NCHRP Report 493 – Evaluation of Traffic Signal Displays for Protected/Permissive Left-Turn Control

Evaluation of Traffic Signal

Displays for Protected/

Permissive Left-Turn Control

NATIONAL

COOPERATIVE

HIGHWAY

RESEARCH

PROGRAM

NCHRP

REPORT 493

TRANSPORTATION RESEARCH BOARD EXECUTIVE COMMITTEE 2003 (Membership as of March 2003)

OFFICERS

Chair: Genevieve Giuliano, Director and Professor, School of Policy, Planning, and Development, University of Southern California,

Los Angeles

Vice Chair: Michael S. Townes, Executive Director, Transportation District Commission of Hampton Roads, Hampton, VA

Executive Director: Robert E. Skinner, Jr., Transportation Research Board

MEMBERS

MICHAEL W. BEHRENS, Executive Director, Texas DOT

JOSEPH H. BOARDMAN, Commissioner, New York State DOT

SARAH C. CAMPBELL, President, TransManagement, Inc., Washington, DC

E. DEAN CARLSON, Secretary of Transportation, Kansas DOT

JOANNE F. CASEY, President, Intermodal Association of North America

JAMES C. CODELL III, Secretary, Kentucky Transportation Cabinet

JOHN L. CRAIG, Director, Nebraska Department of Roads

BERNARD S. GROSECLOSE, JR., President and CEO, South Carolina State Ports Authority

SUSAN HANSON, Landry University Professor of Geography, Graduate School of Geography, Clark University

LESTER A. HOEL, L. A. Lacy Distinguished Professor, Department of Civil Engineering, University of Virginia

HENRY L. HUNGERBEELER, Director, Missouri DOT

ADIB K. KANAFANI, Cahill Professor and Chairman, Department of Civil and Environmental Engineering, University of California

at Berkeley

RONALD F. KIRBY, Director of Transportation Planning, Metropolitan Washington Council of Governments

HERBERT S. LEVINSON, Principal, Herbert S. Levinson Transportation Consultant, New Haven, CT

MICHAEL D. MEYER, Professor, School of Civil and Environmental Engineering, Georgia Institute of Technology

JEFF P. MORALES, Director of Transportation, California DOT

KAM MOVASSAGHI, Secretary of Transportation, Louisiana Department of Transportation and Development

CAROL A. MURRAY, Commissioner, New Hampshire DOT

DAVID PLAVIN, President, Airports Council International, Washington, DC

JOHN REBENSDORF, Vice President, Network and Service Planning, Union Paciﬁc Railroad Co., Omaha, NE

CATHERINE L. ROSS, Executive Director, Georgia Regional Transportation Agency

JOHN M. SAMUELS, Senior Vice President-Operations Planning & Support, Norfolk Southern Corporation, Norfolk, VA

PAUL P. SKOUTELAS, CEO, Port Authority of Allegheny County, Pittsburgh, PA

MARTIN WACHS, Director, Institute of Transportation Studies, University of California at Berkeley

MICHAEL W. WICKHAM, Chairman and CEO, Roadway Express, Inc., Akron, OH

MIKE ACOTT, President, National Asphalt Pavement Association (ex officio)

MARION C. BLAKEY, Federal Aviation Administrator, U.S.DOT (ex officio)

REBECCA M. BREWSTER, President and CEO, American Transportation Research Institute, Atlanta, GA (ex officio)

THOMAS H. COLLINS (Adm., U.S. Coast Guard), Commandant, U.S. Coast Guard (ex officio)

JENNIFER L. DORN, Federal Transit Administrator, U.S.DOT (ex officio)

ELLEN G. ENGLEMAN, Research and Special Programs Administrator, U.S.DOT (ex officio)

ROBERT B. FLOWERS (Lt. Gen., U.S. Army), Chief of Engineers and Commander, U.S. Army Corps of Engineers (ex officio)

HAROLD K. FORSEN, Foreign Secretary, National Academy of Engineering (ex officio)

EDWARD R. HAMBERGER, President and CEO, Association of American Railroads (ex officio)

JOHN C. HORSLEY, Executive Director, American Association of State Highway and Transportation Officials (ex officio)

MICHAEL P. JACKSON, Deputy Secretary of Transportation, U.S.DOT (ex officio)

ROGER L. KING, Chief Applications Technologist, National Aeronautics and Space Administration (ex officio)

ROBERT S. KIRK, Director, Office of Advanced Automotive Technologies, U.S. Department of Energy (ex officio)

RICK KOWALEWSKI, Acting Director, Bureau of Transportation Statistics, U.S.DOT (ex officio)

WILLIAM W. MILLAR, President, American Public Transportation Association (ex officio)

MARY E. PETERS, Federal Highway Administrator, U.S.DOT (ex officio)

SUZANNE RUDZINSKI, Director, Office of Transportation and Air Quality, U.S. Environmental Protection Agency (ex officio)

JEFFREY W. RUNGE, National Highway Traffic Safety Administrator, U.S.DOT (ex officio)

ALLAN RUTTER, Federal Railroad Administrator, U.S.DOT (ex officio)

ANNETTE M. SANDBERG, Deputy Administrator, Federal Motor Carrier Safety Administration, U.S.DOT (ex officio)

WILLIAM G. SCHUBERT, Maritime Administrator, U.S.DOT (ex officio)

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

Transportation Research Board Executive Committee Subcommittee for NCHRP

GENEVIEVE GIULIANO, University of Southern California,

Los Angeles (Chair)

E. DEAN CARLSON, Kansas DOT

LESTER A. HOEL, University of Virginia

JOHN C. HORSLEY, American Association of State Highway and

Transportation Officials

MARY E. PETERS, Federal Highway Administration

ROBERT E. SKINNER, JR., Transportation Research Board

MICHAEL S. TOWNES, Transportation District Commission

of Hampton Roads, Hampton, VA

TRANSPORTATION RESEARCH BOARD

WASHINGTON, D.C.

2003

www.TRB.org

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

NCHRP REPORT 493

Research Sponsored by the American Association of State Highway and Transportation Officials

in Cooperation with the Federal Highway Administration

SUBJECT AREAS

Highway Operations, Capacity, and Traffic Control

Evaluation of Traffic Signal

Displays for Protected/

Permissive Left-Turn Control

CHRIS L. B

REHMER

Kittelson & Associates, Inc.

Portland, OR

KENT C. KACIR

Siemens Energy & Automation, Inc.

Intelligent Transportation Systems

Portland, OR

DAVID A. NOYCE

University of Massachusetts

Amherst, MA

MICHAEL P. MANSER

Texas Transportation Institute

College Station, TX

NATIONAL COOPERATIVE HIGHWAY RESEARCH

PROGRAM

Systematic, well-designed research provides the most effective

approach to the solution of many problems facing highway

administrators and engineers. Often, highway problems are of local

interest and can best be studied by highway departments

individually or in cooperation with their state universities and

others. However, the accelerating growth of highway transportation

develops increasingly complex problems of wide interest to

highway authorities. These problems are best studied through a

coordinated program of cooperative research.

In recognition of these needs, the highway administrators of the

American Association of State Highway and Transportation

Officials initiated in 1962 an objective national highway research

program employing modern scientiﬁc techniques. This program is

supported on a continuing basis by funds from participating

member states of the Association and it receives the full cooperation

and support of the Federal Highway Administration, United States

Department of Transportation.

The Transportation Research Board of the National Academies

was requested by the Association to administer the research

program because of the Board’s recognized objectivity and

understanding of modern research practices. The Board is uniquely

suited for this purpose as it maintains an extensive committee

structure from which authorities on any highway transportation

subject may be drawn; it possesses avenues of communications and

cooperation with federal, state and local governmental agencies,

universities, and industry; its relationship to the National Research

Council is an insurance of objectivity; it maintains a full-time

research correlation staff of specialists in highway transportation

matters to bring the ﬁndings of research directly to those who are in

a position to use them.

