific intersection can only bedetermined after conducting an engineering study thatinvestigates the safety of the intersection as related tored-light running and also the occurrence of red-lightviolations. Additionally, an engineering study willinvestigate the existing design elements of theintersection. After such an investigation, the appropriatecountermeasure for the specific site can be identified.Chapter 4 provides further information on conductingan engineering study.

40 40



�� Placement and number of signalheads

�� Size of signal display

�� Line of sight

Increase Likelihood of Stopping

�� SIGNAL AHEAD signs

�� Advanced-warning flashers

�� Rumble strips�� Left-turn signal sign

�� Pavement surface condition

Eliminate Need to Stop

�� Unwarranted signals

�� Roundabout intersection design

�� Flash mode


�� Redundancy

�� LEDs signal lenses

�� Backplates

�� Strobe lights

Address Intentional Violations

�� Signal optimization

�� Signal-cycle length

�� Yellow-change interval

�� All-Red clearance interval

�� Dilemma-zone protection

Figure 3–24. Summary of Engineering Countermeasures by Category.

4141

Red-Light RunningProblem Identification and

Resolution Process

4chapter

INTRODUCTION

The solution to the red-light running problem involvesa combination of education, enforcement andengineering measures. The focus of this report has beenon engineering countermeasures, which were identifiedin Chapter 3. This chapter presents information on howan agency can identify the existence of a red-lightrunning problem and then select the most appropriatecountermeasure or a combination of countermeasures.

Governmental agencies may first install automated-enforcement systems at red-light running problemlocations without investigating whether or not there areany engineering deficiencies and/or if certainengineering countermeasures can reduce the incidenceof red-light running. Prior to the installation of red-lightrunning cameras, an engineering study should beperformed. This study should determine whether thered-light running problem is a design/operational issue(requiring engineering) or a behavioral issue (requiringenforcement and education). It is first necessary toverify that an intersection has been properly designedand constructed and that there are no engineeringdeficiencies that contribute to the red-light runningviolations.

The intent of this chapter is to provide a process to befollowed to systematically address a red-light runningproblem at a signalized intersection. The process ofinvestigating an intersection, looking for engineeringdeficiencies and implementing engineeringcountermeasures is known as an intersectionengineering study. The goal of an engineering study isto identify the most effective solution to an identifiedproblem. In this case, the problem would be red-lightrunning. The solution could include engineering,education, or enforcement countermeasures.

A distinction should be made between a red-lightrunning “problem” at a specific signalized intersectionor a system-wide problem within a jurisdiction or area.It appears that the incidence of red-light running isincreasing along with other traffic violations,collectively described as aggressive driving behaviors.If true, this increase should be experienced at a majorityof signalized intersections. If that is true, then thereneeds to be a system-wide solution set that wouldconsist of a combination of engineering, education andenforcement measures. The discussion that follows,however, deals with specific intersections.

THE PROCESS

The process for addressing a safety problem related tored-light running is the same as would be for anyidentified safety problem. From an engineeringperspective, it includes the following activities:

1. Confirm that there is a safety problem;

2. Conduct an engineering analysis to identify thefactors that might be causing the problem;

3. Identify alternative countermeasures;

4. Select the most appropriate single or combined setof countermeasures; and

5. Implement the countermeasures and monitorimplementation of the solution to determine theextent of the continuance of the problem.

How these elements can be pursued for a red-lightrunning safety problem is discussed below.

Red-Light Running ProblemIdentification

At any given signalized intersection there is likely to besome amount of red-light violations. There are alsolikely to be some number of crashes related to red-lightrunning, notably angle-type crashes. Logically, the twoare related with increasing violations begettingincreasing crashes; however, the exact relationship hasnot yet been established.

The issue here is whether or not the frequency of one orboth measures, violations and crashes, is high enoughsuch that it signals a red-light running problem—thatbeing a frequency that is higher than what would beexpected. If a specific intersection has a red-lightrunning problem, then how should the engineer, inconcert with law enforcement, address the problemuntil it is sufficiently mitigated?

The initial identification of a red-light running problemcan come from several sources, singularly or incombination as illustrated in Figure 4–1. Citizens,either as drivers, pedestrians, or bicyclists, cancomplain about a specific intersection having too many

motorists running the red light. These complaints can bedirected to either the local engineering office, thepolice, or to elected officials. Police can become awareof problem intersection either through citizencomplaints, their own patrolling and monitoring, orfrom high accident location identification programs oftheir own or the engineering department. Theengineering office may become aware of a potentialproblem in a similar fashion as the police. Quite often,elected officials may be most vocal about a red-lightrunning problem that needs to be corrected.

Figure 4–1. Red-Light Running Problem Identification.

However, to determine if there is indeed a red-lightrunning problem, a traffic engineering analysis shouldbe performed. A specific signalized intersection couldbe considered a red-light running problem if it isexperiencing a level of violations or related crashes thatis greater than some selected threshold value. Thresholdvalue criteria, such as higher than the average or someother statistic, for violations and/or related crashesshould be established and applied to quantify that thereis a red-light running problem. For violations, a valuecould be based on local police experience. Lawenforcement agencies would have data on citationsissued for various traffic violations and may be able toestablish if a given intersection has a higher thanaverage violation rate. Based on the literature, violationrates can vary significantly and are likely dependentupon a number of factors. Some violation rates, in termsof violations per approach volume or per time periodwere presented in Chapter 2.