The program is developed on the basis of research needs

identiﬁed by chief administrators of the highway and transportation

departments and by committees of AASHTO. Each year, speciﬁc

areas of research needs to be included in the program are proposed

to the National Research Council and the Board by the American

Association of State Highway and Transportation Officials.

Research projects to fulﬁll these needs are deﬁned by the Board, and

qualiﬁed research agencies are selected from those that have

submitted proposals. Administration and surveillance of research

contracts are the responsibilities of the National Research Council

and the Transportation Research Board.

The needs for highway research are many, and the National

Cooperative Highway Research Program can make signiﬁcant

contributions to the solution of highway transportation problems of

mutual concern to many responsible groups. The program,

however, is intended to complement rather than to substitute for or

duplicate other highway research programs.

Note: The Transportation Research Board of the National Academies, the

National Research Council, the Federal Highway Administration, the American

Association of State Highway and Transportation Officials, and the individual

states participating in the National Cooperative Highway Research Program do

not endorse products or manufacturers. Trade or manufacturers’ names appear

herein solely because they are considered essential to the object of this report.

Published reports of the

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

are available from:

Transportation Research Board

Business Office

500 Fifth Street, NW

Washington, DC 20001

and can be ordered through the Internet at:

http://www.national-academies.org/trb/bookstore

Printed in the United States of America

NCHRP REPORT 493

Project G3-54(2) FY 1995

ISSN 0077-5614

ISBN 0-309-08757-0

Library of Congress Control Number 2003107173

Price $31.00

NOTICE

The project that is the subject of this report was a part of the National Cooperative

Highway Research Program conducted by the Transportation Research Board with the

approval of the Governing Board of the National Research Council. Such approval

reﬂects the Governing Board’s judgment that the program concerned is of national

importance and appropriate with respect to both the purposes and resources of the

National Research Council.

The members of the technical committee selected to monitor this project and to review

this report were chosen for recognized scholarly competence and with due

consideration for the balance of disciplines appropriate to the project. The opinions and

conclusions expressed or implied are those of the research agency that performed the

research, and, while they have been accepted as appropriate by the technical committee,

they are not necessarily those of the Transportation Research Board, the National

Research Council, the American Association of State Highway and Transportation

Officials, or the Federal Highway Administration, U.S. Department of Transportation.

Each report is reviewed and accepted for publication by the technical committee

according to procedures established and monitored by the Transportation Research

Board Executive Committee and the Governing Board of the National Research

Council.

The National Academy of Sciences is a private, nonproﬁt, self-perpetuating society of distinguished schol-

ars engaged in scientiﬁc and engineering research, dedicated to the furtherance of science and technology

and to their use for the general welfare. On the authority of the charter granted to it by the Congress in

1863, the Academy has a mandate that requires it to advise the federal government on scientiﬁc and techni-

cal matters. Dr. Bruce M. Alberts is president of the National Academy of Sciences.

The National Academy of Engineering was established in 1964, under the charter of the National Acad-

emy of Sciences, as a parallel organization of outstanding engineers. It is autonomous in its administration

and in the selection of its members, sharing with the National Academy of Sciences the responsibility for

advising the federal government. The National Academy of Engineering also sponsors engineering programs

aimed at meeting national needs, encourages education and research, and recognizes the superior achieve-

ments of engineers. Dr. William A. Wulf is president of the National Academy of Engineering.

The Institute of Medicine was established in 1970 by the National Academy of Sciences to secure the

services of eminent members of appropriate professions in the examination of policy matters pertaining to

the health of the public. The Institute acts under the responsibility given to the National Academy of

Sciences by its congressional charter to be an adviser to the federal government and, on its own

initiative, to identify issues of medical care, research, and education. Dr. Harvey V. Fineberg is president

of the Institute of Medicine.

The National Research Council was organized by the National Academy of Sciences in 1916 to associate

the broad community of science and technology with the Academy’s purposes of furthering knowledge and

advising the federal government. Functioning in accordance with general policies determined by the Acad-

emy, the Council has become the principal operating agency of both the National Academy of Sciences and

the National Academy of Engineering in providing services to the government, the public, and the scientiﬁc

and engineering communities. The Council is administered jointly by both the Academies and the Institute

of Medicine. Dr. Bruce M. Alberts and Dr. William A. Wulf are chair and vice chair, respectively, of the

National Research Council.

The Transportation Research Board is a division of the National Research Council, which serves the

National Academy of Sciences and the National Academy of Engineering. The Board’s mission is to promote

innovation and progress in transportation by stimulating and conducting research, facilitating the

dissemination of information, and encouraging the implementation of research results. The Board’s varied

activities annually engage more than 4,000 engineers, scientists, and other transportation researchers and

practitioners from the public and private sectors and academia, all of whom contribute their expertise in the

public interest. The program is supported by state transportation departments, federal agencies including

the component administrations of the U.S. Department of Transportation, and other organizations and

individuals interested in the development of transportation. www.TRB.org

www.national-academies.org

COOPERATIVE RESEARCH PROGRAMS STAFF FOR NCHRP REPORT 493

ROBERT J. REILLY, Director, Cooperative Research Programs

CRAWFORD F. JENCKS, Manager, NCHRP

B. RAY DERR, Senior Program Officer

EILEEN P. DELANEY, Managing Editor

HILARY FREER, Associate Editor II

NCHRP PROJECT G3-54(2) PANEL

Field of Traffic—Area of Operations and Control

ROBERT E. MAKI, City of Surprise, AZ (Chair)

NADER A. AYOUB, Texas Highway Products Corporation, Round Rock, TX

GEORGE L. BUTZER, DLZ Corporation, Columbus, OH

RAJ S. GHAMAN, FHWA

JOSEPH E. HUMMER, North Carolina State University

LOUIS E. LIPP, Aurora, CO

TIMOTHY J. SZWEDO, New Jersey DOT

W. SCOTT WAINWRIGHT, FHWA

JOHN H. WARNER, DMJM + Harris Inc., Tucson, AZ

THOMAS GRANDA, FHWA Liaison Representative

RICHARD A. CUNARD, TRB Liaison Representative

AUTHOR ACKNOWLEDGMENTS

Kittelson & Associates, Inc. (KAI) performed the research proj-

ect reported herein under NCHRP Project 3-54 in association with

Siemens Gardner Transportation Systems, the University of Mass-

achusetts at Amherst (UMass), and the Texas Transportation Insti-

tute (TTI). The project was directed by Kent C. Kacir, formerly of

KAI, now of Siemens Gardner Transportation Systems, as the Prin-

cipal Investigator. Bill Kloos, also formally of KAI, now with the

City of Portland, Oregon, also provided project direction as the Co-

Principal Investigator. The other authors of this report are David A.

Noyce, former Assistant Professor of Civil Engineering at the Uni-

versity of Massachusetts at Amherst, now Assistant Professor of

Civil Engineering at University of Wisconsin-Madison, and

Christopher L. Brehmer of KAI.

The work performed at Kittelson & Associates, Inc., was con-

ducted under the direction of Kent Kacir. Wayne Kittelson, Senior

Principal at KAI, served as the Project Principal. Chris Brehmer of

KAI served as the Senior Engineer on the project. The work per-

formed at UMass was conducted under the supervision of Dr.

Noyce with the assistance of Michael A. Knodler Jr., Graduate

Research Assistant. The work performed at TTI was conducted

under the supervision of Roger J. Koppa, Associate Professor, with

the assistance of Michael P. Manser, Associate Research Scientist,

and Jacqueline Jenkins, Graduate Assistant Research. The authors

wish to recognize the invaluable contributions of the late Daniel B.

Fambro, Professor, TTI, who directed much of the study plan devel-

opment of this research project. His insights and commitment to the

project were signiﬁcant to its success.