For crashes, the investigating agency should isolate red-light running related crashes. The ability to identify acrash that was a result of running a red light is

42 42


dependent upon the type and accuracy of theinformation recorded on the police report. Indicators ofa red-light running related crash can be found in severalsections of the report, depending upon the state, andinclude: contributing cause (e.g. failed to yield right ofway, disregarded traffic signal), collision type (e.g.angle, left, or right turn), traffic control (i.e. presence oftraffic signal), offense charged and the narrative anddiagram. However, all of these data elements may notbe coded or available to the analyst who is using onlythe coded file to identify red-light running crashes. Ifthe analyst does not have access to the police report,angle-type crashes that are coded as occurring at anintersection, with a traffic signal and a driver action thatwould indicate disregard of the signal, would likely bea red-light running related crash.

To be a problem, red-light running related crashes couldbe either high in frequency, high in rate based onintersection entering volume, or high in comparison toother types of crashes related to the intersection.Bonneson (39) indicates that typical intersectionapproaches experience 3 to 5 red-light runners per1,000 vehicles. Using a rate statistic for this assessmentis preferred, but it requires having timely traffic volumedata that may not be readily available. An alternativemethod would be to compare the percentage of red-lightrunning related crashes to other crash types at theintersection. This comparison would be made against anintersection crash type distribution developed by therespective agency, where the data distribution is morerepresentative of local intersection characteristics. If thepercentage of angle crashes was much higher than thevalue for the distribution found for all similarintersections in the jurisdiction, than this could indicatea red-light running problem.

The data should be evaluated to determine if a red-lightrunning problem exists. If not, then attention can beturned to other problems that might exist at theintersection and countermeasures to address those.

Site Evaluation to Identify Deficienciesand Engineering Countermeasures

If there is a confirmed problem, then the engineershould identify the factors that are contributing to theproblem and then evaluate possible countermeasures ina systematic method. The initial step for this evaluationis to conduct a field review and collect the necessarydata that would isolate any deficiencies. As a minimum,the data and assessments that will be needed include:

�� Traffic volumes as turning movement counts (with consideration to truck volumes);

�� Signal timing parameters;

�� Sight distance to the signal;

�� Geometric configuration;

�� Traffic signs and markings and their condition;

�� Pavement condition; and

�� Traffic speed.

The engineer can refer to the ITE publication, Manualof Transportation Engineering Studies (39) regardingmethods and procedures for collecting various trafficrelated data.

In addition, the engineer should spend some timeobserving the traffic flow and the occurrence of red-light running at the intersection. An hour or two of on-site observation could confirm the existence of red-lightrunning; indicate the principal kind of red-light runningevent (whether or not the event appears to beintentional); and possibly how these contribute tocrashes. A formal traffic conflicts study as described inthe ITE Manual of Transportation Engineering Studiescould be conducted as well. These observations willprovide a sense for the traffic operational characteristicsof the intersection and may offer potential clues toproblem identification and solution.

One of the first considerations is to confirm that thetraffic signal is still warranted. It would be unusual fora signalized intersection to have a red-light runningproblem where the signal is not warranted because oflow traffic volumes. Still, this preliminary assessment,which can be made easily, is suggested. If there is a

4343

Chapter 4: Red-Light Running Problem Identification and Resolution Process

possibility that the signal is not warranted, then furtherstudies are recommended. (See reference 57 forguidance on removing unwarranted traffic signals).However, even if it is decided to investigate thepossibility of signal removal, the engineer shouldproceed with the following field review.

A field review of the problem intersection is necessaryto better understand the characteristics of the problem,to isolate deficiencies and to identify potentialcountermeasures. Sufficient time should be allocated toconduct a thorough review of the intersection. Thismeans that the review may have to occur duringdifferent times of the day to observe operations andconditions under different levels of traffic flow andambient light.

Figure 4–2 provides a listing of items that should bechecked while at the intersection. They are discussedbelow with consideration to the alternativecountermeasures discussed in Chapter 3 to resolveidentified deficiencies.

Check for Signal Visibility

There are several visibility features that should bechecked. The sight distance to the traffic signals shouldbe determined and compared to the minimum sightdistance requirements shown in Table 3–1. Keep inmind that this will require a knowledge or goodestimate of the 85th percentile speed of approachingtraffic. If the minimum sight distance is not available,there are a number of countermeasures discussed inChapter 3 to consider:

44 44


Visibility and Conspicuity Features1. Sight distance to signals2. Number of signals3. Positioning of signals—overhead, post-mounted,

near-side, far-side4. Line of sight for visibility restricted signals (programmable)5. Brightness of signals6. Conspicuity of signals

Signal Control Parameters1. Coordination with adjacent signals2. Timing and cycle length3. Yellow change interval4. All-red clearance interval

Geometric Features1. Grade of approach lanes2. Pavement condition

Traffic Operations Features1. Vehicle approach speed2. Right-turn-on-red3. Pedestrian usage4. Truck usage

Items to Check During Site Review

Figure 4–2. Traffic Signal Field Review Checklist.

�� Signal Ahead Sign—the MUTCD requires that theW3-3 sign be used if minimum sight distance is notavailable;

�� Advance warning flashers;

�� Repositioning of signals; and

�� Supplemental near-side signal.

While the visibility distance may be adequate, theremay be obstructions to full and continuous visibility.Oftentimes, as shown in Figure 4–3, utility wires canlimit the visibility to the signals. Resolution of thisproblem may require repositioning of the signalsvertically or horizontally or, as was the case shown inFigure 3–9, installing a supplemental signal.