The project was guided by the panel members listed above, as

well as the following technical advisors:

Mr. Frank Dolan, Bergmann Associates

Mr. John Black, NAZTEC, Inc.

Mr. Ron Cameron, Gray Osborn

Mr. Tom Rathbun, Michigan DOT

Mr. Jim Poston, Meyer, Mohaddes Associates

This report recommends traffic signal displays for protected/permissive left-turn con-

trol. The recommendations are based on a comprehensive evaluation of the safety and

effectiveness of alternative traffic signal displays and phasing through laboratory and

ﬁeld studies. These studies are summarized in the report and detailed information on

them is available on the enclosed CD-ROM, CRP-CD-35. Traffic signal designers and

operators will ﬁnd the report informative, as will those interested in human factors

research. A key audience for the report will be those responsible for the Manual on Uni-

form Traffic Control Devices (MUTCD) because it is intended that the recommenda-

tions be considered for the next edition.

Protected/permissive left-turn (PPLT) traffic controls increase the left-turn capac-

ity and reduce delay at intersections by providing an exclusive turn phase for left turns

as well as a phase during which left turns can be made as opposing traffic will allow.

The protected left turn can either lead (or precede) or lag (or follow) the opposing

through signal phase. PPLT controls have been implemented in a variety of ways,

because the MUTCD provides limited guidance. At least six displays to indicate the per-

missive phase are known to exist in the United States (i.e., straight ﬁve-section head with

circular green, ﬁve-section cluster head with circular green, ﬂashing circular red, ﬂash-

ing circular yellow, ﬂashing red arrow, and ﬂashing yellow arrow). Variations also exist

in the phasing, signal displays, arrangement, signal placement, and use of supplemental

signs. There have been concerns that some of these variations may confuse motorists,

and validation of their relative operational and safety advantages was needed.

A key concern with PPLT control is the “yellow trap,” which occurs during the

change from permitted left turns in both directions to a lagging protected left turn in one

direction. The MUTCD requires that all circular signal indications on an approach to an

intersection display the same color. The left-turning driver whose permitted interval is

ending may try to sneak through the intersection on the yellow indication, not realizing

that the opposing through traffic still has a green indication. To avoid the yellow trap,

most agencies do not use leading/lagging PPLT. An innovation known as “Dallas Dis-

play” allows this operation without the yellow trap by operating the permissive left turns

simultaneously with the opposing through movement. Previous research has shown that

this operation reduces delay and improves safety, but is not easily implemented in all

situations.

Under NCHRP Projects 3-54 and 3-54(2), Kittelson & Associates and their sub-

contractors reviewed the literature and surveyed state and local transportation agencies

to determine what displays are used for PPLT and the prevalance of each. They then

assessed the safety and operational characteristics of each display qualitatively and ana-

lyzed crash data for a more quantitative assessment. Of particular interest was how the

different displays handled the yellow trap. Surveys of drivers using both static pho-

FOREWORD

By B. Ray Derr

Staff Officer

Transportation Research

Board

tographs and video were conducted to better understand how well the various displays

are comprehended. Following a meeting with the oversight panel to select the best dis-

plays for further study, driving simulator testing was used to gain a better understand-

ing of driver comprehension in a dynamic environment. Because the research conducted

pointed to the ﬂashing yellow arrow as a promising display, pilot installations were made

in several cities to determine how well it operates in real-life conditions and to identify

implementation issues.

In the interest of brevity, this report presents only the highlights of a very compre-

hensive project. Working papers from the individual studies mentioned above are

included on the enclosed CD-ROM, CRP-CD-35.

1 SUMMARY

13 DEFINITIONS

15 CHAPTER 1 Introduction

Research Problem Statement, 15

PPLT Displays, 15

Lead-Lag Phasing with PPLT Control, 16

NCHRP 3-54 Research Objective, 17

Development of Research Plan, 19

Review of Key Work Plan Activities, 19

Identify Technical Advisors, 19

Agency Survey, 20

Photographic Driver Studies, 20

Field Traffic Operations Studies, 22

Crash Data Analysis, 22

Engineering Assessment, 22

Data Analysis and Report of Preliminary Findings, 22

Driver Conﬁrmation Studies, 22

Field Implementation Studies, 22

Report Organization, 22

23 CHAPTER 2 Background

Modes of Left-Turn Control, 23

Current Standards for Protected/Permissive Control, 23

Signal Display Placement, 23

Operation, 26

Design, 27

Maintenance, 27

Uniformity, 27

Signal Phasing, 27

Alternative Displays Used in the United States, 29

Flashing Red Display, 29

Flashing Yellow Display, 30

Lead-Lag Displays, 30

Practices Outside the United States, 30

Canada, 31

Europe, 31

Australia, 32

The Left-Turn Problem, 32

Advantages of Protected/Permissive Left-Turn Phasing, 32

Disadvantages of the MUTCD Circular Green Display, 32

Summary, 33

34 CHAPTER 3 Research Activities and Findings

Agency Survey, 34

Objective, 34

Methodology, 34

Results, 34

Findings, 34

Photographic Driver Study, 35

Objective, 35

Methodology, 35

Results, 39

Findings, 40

Field Traffic Operations Study, 44

Objective, 44

Methodology, 44

Results, 44

Findings, 44

Field Traffic Conﬂict Study, 46

Objective, 46

Methodology, 46

Results, 48

Findings, 48

CONTENTS

Crash Data Analysis, 48

Objective, 48

Methodology, 48

Results, 49

Data Analysis, 49

Findings, 49

Driver Conﬁrmation Study, 50

Objective, 51

Signal Displays Studied, 51

Simulator Environment, 51

Simulator Study Methodology, 51

Opposing Traffic, 53

Video-Based Static Evaluation, 55

Conﬁrmation Study Sample Size, 55

Conﬁrmation Study Findings, 55

Driving Simulator Findings, 56

Static Evaluation Findings, 57

Driving Simulator and Static Evaluation Comparison Findings, 57

Implications of Driver Comprehension Study Findings, 58

Field Implementation Study, 58

Objective, 58

Methodology, 58

Implementation Study Findings, 59

Conﬂict Analysis Findings, 62

Field Observations During Flashing Yellow Arrow Activation, 63

Post-Implementation Survey of Volunteer Agencies, 63

Agency Feedback, 69

Public Reaction, 69

Agencies Declining Participation, 69

Engineering Assessment, 69

Objective, 69

Methodology, 69

Findings, 74

75 CHAPTER 4 Discussion

Findings of the Conﬁrmation Study, 75

Findings of the Field Implementation Study, 76

Operational Advantages, 76

Protected-Only Operation, 76

Protected-Permissive Operation, 76

Permissive-Protected Operation, 76

Permissive-Only Operation, 76

Other Considerations, 77

Supplemental Display Arrangements, 77

Right-Turn Overlap Display, 77

Universal Application, 78

A Need for Change, 79

Public Support for the Display, 80

Implications for Nonconforming PPLT Displays, 80

81 CHAPTER 5 Recommendations

Recommendation #1: Incorporate Flashing Yellow Arrow Display into the

MUTCD, 81

Recommendation #1A: Displays, 81

Recommendation #1B: Location, 81

Recommendation #1C: Supplemental Signs, 81

Recommendation #1D: Phasing, 81

Recommendation 2: Conduct Follow-Up Study, 82

Recommendation 3: Restrict Use of Flashing Red Indications, 82

83 REFERENCES

The NCHRP 3-54 Project is the culmination of extensive research efforts to identify

the “best” traffic signal display for protected/permissive left-turn (PPLT) control. The

research efforts respond to decades of practice wherein practitioners have experimented

with various displays and signal phasing schemes that either avoided—by design—

safety problems (e.g., yellow trap) or attempted to convey a clearer message to the

driver on the correct right-of-way (i.e., permissive movement).

Conducted over a 7-year period, the NCHRP 3-54 project is the most comprehen-

sive study of the PPLT display to date. The research team members surveyed current

practice, studied driver understanding of known permissive displays in the United

States, analyzed crash data, analyzed operational data, studied the implementation of

an experimental permissive display, and conducted a conﬁrmation study using a full-

scale driving simulator to study driver understanding of the most promising permissive

displays.

Unlike previous research, the NCHRP 3-54 study focused heavily on human factors

and the techniques used to observe human reactions. Previous studies typically relied

on crash data to indicate how well drivers understood one display compared with other

displays or traffic control devices. However, crash data generally are unreliable for

assessing driver reaction to different displays because the level of detail of individual

crash reports typically does not identify the traffic signal control indications illumi-

nated at the time of the crash.

The NCHRP 3-54 study used modern techniques to present visual images or inter-

active situations whereby human reaction could be evaluated independently. Each of the

14 individual study tasks was successful in gathering pertinent data. The study task

results and ﬁndings led the research team to develop a concise set of recommendations.

Additionally, many ﬁndings from this study will lead to safer implementations of PPLT

control. For example, this study identiﬁed how well drivers understand multiple indica-

tions illuminated at the same time within the same display arrangement. How adjacent

signal display indication (e.g., through movement) affects drivers’ interpretations of the

left-turn display indications was also identiﬁed. Further, differences in drivers’ reactions

to ﬂashing indications versus steady indications were explored and documented.

SUMMARY

EVALUATION OF TRAFFIC SIGNAL

DISPLAYS FOR PROTECTED/

PERMISSIVE LEFT-TURN CONTROL

RESEARCH PROBLEM STATEMENT

The increase in traffic volume on urban roadways has led engineers to develop inno-

vative means to control traffic. With an increase in traffic volume, a driver has fewer

available gaps in the opposing through traffic to execute a left-turn maneuver safely.

To alleviate this situation, signal phasing was designed to provide a protected left-turn

phase for part of the signal cycle. The resulting increase in safety came at the expense

of operational efficiency.

To regain some of the lost efficiency, traffic signals were designed to protect the left-

turn movement during a portion of the signal cycle and to allow a permissive movement

during the remainder of the signal cycle, resulting in left-turn control that is commonly

known as protected-permissive left-turn control, or simply PPLT control (or phasing).

If a protected movement is warranted, PPLT control has been shown to increase left-

turn capacity and reduce delay at intersections (as compared with protected-only con-

trol) by providing an exclusive turn phase for left turns as well as a permissive phase

during which left turns can be made if gaps in opposing through traffic will allow, all

within the same cycle. The left-turn phase (interval) can precede (lead) or follow (lag)

the through phase.

PPLT Displays

Over the years, PPLT control has been implemented in various ways. Variability

occurs in signal display arrangement, placement, and permissive indications. The vari-

ance in implementation has been consistent with the Manual on Uniform Traffic Con-

trol Devices (MUTCD) because, historically, the manual provided limited guidance. The

Federal Highway Administration (FHWA) recommends a ﬁve-section signal display.

Consistent with the FHWA’s recommendation, many states have adopted the five-

section cluster or “doghouse” signal display arrangement as their standard. The five-

section cluster is located in a shared overhead position, between the through and turn-

ing lanes, providing a green arrow indication for the protected phase and a circular green

indication for the permissive phase. The circular green indication is shared with the

through movement.

Several transportation agencies within the United States have designed and imple-

mented unique PPLT phasing displays with the intent to convey a clearer message of

drivers’ left-turn control. Several unique displays have been implemented as experi-

mental traffic control devices with approval by FHWA. To date, at least ﬁve variations

of the permissive indications are in use in the United States: the MUTCD standard cir-

cular green indication; the ﬂashing circular red indication, the ﬂashing circular yellow

indication, the ﬂashing red arrow indication, and the ﬂashing yellow arrow indication.

Lead-Lag Phasing with PPLT Control

The left-turn phase can lead or lag the opposing through movement. Traffic engi-

neers often would like to increase operational efficiency on roadways by using lead-

lag signal phasing, but cannot safely do so if PPLT control is used. The MUTCD does

not preclude the use of lead-lag left turns with PPLT control; however, doing so will

create what is known as the “yellow trap.” The yellow trap condition essentially leads

the left-turning driver into the intersection when it is unsafe to do so, even though the

signal displays are correct. During the signal change from permissive movements in

both directions to a lagging protected movement in one direction, a yellow trap is pre-

sented to the left-turning driver whose permissive left-turn phase is terminating. The

yellow trap occurs when a signal changes from the permissive left-turn intervals in both

directions to a lagging protected movement in only one direction. A driver attempting

to make a left turn on the permissive circular green indication becomes trapped in the

intersection when the circular green indication turns yellow for the change interval (for

the through traffic). The left-turn driver who is attempting to clear the intersection sees

the adjacent through lanes receive the circular yellow indication for their change inter-

val. The left-turner mistakenly believes that the opposing traffic also has the yellow

change interval and so makes the left turn, in effect becoming a “sneaker.” The yellow

trap occurs because the opposing traffic does not, in fact, receive a yellow change inter-

val, but instead has a circular green indication in the through lanes and a protected left-

turn arrow indication. The potential for serious conﬂict occurs between the sneaker

vehicle and the opposing, non-stopping, through traffic.

To avoid the hazardous yellow trap situation, traffic engineers use either simultane-

ous leading (lead-lead) or lagging (lag-lag) left-turn phasing. With lead-lead (protected-

permissive) left-turn phasing, both left-turn phase indications are initially illuminated

together. With most modern signal controllers, if the left-turn demand diminishes on one

side, the opposing through-lane traffic can proceed concurrently with the remaining left-

turn phase movement traffic. Conversely, for lag-lag (permissive-protected) left-turn

phasing, both left-turn phase indications may not be illuminated together; however both

left-turn phases must also terminate together. The lag-lag left-turn phasing results in a

potential decrease in capacity and increase in delay. For both the protected-permissive

(lead-lead) and permissive-protected (lag-lag) signal phasing operation, the permissive

left-turn circular green indication can be illuminated for the through movement and the

left-turn movement. The driver making the left-turn movement may proceed if there is

an acceptable gap in the opposing traffic stream. The lead-lead and lag-lag signal phas-

ing operation has been in use for many years.

Since approximately the mid-1980s, some traffic engineers have implemented an

innovative signal phasing operation known as the “Dallas Display.” The Dallas Display

permits phase overlaps and was designed to eliminate the potentially unsafe yellow trap

situation by allowing a continued permissive left-turn during the opposite approach lead-

ing and lagging protected left-turn phase. The Dallas Display advances traffic engineers’

ability to maximize signal coordination by allowing protected-permissive and lead-lag

operation within the same signal cycle.

NEED FOR RESEARCH

Traffic engineers have long cited the advantages of implementing PPLT control

(e.g., improved operational efficiency and traffic progression, reduced vehicle delay,

reduced fuel consumption, and reduced air pollution). The disadvantage, some argue,

is that PPLT control can be deployed in such a way that the yellow trap is created, and

there is potential for driver confusion. Nevertheless, many practicing agencies have

found the advantages to outweigh the potential disadvantages. Additionally, several

agencies in the United States have deployed various types of signal phasing techniques

to avoid the yellow trap and/or resolve the potential driver confusion problem. Over the

past two decades (in some limited cases, three decades), some agencies have been

granted approval from FHWA to implement unique displays, or display arrangements,

on an experimental basis. The premise for these implementations was the potential for

a safer or more efficient traffic control device. Examples of these unique displays are

the ﬂashing red and ﬂashing yellow circular and arrow indications.