Figure 4–3. Example of Signal View Restrictedby Utility Wires.

The next visibility feature would be the positioning ofthe signals to ensure they meet the cone of visionrequirements and the minimum and maximum distancesfrom the stop bar. Resolution of any deficiencies notedhere would include:

�� Placing signals overhead (if not already); and

�� Repositioning signals.

Next, would be to check that there is an adequatenumber of signal heads. While a minimum of two arerequired for the major movement, consideration shouldbe given to providing one for each lane where there arethree or more lanes and that they are centered over thelane.

When programmable signals are used to avoidconfusion, their visibility sight line should be checked.The provided sight line should be as long as possiblewithout conflicting with other signal displays.

The brightness level of the signals should be viewedduring varying ambient light conditions. Standardincandescent bulbs will deteriorate over time and needto be checked on a periodic schedule. The solution willbe to replace the bulbs in a timely fashion or consideruse of LED signals. While these types of signals willeventually fail, they hold their brightness level for amuch longer period.

Check for Signal Conspicuity

While the signal display may meet all the visibilityrequirements noted above, the signals still may not beconspicuous—the ability to stand out amongstcompeting light sources or other information sourcesthat compete for the motorists’ visual and drivingattention. Often times when the environment around theintersection is visually cluttered, the motorist can bedistracted and not see the traffic signal until well intothe dilemma zone. Visual competition can come inmany forms. In suburban, high-density commercialcorridors, there may be many other light sources fromadvertising signs and the like. Large overhead trafficguide signs may draw the attention of unfamiliarmotorists placing more attention to navigation thanintersection control. Whatever the situation, there are anumber of countermeasures designed to draw theattention to the traffic signal. In addition to those notedfor improving signal visibility, these would include:

�� Use of double red signal displays;

�� Use of backplates, and if the problem is moresevere at night, the use of reflective tape; and

�� Use of 12-in. signals if not already being used.

In addition to making the signals as visible andconspicuous as possible, the engineer may determinethat it is necessary to get the motorists’ attention as theyapproach the intersection. This can be accomplished anumber of ways to include amber flashers on aSIGNAL AHEAD sign and advance warning flashers.

4545


On rare occasions, some engineers have installedstrobes in their red lights for the purpose of itsattention-getting values. However, as noted in Chapter3, these are not sanctioned by the MUTCD and,therefore, their use should be restricted to specialcircumstances. Also, rumble strips have been used inrare instances to gain the attention of approachingtraffic and to reduce speed. As noted in Chapter 3, thereare no known evaluations of their effectiveness and,hence, their use should be restricted to specialsituations. They are not suggested for high-speed(greater than 45 mph) facilities.

Check Signal Control Parameters

Having identified any visibility or conspicuity issues,the engineer should review the various traffic signalcontrol parameters. The occurrence of red-lightviolations is affected, in part, by signal controlparameters, specifically the change period interval andthe cycle length and phasing. These are discussedbelow.

Change Interval Review

Probably the most significant factor that affects theincidence of red-light running is the duration of theyellow change interval. Hence, this aspect of signaloperations should be reviewed for adequacy. Theyellow change and all-red clearance interval wasdiscussed in Chapter 3. Using either the agency’sestablished procedures or procedures provided inpublications such as the ITE Traffic EngineeringHandbook (44), the engineer should determine if theyellow change interval is appropriate for the conditions.If the yellow change interval is within the guidelines,then the engineer may want to consider furtherincreasing the interval, but likely no more than 1additional sec. If it is below the guidelines, then theyellow change interval should be increased to that level.Care must be exercised when using a long yellowchange interval, say 5 sec. or greater. Frequent driversmay realize the signal has a long yellow display andtake advantage of it by continuing through theintersection, instead of coming to a stop. As a result, itmay be appropriate to shorten the yellow changeinterval (yet, with respect to guidelines) and add or

increase an all-red clearance interval to discourageinappropriate driver behavior.

Additionally, the all-red clearance interval should bereviewed, assuming one is being used. As discussed inChapter 3, an all-red clearance interval is that portion ofa traffic signal cycle where all approaches have a red-signal display. The purpose of the all-red clearanceinterval is to allow time for vehicles that entered theintersection during the yellow change interval to clearthe intersection before the traffic-signal display for theconflicting approaches turns to green. While the use ofan all-red does not eliminate red-light running, its usecan prevent a crash for the violator entering theintersection just as the signal turns red. Please refer tothe ITE publication, Traffic Engineering Handbook(44), concerning procedures on the application of theall-red clearance interval.

Cycle Length/Phasing Review

The traffic-signal cycle length and phasing should bereviewed. If the intersection operates within acoordinated signal system, then these two componentswould not likely be changed so as to disrupt progressionand overall system efficiency. If the intersection isoperated in isolation, then the engineer may want toconsider reducing the cycle length if it is long (180 sec.or longer), or increasing the length if short (60 to 90sec.). The possible effect of long and short cycle lengthsis discussed in Chapter 3. Unexpected signal phasingsequences may contribute to red-light running and thisshould be examined as well.

Check Geometric Features

There are at least two geometric features that should bereviewed as they may have an effect on red-lightrunning, namely:

�� Approach grade—Grade is a factor in determiningthe yellow clearance and all-red intervals and isparticularly important when present on high-speedfacilities. The braking distance for high-speedvehicles, especially trucks, on a downgrade aresignificantly longer; and

46 46


�� Pavement condition—The condition of thepavement on the approach should be checked atleast visually. Motorists may be reluctant to cometo a “quick” stop if the pavement is unusuallyrough or appears slippery.