The National Committee on Uniform Traffic Control Devices (NCUTCD), which

provides guidance to the FHWA on the MUTCD, has expressed concern that the vari-

ety of PPLT controls currently in use may confuse drivers traveling throughout the

United States and has long proposed a comprehensive national study that would vali-

date the operational advantages and safety aspects of the various PPLT control devices

and signal arrangements.

NCHRP 3-54 RESEARCH OBJECTIVE

The objective of the NCHRP 3-54 project was to evaluate the safety and effective-

ness of different signal displays and phasing for PPLT control through laboratory and

ﬁeld studies. Study activities were designed to gather, analyze, and interpret data that

would serve as the basis for recommending a uniform display for PPLT control. The

study considered all current applications of PPLT control in the United States, includ-

ing arrangement, indications, placement, phasing sequence, and safety considerations

(e.g., the yellow trap).

NCHRP Project 3-54 did not develop any guidelines, warrants, or recommendations

for the use of PPLT control. The underlying assumption was that the traffic engineer

had decided that PPLT control is an appropriate left-turn treatment. The goal of this

research project was to identify the “best” or most appropriate signal display, includ-

ing arrangement and indications.

RESEARCH PLAN

NCHRP Project 3-54 consisted of individual study tasks as identiﬁed in Figure S-1. In

October of 1999, the research team and project panel met to review the study PPLT dis-

plays and to determine which displays showed the most potential. As part of the meet-

ing, several decisions were made, including the decision to reduce the number of dis-

plays to those with the most potential for further study. Additionally, the project panel

directed the research team to conduct a thorough evaluation of one particular display

through ﬁeld implementation. Brief descriptions of the key study activities follow.

(Chapter 3 provides detailed information on each study task.)

Identify Technical Advisors

Before beginning the formal study effort, potential members of a Technical Advi-

sory Group were sought. The Technical Advisory Group was to consist of at least ﬁve

knowledgeable professionals who would provide the project panel with ﬁrst-hand

experience and expertise in the evaluation of PPLT signal displays. Members of the

advisory group were to be well-respected and experienced traffic engineers knowl-

edgeable about the PPLT issue. The research team also sought regular interaction with

the Signals Technical Committee of the NCUTCD to provide a regular ﬂow of infor-

mation to the full committee.

Agency Survey

To assess the current state of the practice, the research team conducted an extensive

literature review to investigate the state of the art in PPLT display. Both published and

unpublished literature was evaluated, including literature from international sources.

This task also included the administration of a survey of transportation agencies to

solicit information regarding the number and type of left-turn traffic control devices in

use across the United States. The objective of the agency survey was to identify and

quantify the different types of PPLT displays in use. The survey was administered to

transportation professionals at the state and city levels who were directly involved with

PPLT design and installations. Data on PPLT installations in all 50 states and parts of

Canada were obtained.

Agency

Survey

Photographic

Driver Studies

Field Traffic

Operations

Studies

Field Traffic

Conflict

Studies

Crash Data

Analysis

Report

Preliminary

Findings

Driver

Confirmation

Studies

Field

Implementation

Studies

Identify

Displays for

Further

Study

Refine

Research

Techniqu

2nd Interim

Report

and

2nd Panel

Meetin

Final Report

Findings and

Recommendations

Present

Recommendations to

NCUT

Engineering

Assessmen

Identify

Technical

Advisor

Provide Input

Refine

Recommendatio

Provide Input

Figure S-1. Final work plan.

The signiﬁcant ﬁndings of the agency survey are as follows:

•

Nearly 110,000 traffic signals in the United States (29% of all signals in the United

States) have at least one approach with PPLT operation.

•

In 34 of the 50 states, the ﬁve-section cluster PPLT display is used most, with 41%

of the states reporting the use of only one PPLT as the state standard. Speciﬁcally,

63% of all PPLT displays are the ﬁve-section cluster display.

•

A total of 40% of the responding agencies always use the PPLT display for one of

the two required through movement displays, and another 37% sometimes do.

•

A total of 50% of the reporting agencies use supplemental displays (e.g., additional

pole-mounted display on the far left side of the intersection); therefore the left-turn

display can be visible to the through motorist.

•

The circular green PPLT indication was used in 165 of the 168 reporting agencies.

Other indications used were ﬂashing circular red, ﬂashing red arrow, ﬂashing cir-

cular yellow, and the ﬂashing yellow arrow.

Photographic Driver Studies

Photographic driver studies evaluated driver understanding of the circular green,

ﬂashing yellow arrow, ﬂashing circular yellow, ﬂashing red arrow, and ﬂashing circu-

lar red permissive displays in association with different arrangements, placements, and

traffic/lane conﬁgurations. Computer-generated PPLT signal displays on static back-

ground photographs were used to represent the dynamic aspects of the PPLT displays.

More than 300 drivers participated in each of eight geographic locations around the

United States. All study participants were licensed drivers.

The following ﬁndings were key to reﬁning the research plan and developing rec-

ommendations:

•

Simultaneous illumination of two display indications signiﬁcantly reduced driver

understanding.

•

The circular green indication had the lowest level of driver comprehension (nearly

50%) of all PPLT displays studied.

•

The through indication had little effect on driver comprehension of the left-turn

PPLT indication.

•

Age had a signiﬁcant effect on driver comprehension. In at least one PPLT scenario,

drivers over the age of 65 only responded correctly 20% of the time. Older drivers

responded more correctly to the ﬂashing circular red and yellow indications.

•

Flashing indications were understood better than steady indications, as evidenced

by overall time to respond to the indication in question.

•

The circular indication was better understood when compared with arrow indi-

cations.

•

The red indications were understood better than the other displays studied.

Field Traffic Operations Studies

The operations studies consisted of quantifying the capacity and delay associated

with various PPLT displays by analyzing saturation ﬂow rates, lost times, response

times, and follow-up headways. The traffic studies were completed in the same eight

geographic locations in which the photographic driver studies were conducted to pro-

vide additional insights into driver behavior related to alternative PPLT displays. As

part of this effort, the research team worked with local traffic engineers to identify rep-

resentative sites with PPLT displays and to then gather crash, traffic, geometry, and

other relevant data for each site. More than 8 hr of conﬂict data per intersection were

collected at three intersections within each geographic study area.

This work activity identiﬁed that study location was a signiﬁcant factor, but that

PPLT signal display arrangement and phasing was not a signiﬁcant factor affecting the

saturation ﬂow rate. Additionally, this work activity showed that signal phasing sig-

niﬁcantly inﬂuenced the start-up lost time, but that PPLT signal display arrangement

or study location did not.

Field Traffic Conﬂict Studies

The traffic conﬂict studies focused solely on the left-turn movement. Conﬂict stud-

ies were completed in the same eight geographic locations as the photographic driver

and ﬁeld traffic operations studies. The purpose of these studies was to quantify left-

turn conﬂict rates and event rates for different PPLT signal displays and indications.

The conﬂict studies involved the collection and analysis of 8 hr of conﬂict data on a

typical weekday at 24 study intersections. At each intersection, traffic conﬂicts were

observed and classiﬁed into one of six types.

The conﬂict analysis study task showed that, overall, traffic conﬂicts were low for

all PPLT displays evaluated, and few left-turn conﬂicts were associated with the PPLT

display.

Crash Data Analysis

The research team conducted a review of crash data to determine and compare left-

turn crash rates associated with various PPLT displays. In addition, selected compo-

nents of a crash database created in 1988 as part of an FHWA study were examined.

The crash analysis showed that, based on (1) the average crash rate per year and

(2) average crashes per 100 left-turning vehicles, the four-section ﬂashing circular yel-

low indication used in Seattle, Washington, experienced the fewest number of crashes

per year. The analysis of the average crash rate by intersection indicated that the ﬂash-

ing red arrow used in Cupertino, California, had the lowest average crash rate. The cir-

cular green indication had the highest average statistic in three of the four evaluations

studied and reported above.