Check Traffic Operational Features

As a minimum, the following traffic operationalfeatures should be examined:

�� Vehicle approach speed—Visibility requirementsand clearance intervals are dependent upon vehiclespeed. The preferred measure is the 85th percentilespeed, but this requires a speed survey for anaccurate determination. When this is not practical,the engineer should observe traffic during non-peak conditions to make an approximation;

�� Right turn on red (RTOR)—Red-light runningcrashes have been mostly associated with throughtraffic and left-turning traffic, however violationsof RTOR can be a special form of this problem.The occurrence of conflicts or violations of a no-RTOR signing should be observed while in the field;

�� Pedestrian usage—The engineer should make noteof pedestrian traffic, conflicts with vehicles(especially with red-light violators) and thepresence of pedestrian accommodations.Improvements to pedestrian accommodationsshould be considered if this appears as a problem;

�� Truck usage—The assumed deceleration rate usedin the formula for determining the yellow changeinterval may need to be decreased if there is a largepercentage of trucks in the traffic stream.

Other Countermeasures To Consider

If after all the viable relatively low-cost engineeringcountermeasures described in the first part of thischapter and in Chapter 3 have been tried withoutsuccess in eliminating the problem, then there are avariety of additional measures to consider. Theseinclude more extensive re-engineering measures, orenforcement countermeasures.

Re-Engineering of Intersection

Consideration should be given to a redesign of theintersection, if appropriate. This may involvephysically improving the sight distance, if that was theproblem, by a change in approach curvature and/orgrade profile. This is obviously an expensivecountermeasure and would require an in-depthengineering analysis to support such a decision. Also,the agency may want to consider replacing thesignalized intersection with an alternative intersectiondesign or possibly a roundabout, but again, a morecomprehensive study would be needed, and quite likelythere would have to be other problems beyond the red-light running issue to justify such an expensivetreatment.

Enforcement

Increased enforcement should be considered if theengineering measures do not resolve the problem oruntil the engineering measures can be installed.Traditionally this would include selective enforcementby the police. Some cities have begun to developspecific task forces to address traffic issues andviolations. Some of these special tasks often includetarget enforcement of red-light violations at particularlydangerous locations with a high number of violations.However, this type of measure is usually transitory ineffectiveness, and can itself be hazardous becausepolice have to follow the offender through theintersection exposing them to potential collisions.

To counter this problem, some jurisdictions use a red-light detector and enforcement light, known as a “ratbox” or “red eye” unit. Figure 4–4 shows the placementof a rat box at an intersection. Figure 4–5 shows theenforcement light being used in the City of Richardson,TX. With this device, a light is attached to a pole that isactivated when the red is on. This allows the police toposition themselves on the far side of the intersection,which precludes the need to follow the offender throughthe intersection.

If the jurisdiction already has an automated-enforcement program using cameras, then they shouldconsider adding the problem location into their

4747


program. In so doing, the jurisdiction might want to usethe following examples as a guide.

Maryland State HighwayAdministration Criteria for Installationof Automated Systems

The Maryland State Highway Administration(MDSHA) does not install automated systems, but doesallow local jurisdictions to use them at intersectionswith a state road under certain requirements. MDSHAhas developed a number of principles for the use of red-light camera systems, which are enumerated below:

1. Use of camera system at a specific site must serve ahighway safety purpose;

2. Site must be studied to disclose engineeringdeficiencies and ascertain potential improvements,and deficiencies corrected and improvements madeprior to the red-light camera use;

3. Traditional enforcement proven ineffective orinefficient prior to red-light camera use;

4. Red-light camera system must be proven technology,reliable, properly installed and maintained;

5. Processing of images and issuance of citationaccurate, efficient and fair;

6. Effective and fair adjudication of offenders who go tocourt;

7. Effectiveness is continually evaluated; and

8. Public awareness maintained.

State of North Carolina Policy on theUse of Automated Systems

In developing a recommended policy for the use ofautomated systems for the state of North Carolina,Milazzo et al. (10) recommended an eight-stage processto be followed for systematically solving a red-lightrunning problem. It is enumerated below:

1. Conduct a traffic engineering study to verifyexistence, extent and causes of the problem;

2. If feasible, implement traffic engineeringcountermeasures;

3. Consider implementation of traditional enforcementmeasures, perhaps with “rat boxes”;

4. If engineering countermeasures and/or traditionalenforcement proves to be unsuccessful or unfeasible,then select appropriate red-light camera locations;

5. Choose a financing arrangement to ensure that publicsafety will remain the primary goal;

6. Conduct a detailed, perpetual public information andeducational effort regarding the program;

7. Implement red-light cameras at intersections withhighest potential for crash reduction benefits; and

8. Monitor camera-controlled intersections, and indeedall countermeasures, for progress over time.

Figure 4–4. Placement of Rat Box at an Intersection.

Figure 4–5. Enforcement Light Installation.

48 48


Milazzo et al. also suggest the following types oflocations may be appropriate for camera installationpending the results of an engineering study:

�� Through lane when opposed by a permitted orprotected-permitted left turn. The reason for this isthe absence of an all-red interval between thesemovements;

�� Through lane when conflicting traffic is likely to bemoving at green (either due to progression ormultiple lanes on the conflicting approach);

�� Through lane when conflicting traffic couldattempt to anticipate the green (due to fixed signaltiming or signal heads visible from conflictingapproach);

�� A selection from among high accident locations ispermissible; and

�� Other situations in which engineering deficienciescannot be reasonably implemented, orimplemented in a reasonable amount of time.