Ultimately, the crash data analysis did not identify any safety problems among the

PPLT displays studied. The data showed that the ﬂashing circular yellow indication

typically had the lowest (best) statistic and the circular green indication (MUTCD stan-

dard) had the highest (worst) statistic.

Engineering Assessment

The engineering assessment sought to identify the objective and subjective infor-

mation needed to evaluate the proposed displays. To assist the assessment, an evalua-

tion matrix was developed. The evaluation matrix included considerations addressing

the key issues of safety, operations, implementation, human factors, and versatility and

was updated as each study activity was completed.

The engineering assessment was conducted continuously over the course of the

study. At the conclusion of the research tasks discussed above, the research team made

the following assessments:

•

The ﬂashing yellow arrow and the circular green indication can be used in both

exclusive and shared PPLT signal displays, whereas the ﬂashing yellow circular

indication and ﬂashing red arrow indications can only be used in an exclusive

PPLT display.

•

The ﬂashing yellow arrow indication and circular green indication present the best

displays for further study.

Report of Preliminary Findings

Using the data collected to date, the research team prepared a comprehensive

analysis of the results obtained from the agency evaluation, photographic driver stud-

ies, conflict studies, and crash studies. This information was presented to the mem-

bers of the project panel, who in turn reviewed the findings and ultimately reduced

the number of PPLT displays for future consideration. In addition to selecting a sub-

set of promising PPLT displays to be studied further, the project panel provided

direction to complete the confirmation study and to conduct a field implementation

study.

Driver Conﬁrmation Studies

Before the research team could make a ﬁrm recommendation on a “best” display to

be used as a national standard, the research team conducted a conﬁrmation study

whereby human subjects (drivers) were tested in a full-scale simulator. The conﬁrma-

tion study evaluated 12 PPLT signal display scenarios, each with a different permis-

sive indication, display face, location, and through movement indication. Each of the

PPLT signal displays included only the circular green indication and/or ﬂashing yel-

low permissive indications. Two separate conﬁrmation studies were conducted—one

at the University of Massachusetts and one at the Texas A&M University. The two con-

ﬁrmation studies were created to be near identical, within the practical budget limits of

the study.

The driver conﬁrmation study demonstrated the following:

•

There was a high level of comprehension with no variation between the different

PPLT displays tested.

•

The data showed no statistical difference in driver comprehension when the data

were cross-analyzed by permissive indication, display arrangement, through indi-

cation, and/or location of the display.

•

Males and females had statistically equivalent levels of comprehension.

•

Drivers over the age of 45 had signiﬁcantly fewer fail-critical responses than those

45 and younger.

•

Drivers who drive between 10,000 and 20,000 miles per year had signiﬁcantly

more responses that were correct and signiﬁcantly fewer fail-critical responses

than those who drove fewer than 10,000 miles per year.

•

The level of education that the drivers possessed did not have a signiﬁcant effect

on comprehension.

•

PPLT displays with the four-section vertical display face resulted in a signiﬁcantly

greater number of correct responses compared with the ﬁve-section vertical and

ﬁve-section cluster display arrangements.

•

The location of the PPLT signal display did not result in statistically signiﬁcant

differences.

Field Implementation Studies

An experimental ﬂashing yellow arrow display was ﬁeld tested because it had shown

promise from both safety and driver comprehension perspectives in previous task activ-

ities and had ranked high in the Engineering Assessment. Volunteer agencies were

sought from across the United States to install and operate the experimental ﬂashing

yellow arrow display. In conjunction with the use of the experimental display (which

required permission from FHWA to operate), “before” and “after” studies were com-

pleted at each intersection where the ﬂashing yellow arrow display was installed and

at nearby control sites. These “before” and “after” studies allowed the research team to

quantify the impact of the changeover from the MUCTD circular green indication to

the ﬂashing yellow arrow indication.

The ﬂashing yellow arrow indication was implemented in the following six U.S.

jurisdictions:

•

Montgomery County, Maryland;

•

Tucson, Arizona;

•

Woodburn, Oregon;

•

Jackson County, Oregon;

•

Beaverton, Oregon; and

•

Broward County, Florida.

With the exception of the PPLT display implemented in Tucson, Arizona, all PPLT

displays were implemented with few problems and remain operational as of the date of

this report. As explained in greater detail in Chapter 3, the PPLT display implemented

in Tucson, Arizona, was removed from operation because of safety concerns expressed

by city management. For each of the 15 intersections under study, only the PPLT dis-

play was changed from the MUTCD standard circular green indication to the ﬂashing

yellow arrow display.

Findings of the ﬁeld implementation study include the following:

•

Review of before and after ﬁeld conﬂict data showed no differences attributable

to the change in the PPLT display.

•

Before and after observations of follow-up headway and ﬂow rate data for the

study and control intersections demonstrated that the change in PPLT display had

negligible impact.

•

The implementation study demonstrated that the ﬂashing yellow arrow display can

be ﬁeld implemented (albeit with minor issues) using existing traffic control hard-

ware and software. Technical implementation issues identiﬁed during the ﬁeld

implementation process could be dealt with appropriately in any future develop-

ments of hardware and/or software, should the ﬂashing yellow arrow display become

a standard.

•

Traffic engineers who participated in the implementation study (and those who

declined participation) generally expressed their approval of the ﬂashing yellow

arrow display because of the following:

– It provides an exclusive signal display for the left-turn control.

– The indication was ﬂashing, which attracted more attention.

– The indication provided enhanced operation control.

•

Public comments from citizens who experienced the ﬂashing yellow arrow display

in the ﬁeld were generally positive. Several implementing agencies reported

receiving e-mails or written letters from the motoring public with most, if not all,

in support of the ﬂashing yellow arrow display.

Engineering Assessment (Update)

At the completion of the driver conﬁrmation studies and the ﬁeld implementation

studies, the research team reviewed the engineering assessment ﬁndings to reﬂect cur-

rent data. The ﬁnal ﬁndings are reported below:

•

The ﬂashing yellow arrow display was shown to offer the highest level of safety.

•

The circular green indication using the Dallas Display and the ﬂashing yellow

arrow display was shown to rank “best” in the category of operations.

•

The circular green indication was shown to rank “best” as being implementable.

•

The ﬂashing yellow arrow display was shown to be the “best” in the category of

human factors.

•

The ﬂashing yellow arrow display was shown to have the most versatile charac-

teristics and the circular green indication was the least versatile.

RECOMMENDATIONS

Based on the results and ﬁndings of the various research activities, the research team

and the project panel identiﬁed the following three recommendations:

Recommendation 1: The ﬂashing yellow arrow display should be included in the

MUTCD as an allowable alternative display to the circular green indication when used

in PPLT control/operation.

Displays. The four-section, all-arrow display face should be the only display allowed.

The only display that justiﬁes an exception to this recommendation is the three-section

display face with bi-modal lens. The three-section display face with bi-modal lens should

also be allowed given that it operates the same as does the four-section display face.

Only one indication shall be illuminated at any time.

Location. The ﬂashing yellow arrow operation shall only be used in an exclusive

signal arrangement. It is recommended, but not required, that the left-turn signal face

be placed over the left turn lane.

Supplemental Signs. Supplemental signing is not warranted with ﬂashing yellow

arrow display. Use of supplemental signing is optional.

Phasing. When used for left-turn treatments, the ﬂashing yellow arrow display shall

be tied to the opposing through green indication/display.

Recommendation 2: Conduct Follow-Up Study

It is recommended that a follow-up study to this project be conducted. The follow-

up study should be conducted after there has been ample time for an implementation

trial period by agencies currently participating in the ﬁeld implementation as well as

other potential additional agencies that may choose to implement based on the ﬁndings

of this research effort. Sufficient time should also be allowed such that before and after

crash data can be acquired at the study intersections and corresponding control sites.

The follow-up study should

•

Analyze all available crash data for the experimental ﬂashing yellow arrow dis-

plays implemented as part of this study;

•

Identify whether the ﬂashing yellow arrow display should become the only dis-

play allowed in the MUTCD for PPLT operation;

•

Identify whether, if the ﬂashing yellow arrow is selected to become the only dis-

play allowed for PPLT, the MUTCD should also be changed to add the following

prohibition: “For Permissive Only Mode operation, a signal face displaying a cir-

cular green indication shall not be located directly over or in line with a left-turn

lane”; and

•

Identify an implementation plan.

Recommendation 3: Restrict Use of Flashing Red Indications

The use of the ﬂashing red indication should only be implemented at locations where

an engineering study has identiﬁed that all drivers must come to a complete stop before

proceeding on the permissive interval.

PROJECT FINDINGS SUMMARY

Figure S-2 illustrates the ﬂow of work activities and the relevant ﬁndings that lead

to the study recommendations.

Agency Survey:

PPLT operation has widespread application across the U.S.

The five-section cluster is used in most PPLT displays.

The PPLT display is used often as a shared display.

Many agencies use supplemental displays to augment the main

display for the left-turn movement.

Driver Survey:

Driver understanding is reduced when two indications are illuminated

simultaneously.

The circular green indication used for the permissive left-turn

movement had the lowest level of understanding.

The circular flashing red indication had the highest level of

understanding.

Flashing indication is understood better than steady indications.

The circular indication is better understood than the arrow indication.

Field Traffic Operations Study:

The PPLT indication does not have an effect on saturation flow rate

and start up lost time.

Follow-up headways for the circular flashing red indications were a

factor of local signal operation (lead versus lag).

Conflict Study:

Few left-turn conflicts were associated with the PPLT display.

Many left-turn events associated with the PPLT display were related

to the simultaneous illumination of the circular red indication and the

protected green arrow indication.

The PPLT display has little effect on left-turn conflicts.

Crash Analysis:

The flashing circular yellow indication typically had the lowest (best)

crash statistic and the circular green indication had the highest

(worst) statistic.

Engineering Assessment:

The flashing yellow arrow and the circular green indication can be

used in both exclusive and shared PPLT signal displays, whereas the

flashing yellow ball and flashing red arrow indications can only be

used in an exclusive PPLT display.

The flashing yellow arrow indication and circular green ball indication

present the best displays for further study.

Total Population of PPLT Displays

Analysis of Findings

First Interim Report

First Panel Meeting

Identify Displays/Indications for more study

(Circular green and flashing yellow arrow indications)

Driver Confirmation Study:

High level of comprehension among all

PPLT displays studied.

No significant difference among permissive

indication, display arrangement, or location of

display.

Older drivers and drivers who drive more

have fewer fail-critical responses.

PPLT displays with the four-section vertical

face had a significantly greater number of

correct responses.

Field Implementation:

Conflict data showed no statistical differences

between circular green and flashing yellow

arrow indication.

In general, traffic engineers express their

approval of the flashing yellow arrow due to

greater operational control without a

degradation of safety.

Public reaction to the flashing yellow arrow

display has been mostly positive.

Analysis of Findings

Identify Recommendations

Second Interim Report

Second Panel Meeting

Engineering Assessment:

The flashing yellow arrow indication presented

better qualities in the categories of safety,

operations, human factors, and versatility.

The circular green indication cannot be

applied universally.

There are inherent safety problems associated

with the circular green indication.

The red indications can violate the meaning of

the stop condition.

The flashing yellow ball cannot be applied to

left- and right-turn treatment

Refine Recommendations

Final Report

Recommendations:

The flashing yellow arrow display should be

included in the MUTCD as an allowable

alternative display to the circular green

indication when used in PPLT control/

operation.

Conduct a follow-up study to analyze 'after'

crash data related to study intersections.

Restrict use of flashing red indications.

Study of Many Displays

Focus on a Few Displays/ Refined Study Focus Recommendations

Figure S-2. Work ﬂow of activities and ﬁndings that led to the development of recommendations.

DEFINITIONS

Exclusive Display: A display on a single approach for con-

trolling only the left-turn movement.

FHWA: Federal Highway Administration

Horizontal Arrangement: A particular arrangement of sig-

nal indications in a horizontal position. This term is the same

as the horizontal display face in the MUTCD.

Interval: A discrete portion of the signal cycle during which

the signal indications remain unchanged.

Lagging Left Turn: A phase sequence in which a protected

left-turn phase follows the opposing through-movement phase.

Lead Left Turn: A phase sequence in which a protected left-

turn phase precedes the opposing through-movement phase.

Lead-Lead Left Turn: Also called dual leading left turns.

Indicates a phase sequence in which two left-turn movements

from opposite directions of a roadway are both served by

leading protected phases. When both streets at the intersec-

tion are serviced in this manner, the operation is referred to

as quad left-turn phasing.

Lead-Lag Left Turn: A phase sequence in which one left-

turn movement is served by a leading protected phase and the

other left-turn movement (from the opposite direction of the

same street) is served by a lagging protected phase.

Lag-Lag Left Turn: A phase sequence in which two left-

turn movements from opposite directions of a street are both

served by lagging protected phases.

MUTCD: Manual on Uniform Traffic Control Devices

NCUTCD: National Committee on Uniform Traffic Control

Devices

Permissive Mode: A mode of traffic control signal opera-

tion in which left or right turns may be made on a circular

green signal indication after yielding to oncoming traffic and

pedestrians.

Phase: A part of the traffic signal time cycle allocated to any

combination of traffic movements receiving right-of-way

simultaneously during one or more intervals.

Phase Sequence: The order in which a controller cycles

through all phases.

The following list of deﬁnitions is intended to clarify the

manner in which they are used in this report. Some of the

terms listed may not appear in this document. The research

team has attempted to use terminology consistently through-

out this ﬁnal report and in the various working papers docu-

menting the detailed progression of the project; however,

there has been a change in direction within the industry and,

in particular, within the National Committee on Uniform Traf-

ﬁc Control Devices regarding terminology.

Change Interval: The yellow change interval follows the

green interval to warn traffic of an impending change in the

right-of-way assignment. The yellow change interval may be

followed by a red clearance interval.

Cluster Arrangement: Cluster arrangement refers to a par-

ticular arrangement of signal indications where four of them

are clustered in the form of a square, and the ﬁfth signal indi-

cation (circular red) is mounted directly on the top, either

centered or off to one side.

Cycle Length: See signal cycle.

Dallas Display: Also called Dallas Phasing. Also call Per-

missive Lead-Lag. A unique signal operation designed to

eliminate the “yellow trap” situation by allowing a continued

permissive left-turn during the opposite approach lagging

protected left-turn phase. In order to accomplish this, a left-

turn circular green indication is displayed, in an exclusive

display, during the adjacent through movement green and

yellow indications and continues to be displayed during the

lagging protected left-turn green arrow and through circular

green (or straight green arrow) indications for the opposite

approach. The continuing circular green left-turn indication is

terminated by a circular yellow indication that is simultane-

ously displayed with the opposing through movement yellow

clearance. This type of operation requires the use of visibility-

limited (e.g., louvered) signal faces to shield the circular green

and yellow indications intended for the left-turn lane traffic

from the adjacent through traffic.

Display: The signal face as a unit (assembly) that conveys

the message to the driver. The display consists of the indi-

vidual section, regardless of conﬁguration. Historically, this

term has been used in reference to the signal head.

Display Face: The part of a signal head provided for con-

trolling traffic in a single direction. Same as “signal face” and

“display.”

Doghouse: See cluster arrangement.

Preemption: The term used when the normal signal sequence

at an intersection is interrupted or altered in deference to a spe-

cial situation such as the passage of a train, a bridge opening,

or the granting of the right-of-way to an emergency vehicle.