SUMMARY

What has been presented in this chapter is a process fordetermining if a red-light running problem exists andwhat types of countermeasures could be implementedin a logical and systematic manner. Individual agenciesmay have already established procedures forconducting audits and review of problem intersectionsthat may accomplish the same objective. The goal ofthis process is to identify the most effective solution toan identified red-light running problem. The solutioncould include engineering, education, or enforcementcountermeasures.

4949


5151

Future Needs5chapter

This report has been prepared to provide a betterunderstanding of the red-light running problem and toprovide information and case studies regarding howvarious engineering measures can be implemented toreduce the extent of red-light running. The solution tothe red-light running problem also requires educationand enforcement measures. An enforcement measurethat has emerged in several jurisdictions throughout theUnited States is the use of automated-camera systems.These automated systems can be a viablecountermeasure to red-light running violations and toresulting crashes. However, jurisdictions now using orcontemplating using automated systems should ensurethat candidate intersections have had engineeringdeficiencies corrected. In many cases, the engineeringmeasures discussed in this report can provide a lastingand acceptable solution to a red-light running problem.

Further improvements in red-light running violationsand crash reductions can be achieved through thefollowing future activities:

�� Research and development;

�� Improved data related to red-light running crashes;

�� Improved guidelines and standards; and

�� Improved procedures and programs.

These activities are discussed below as concludingremarks to this red-light running countermeasurestoolbox.

RESEARCH AND DEVELOPMENT

Research and development is suggested in thefollowing areas:

�� Better understanding of root causes of red-lightrunning. With heightened awareness of theproblem of red-light running, we are starting tobecome more knowledgeable as to why motoristsintentionally or unintentionally run red lights.However, more research is needed to betterunderstand the root causes of why motorists run redlights at traffic signals. Are the causes related to:(1) driver behavioral factors—impatience and/ordisrespect for traffic laws; (2) driver capability andperformance factors such as diminished vision andperception-reaction time; or (3) roadway and trafficcontrol deficiencies such as inadequate signalvisibility and/or improper signal timing? No doubt,all three factors act independently or incombination to cause a motorist to run a red light.Researchers in sociology, human factors and trafficengineering need to combine their expertise toidentify the root causes of a red-light runningproblem. A complete understanding of thisphenomenon will allow the safety community toidentify appropriate countermeasures, whetherthey are engineering, education, or enforcement.

�� Quantification of crash reduction potential ofvarious countermeasures. In Chapter 3, severalengineering countermeasures were presented, andwhere known, their effectiveness in reducing red-light running related crashes was documented.Unfortunately, only limited information is

52 52


available that would provide reliable estimates ofcrash reduction potential for each of the measuressingularly or in combination. As jurisdictions beginto implement red-light running countermeasures,they should conduct evaluations. A clearinghousefor receiving and distributing information onevaluations would help transportation engineers todecide which countermeasures to deploy.

The safety effectiveness of automated-enforcementsystems need to be fully understood and guidelinesfor where they should or should not be used shouldbe developed. As noted in this toolbox, it has beenshown that automated systems can reduceviolations and resulting crashes. However, notenough is known about how different operatingfeatures associated with these systems, such asadvance warning signs, types of intersections, levelof fines, public information programs, etc.,influence the level of effectiveness. Knowledge ofthese relationships will allow for better deploymentguidelines.

�� Development and evaluation of IntelligentTransportation Systems (ITS) technologies. Thecontinuing development of a variety of ITStechnologies, both for the infrastructure and thevehicle, hold promise for providing a safer roadsystem and specifically for the red-light runningproblem. The Federal Highway Administration has

a comprehensive research and developmentprogram for developing intersection collisionavoidance systems. From the roadway side, thisinvolves vehicle detection systems (in-pavementand overhead) and dynamic warning signs, placedroadside or adjacent to signal heads, that willdetermine if a motorist is likely to run a red lightand give warning to the cross street motorist.

There already exists technology, known as “red-light hold” systems that will extend the cross streetred signal momentarily under the same conditions;improvements can be expected soon. ITS systemsare currently being developed that can predictwhen a vehicle will violate a signal and thenprovide a warning to that vehicle. An infrastructure-based warning system is illustrated in Figure 5–1.

It is anticipated that the next generation of collisionavoidance systems will include in-vehicle warningsystems to accompany infrastructure detectionsystems. The system illustrated in Figure 5–1 couldbe modified to a cooperative system such that theinfrastructure would detect that the vehicle was indanger of violating the signal and the vehiclewould provide the warning to the driver.Eventually, vehicles will have the technology toprovide vehicle-to-vehicle dynamics and providewarnings of possible intersection collisions.

Figure 5–1. Infrastructure-Based ITS System that Warns Potential Red-Light Violators Approaching Intersection.Source: Presentation by Robert Ferlis to the AASHTO Standing Committee on Highway Traffic Safety.

IMPROVED CRASH DATA FORRED-LIGHT RUNNING

In Chapter 2, statistics on the frequency andcharacteristics of red-light running crashes wereprovided. However, these statistics must be viewed asestimates, albeit reasonable, because existing crashdatabases do not allow accurate identification ofcrashes attributed to red-light running. The data comefrom the police crash report and police are sometimesreluctant to cite the motorist for running a red light,especially if they cannot determine for certain who isthe offending party. A review of the narrative ordiagram requires confirmation that the crash didinvolve a red-light violator. While a change in the crashreporting form to deal with this issue is desirable, atleast more caution in entering the data into electronicdatabases is needed. Also, agencies need an efficientdata retrieval system that will allow the continuousmonitoring suggested in Chapter 4.

IMPROVED GUIDELINES AND/ORSTANDARDS

Hopefully, this informational report has given thoseresponsible for operation of traffic signals guidance onhow to identify a red-light running problem and whatcountermeasures, especially engineering related, couldbe used to mitigate the problem. As best that could bedone based on available information, guidelines areprovided where a specific measure is most appropriate.However, better guidance on what measure is mostappropriate for a given situation is needed. Thisguidance can follow from the research and developmentprogram noted above and from the experiences gainedby the traffic engineering community.

The MUTCD provides standards and guidance relatedto traffic-signal design and operations and theassociated traffic signs and markings, which draw fromresearch and field experience. Adherence to thesestandards and guidance provides for uniform andconsistent application of traffic-control devices. Whilethis is generally true, there are significant variations inpractices across the country, which can lead to motoristconfusion and misunderstanding that might be reflected

in traffic violations. Consistency in the design—number, placement and configuration of the signals—and the operation—signal phasing, clearance intervals,etc.—would be beneficial to citizens that frequentlydrive in many states. While the unique requirements ofa specific location will always need to be considered bythe engineer, the focus of traffic-signal design andoperation should be to deliver consistency (anduniformity) to the motorist in terms of head placement(signal visibility) and operation such as the length ofyellow change and all-red clearance intervals.

IMPROVED PROCEDURES ANDPROGRAMS

The solution to the red-light running problem requires acomprehensive and coordinated program that involvesthose stakeholders responsible for providing a drivingenvironment that is as safe as possible. From the start,driver-licensing agencies should ensure that newdrivers understand basic rules of the road and themeanings and operations of traffic control devices. Thisis normally accomplished through driver manuals anddriver testing for licensing. Education officials have arole in ensuring this information is acquired throughdriver training. In the case of red-light running,education continues for experienced drivers throughpublic information and awareness campaigns thathighlight the problem and its consequences. Educationand public information programs are especiallyrequired when automated enforcement is utilized.Automated-enforcement programs are better acceptedby the community and are more effective, if the publicunderstands why they are being used, that othermeasures have been used and have not solved theproblem, and that the program is carried out fairly.

Enforcement officials have the responsibility ofassuring that road users adhere to traffic laws and takecorrective action when they do not. Coordinationbetween the enforcement and engineering community isneeded to identify where there are high incidences ofviolations that are resulting in crashes. Enforcementofficials should have a basic understanding of traffic-control devices and recognize where there areengineering deficiencies that may contribute to theviolations.

5353

Chapter 5: Future Needs

The public works and engineering professionalsresponsible for the streets need to be aware of acceptedstandards and guidance that relate to the design andoperation of traffic signals. They have a responsibilityfor monitoring the crash experience of their streetsystem so that they can identify when a problem isemerging.

The stakeholders representing engineering, educationand enforcement need to work together, developingprograms and procedures that would allow for theseactions to be carried out efficiently and effectively.Sometimes, this needed alliance can be achievedthrough partnerships formed between public agenciesand private entities. Frequently, added funding can alsobe obtained when private and quasi-public entitiesinterested in safety participate. An example of this canbe seen in Michigan where an alliance of several groupswas forged through the efforts of AAA Michigan.

Red-light running continues to be a significant nationalsafety problem. The occurrence of red-light running andmoreover the crashes that result from red-light runningcan be reduced at intersections through education,enforcement and engineering. This report providesinformation to engineers, law enforcement officials,elected and appointed officials and the general public tohelp accomplish the goal to reduce red-light runningcrashes.

54 54


5555

references1. Information provided by the Federal Highway Administration in September 2001.

2. H. W. McGee and K. A. Eccles, NCHRP Synthesis of Highway Practice 310: Impact of Red Light CameraEnforcement on Crash Experience, (Washington, DC: Transportation Research Board, National ResearchCouncil, 2002).

3. National Committee on Uniform Traffic Laws and Ordinances, Uniform Vehicle Code and Model TrafficOrdinance (Evanston, IL: Revised, 1992).

4. P. A. Parsonson, NCHRP Synthesis of Highway Practice 172: Signal Timing Improvement Practices,(Washington, DC: Transportation Research Board, National Research Council, 1992).

5. L.M. Wissinger, “Issues Surrounding the Operation and Installation of Red Light Cameras” (master’s thesis,North Carolina State University, 2000).

6. R.A. Retting, A.F. Williams and M.A. Greene, “Red-Light Running and Sensible Countermeasures: Summaryof Research Findings,” Transportation Research Record 1640 (1998): 23-26.

7. A. Kamyab et al., “Red Light Running in Iowa: The Scope, Impact and Possible Implications,” (final report,Center for Transportation Research and Education, Ames, IA, 2000).

8. Y.M. Mohamedshah, L.W. Chen and F.M. Council, “Association of Selected Intersection Factors with Red-Light Running Crashes,” (Highway Safety Information System Summary Report, Federal HighwayAdministration, 2000).

9. R.A. Retting, R. G. Ulmer and A.F. Williams, “Prevalence and Characteristics of Red Light Running Crashesin the United States,” Accident Analysis & Prevention (1999): 31, 687-694.

10. J.S. Milazzo, J.E. Hummer and L.M. Prothe, “A Recommended Policy of Automated Electronic TrafficEnforcement of Red-Light Running Violations in North Carolina,” (North Carolina Governor’s HighwaySafety Program, North Carolina State University, 2001).

11. “Stop Red Light Running,” Federal Highway Administration Safety Web site:safety.fhwa.dot.gov/programs/srlr.htm, 2002.

12. H. Thompson and L. Adlard, “Photo Red Light Enforcement Demonstration Project Evaluation: City ofBeaverton,” (City of Beaverton, OR, 2001).

56 56


13. B. Persaud and F. Council, “Red Light Camera Safety Evaluation—Experimental Design,” (draft final report:FHWA Intelligent Transportation Systems Joint Program Office under Project Number DTFH61-96-C-00077,2002).

14. J. L. Fleck and B.B. Smith, “Can We Make Red Light Runners Stop? Red Light Photo Enforcement in SanFrancisco, California,” (paper presented at the 78th annual meeting of the Transportation Research Board,Washington, DC, January 1999).

15. Unpublished report documenting the methodology and overview of findings for focus groups conducted in SanFrancisco on June 25, 2001 conducted by Jeff Henne, President of The Henne Group.

16. R. A. Retting and A.F. Williams, “Characteristics of Red Light Violators: Results of a Field Investigation,”Journal of Safety Research (1996): 27.1, 9-15.

17. B. E. Porter et al., “A Nationwide Survey of Red Light Running: Measuring Driver Behaviors for the ‘StopRed Light Running’ Program,” 1999. Available: www.fhwa.dot.gov/safety.

18. J. Bonneson, M. Brewer and K. Zimmerman, “Review and Evaluation of Factors that Affect the Frequency ofRed Light-Running,” FHWA/TX-02/4027-1 (Washington, DC: Federal Highway Administration, 2001), 78.

19. Institute of Transportation Engineers, Traffic Engineering Handbook, (Washington, DC: Federal HighwayAdministration, 1999).

20. R. Van der Horst and A. Wilmink, “Drivers’ Decision-Making at Signalized Intersections: An Optimization ofthe Yellow Timing,” Traffic Engineering & Control, (Crowthorne, England: 1986) 615-622.

21. Results of unpublished research conducted by BMI for the Federal Highway Administration.

22. Federal Highway Administration, Manual on Uniform Traffic Control Devices for Streets and Highways,Millenium Edition, (Washington, DC: Federal Highway Administration, 2001).

23. G. F. King, “Guidelines for Uniformity in Traffic Control Signal Design Configurations,” Research ResultsDigest, (Washington, DC: National Cooperative Highway Research Program, Transportation Research Board,National Research Council, 1977).

24. G. B. Thomas and D. J. Smith, “Effective of Roadway Safety Improvements,” (final report: Center forTransportation Research and Education, Iowa State University, 2001).

25. B. P. Russi, “Impact of Mast-Mounted Signal Heads on Accident Reduction,” ITE Journal (October 1991).

26. Based on unpublished information provided by Stan Polanis of Winston-Salem, North Carolina. Some of thisinformation was updated and subsequently compiled in a paper entitled, “Improving Intersections SafetyThrough Design and Operations (Examples),” by Stanley F. Polanis, Assistant Director of Transportation, Cityof Winston-Salem, Department of Transportation.

27. E. Felipe, D. Mitic, and S. R. Zein, “Safety Benefits of Additional Primary Signal Heads,” (British Columbia:Insurance Corporation of British Columbia, 1998).

5757

28. Unpublished Information provided by Naperville, IL.

29. Unpublished information obtained from John Abraham, Engineer, City of Troy, MI.

30. T. Sayed, W. Abdelwahab and J. Nepomuceno, “Safety Evaluation of Alternative Signal Head Design”,Transportation Research Record 1635, (Washington, DC: Transportation Research Board, 1998).

31. B. Marshall, “How Stuff Works—Question 178,” howstuffworks.com, May 2002.

32. Institute of Transportation Engineers, Equipment and Materials Standards, (Washington, DC: Institute ofTransportation Engineers, 2001).

33. W. J. Moore and C. You, “Wire Guided,” Traffic Technology International, 2001.

34. J.D. Bullough, et al., “Response to Simulated Traffic Signals Using Light-Emitting Diode and IncandescentSources,” Transportation Research Record 1724, (Washington, DC: Transportation Research Board, 2000).

35. “Light Emitting Diodes Being Used for Traffic Signal Lights,” www.co.scottsdale.az.us/trafficsignals/ledpur,City of Scottsdale, AZ.

36. “Visibility of Traffic Signal Displays for Aging and Colour Deficient Drivers—Final Report,” (final report:National Committee on Uniform Traffic Control, September 2001).

37. W.J. Styles, “Evaluation of the Flashing Red Strobe Signal,” (report FHWA/MD-83/07: Maryland StateHighway Administration, June 1982).

38. B.H. Cottrell, Jr., “Evaluation of the Use of Strobe Lights in the Red Lens of Traffic Signals,” TechnicalAssistance Report.

39. J. Bonneson, M. Brewer and K. Zimmerman, “Engineering Countermeasures to Reduce Red Light Running,”FHWA-TX-03/ 4027-2, (Washington, DC: Federal Highway Administration, 2002) 122.

40. B.A. Farraher, R. Weinholzer and M.P. Kowski, “The Effect of Advanced Warning Flashers on Red LightRunning—A Study Using Motion Imaging Recording System Technology at Trunk Highway 169 and PioneerTrail in Bloomington, Minnesota,” www.dot.state.mn.us/metro/trafficeng/signals/reports.html.

41. T. Sayed, H. Vahidi, and F. Rodriguez, “Advanced Warning Flashers: Do They Improve Safety?”Transportation Research Record 1692, (Washington, DC: Transportation Research Board, 1999).

42. National Center for Statistics and Analysis, “Traffic Safety Facts 2000: A Compilation of Motor Vehicle CrashData from the Fatality Analysis Reporting System and the General Estimates System,” DOT HS 809 337,(NHTSA, National Center for Statistics and Analysis, U.S. Department of Transportation: December 2001).

43. G.A. Ali, R. Al-Mahrooqi, and R. Taha, “Measurement, Analysis, Evaluation and Restoration of SkidResistance on the Streets of Muscat,” (Washington, DC: Transportation Research Board, 1999).

44. Institute of Transportation Engineers, Traffic Engineering Handbook, (Washington, DC: Institute ofTransportation Engineers, 1999).

References

58 58


45. Institute of Transportation Engineers, Manual of Traffic Signal Design, (Washington, DC: Institute ofTransportation Engineers, 1998).

46. American Association of State Highway and Transportation Officials, A Policy on Geometric Design ofHighways and Streets, (Washington, DC: American Association of State Highway and Transportation Officials,1990).

47. R.A. Retting and M.A. Greene, “Influence of Traffic Signal Timing on Red Light Running and PotentialVehicle Conflicts at Urban Intersections,” Transportation Research Record 1595, (Washington, DC:Transportation Research Board, 1997) 1-7.

48. Institute for Transportation Engineers, “Determining Vehicle Change Intervals: A Proposed RecommendedPractice,” (Washington, DC: Institute of Transportation Engineers, 1985).

49. R.H. Wortman, J.M. Witkowski, and T.C. Fox, “Traffic Characteristics During Signal Change Intervals,”Transportation Research Record 1027, (Washington, DC: Transportation Research Board, 1985) 4-6.

50. R.A. van der Horst, “Driver Decision Making at Traffic Signals,” Transportation Research Record 1172,(Washington, DC: Transportation Research Board, 1988) 93-97.

51. The Urban Transportation Monitor, Vol. 15, No. 22 (Lawley Publications, Nov. 23, 2001) 9-11.

52. “Synthesis of Safety Research Related to Traffic Control and Roadway Elements—Volume 1,” FHWA-TS-82-232, (Washington, DC: Federal Highway Administration, 1982) 193.

53. T.K. Datta, K. Schattler and S. Datta, “Red Light Violations and Crashes at Urban Intersections,”Transportation Research Record 1734, (Washington, DC: Transportation Research Board, 2000).

54. Institute of Transportation Engineers, The Traffic Safety Toolbox: A Primer on Traffic Safety, (Washington, DC:Institute of Transportation Engineers, 1999).

55. P.T. McCoy and G. Pesti, “Dilemma Zone Protection with Advance Detection and Active Warning Signs,”(paper presented at the annual meeting of the Institute of Transportation Engineers, August 2002).

56. C.V. Zegeer and R.C. Deen, “Green-Extension Systems at High-Speed Intersections,” ITE Journal,(Washington, DC: Institute of Transportation Engineers, Nov. 1978).

57. J.L. Kay, L.G. Neudorff and F.A. Wagner, “Criteria for Removing Traffic Signals,” Report Number FHWA/RD-80/104, (Washington, DC: Department of Transportation, Sept. 1980).

58. B. Persaud et al., “Crash Reductions Related to Traffic Signal Removal in Philadelphia,” Accident Analysis &Prevention, Vol. 29, 1997.

59. Institute of Transportation Engineers, Traffic Control Devices Handbook, (Washington, DC: Institute ofTransportation Engineers, 2001).

60. Robinson, et al., “Roundabouts: An Informational Guide,” Report No. FHWA-RD-00-067, (Washington, DC:Federal Highway Administration, June 2000).

5959

61. M.E. Niederhauser, B.A. Collins and E.J. Myers, “The Use of Roundabouts: Comparison with AlternateDesign Solution.”

62. B.N. Persaud et al., “Crash Reductions Following Installation of Roundabouts in the United States,”(Arlington, VA: Insurance Institute for Highway Safety, March 2000).

63. K.C. Kacir, R.J. Benz and H.G. Hawkins, “Analysis of Flashing Signal Operation,” Transportation ResearchRecord 1421, (Washington, DC: Transportation Research Board, 1993) 21-29.

64. S.F. Polanis, “Right-Angle Crashes and Late-Night/Early-Morning Flashing Operation: 19 Case Studies,” ITEJournal, (Washington, DC: Institute of Transportation Engineers, April 2002).

65. Institute of Transportation Engineers, Manual of Traffic Engineering Studies, (Washington, DC: Institute ofTransportation Engineers, 1994).

References

1099 14th Street, NW, Suite 300 WestWashington, DC 20005-3438 USATelephone: +1 202-289-0222Fax: +1 202-289-7722ITE on the Web: www.ite.org

ISBN 0-935403-76-0

Making Intersections Safer: A Toolbox of Engineering ... · MAKING INTERSECTIONS SAFER: A TOOLBOX OF ENGINEERING COUNTERMEASURES TO REDUCE RED-LIGHT RUNNING Improve Signal Visibility

Documents