Protected Mode: A mode of traffic control signal operation

in which left or right turns may be made when a left or right

green arrow signal indication is displayed.

Protected/Permissive Mode: A mode of traffic signal oper-

ation in which the left-turn movement is presented during

both the protected mode and the permissive mode on an

approach during the same signal cycle.

In this report, the term protected/permissive does not

necessarily imply a particular phasing order (i.e., protected/

permissive or permissive/protected). In practice, the phase

sequence is important.

Protected Permissive Left-Turn Display: Also referenced

as PPLT display. This term is used to reference speciﬁcally

or generically the indication or display used in conjunction

with protected permissive left-turn control.

Protected Permissive Left-Turn Control: Also referenced

as PPLT control or PPLT operation. This term is used to

identify the signal operation that uses protected permissive

signal phasing.

Shared Display: A signal display is considered to be a shared

display when it constitutes one of the two required displays

for the through movement and provides the left-turn move-

ment indication.

Signal Cycle: The total time required to complete either one

sequence of signal phases at a signalized intersection with

pretimed operation or a sequence of those phases with traffic

demand at a signalized intersection with traffic-actuated

operation.

Signal Display Arrangement: The signal arrangement as

a unit (assembly) that conveys the message to the driver.

The display arrangement consists of the individual sections,

regardless of conﬁguration. Historically, this term has been

used in reference to the signal head.

Signal Face: The part of a signal head provided for con-

trolling traffic in a single direction. This is the same as a dis-

play face.

Signal Head: Historically, this term has referred to an assem-

bly containing one or more signal faces that may be desig-

nated as one-way, two-way, and so forth. See also signal dis-

play or signal display arrangement.

Signal Indications: The illumination of a signal lens, such

as the circular green indication.

Signal Lens: The part of a signal section through which light

from the light source or reﬂectors passes. In doing so, light is

directed into a prescribed pattern, is ﬁltered to a prescribed

color, and, where necessary, is provided with a prescribed

symbol or message.

Signal Section: The assembly of a housing, lens, and light

source with necessary components and supporting hardware

providing one signal indication.

Signal System: Two or more traffic control signals operat-

ing in coordination.

Vertical Arrangement: A particular arrangement of signal

indications in a vertical position. This term is the same as the

vertical display face in the MUTCD.

Yellow Trap: A situation where the driver sees a yellow

indication for a change interval in the left-turn signal face

and the adjacent through signal face and assumes that the

opposing traffic also receives the yellow change interval.

CHAPTER 1

INTRODUCTION

This report is the culmination of extensive research efforts

to identify the “best” traffic signal display for protected/

permissive left-turn (PPLT) control. The research efforts

respond to decades of practice whereby practitioners have

experimented with various displays and signal phasing schemes

that either avoided—by design—safety problems (e.g., yellow

trap) or attempted to convey a clearer message to the driver

on the correct right-of-way (i.e., permissive movement).

Conducted over a 7-year period, the NCHRP 3-54 study is

the most comprehensive study of the PPLT display to date.

In the course of the study, the research team has surveyed

current practice, studied driver understanding of known per-

missive displays in the United States, analyzed crash data,

analyzed operational data, studied the implementation of an

experimental permissive display, and conducted a conﬁrma-

tion study using a full-scale driving simulator to study driver

understanding of the most promising permissive displays.

Unlike previous research, the NCHRP 3-54 study focused

on human factors and the techniques used to observe human

reactions. Previous studies typically relied on crash data to

indicate how well drivers understood one display compared

with other displays or traffic control devices. However, crash

data are generally unreliable for assessing driver reaction to

different displays because the level of detail of individual

crash reports typically does not identify the traffic signal con-

trol indications illuminated at the time of the crash.

The NCHRP 3-54 study used modern techniques to pre-

sent visual images and interactive situations whereby human

reaction could be independently evaluated. As detailed in

Chapter 3 of this report, 14 individual study tasks were com-

pleted, and each produced useful data. Through analysis and

review of the study task results and ﬁndings, the research

team was able to develop a concise set of recommendations.

Additionally, many ﬁndings that have come from this study

will lead to safer implementations of PPLT control. For exam-

ple, this study has identiﬁed how well drivers understand

multiple indications illuminated at the same time within the

same display arrangement. The effects of adjacent signal dis-

play indication (e.g., through movement) as they relate to

driver interpretation of the left-turn display indications were

also identiﬁed. Further, differences in driver reactions to ﬂash-

ing indications versus steady indications were explored and

documented.

As previously mentioned, this report is a culmination of

many tasks conducted over several years. Each major study

task was documented by a working paper. The reader is

encouraged to review the various working papers contained

as appendixes to this report and provided on the accompany-

ing CD-ROM for the task design, results, and ﬁndings.

The remainder of this chapter discusses the research prob-

lem statement, the need for more research, the objective of

the research project, the development of the research plan,

and an overview of study tasks completed in conjunction

with the overall study.

RESEARCH PROBLEM STATEMENT

The increase in traffic volume on urban roadways has led

engineers to develop innovative means to control traffic.

With an increase in traffic volume, a driver has fewer avail-

able gaps in the opposing through traffic to execute a left-

turn maneuver safely. To alleviate this situation, signal phas-

ing was designed to provide a protected left-turn phase for a

portion of the signal cycle. The resulting increase in safety

came at the expense of operational efficiency.

To regain some of the lost efficiency, traffic signals were

designed to protect the left-turn movement during a portion of

the signal cycle and allow a permissive movement during the

remainder of the signal cycle, thereby resulting in left-turn

control that is commonly known as protected-permissive left-

turn control, or simply PPLT control (or phasing). If a pro-

tected movement is warranted, PPLT control has been shown

to increase left-turn capacity and reduce delay at intersec-

tions (as compared with protected-only control) by providing

an exclusive turn phase for left turns as well as a permissive

phase during which left turns can be made if gaps in oppos-

ing through traffic will allow, all within the same cycle. The

left-turn phase (interval) can precede (lead) or follow (lag)

the through phase.

PPLT Displays

Over the years, PPLT control has been implemented in

various ways. Variability occurs in signal display arrange-

ment, placement, and permissive indications. The variance in

implementation has been consistent with the Manual on Uni-

form Traffic Control Devices (MUTCD) because the manual

historically provided limited guidance (1). The FHWA rec-

ommends a ﬁve-section signal display (2). Consistent with

the FHWA’s recommendation, many states have adopted the

ﬁve-section cluster or “doghouse” signal display arrange-

ment as their standard. The ﬁve-section cluster is located in

a shared overhead position typically between the through and

turning lanes, providing a green arrow indication for the pro-

tected phase and a circular green indication for the permis-

sive phase. The circular green indication is shared with the

through movement.

Several transportation agencies in theUnited States have

designed and implemented unique PPLT phasing displays so

as to convey more clearly drivers’ left-turn control. Several

unique displays have been implemented as experimental traf-

ﬁc control devices with approval by FHWA. To date, at least

ﬁve variations of the permissive indications are in use in the

United States: the MUTCD standard circular green indica-

tion; the ﬂashing circular red indication, the ﬂashing circular

yellow indication, the ﬂashing red arrow indication, and the

ﬂashing yellow arrow indication. Figure 1-1 illustrates the

various permissive display indication and display arrange-

ments that have been or are in use in the United States. Chap-

ter 3 of this report provides a detailed discussion of the PPLT

displays shown in Figure 1-1.

Lead-Lag Phasing with PPLT Control

The left-turn phase can lead or lag the opposing through

movement. Traffic engineers often want to increase opera-

tional efficiency on roadways by using lead-lag signal phas-

ing but cannot safely do so if PPLT control is used. The

MUTCD does not preclude the use of lead-lag left turns with

PPLT control; however, doing so will create what is known

as the “yellow trap.” The yellow trap condition essentially

leads the left-turning driver into the intersection when it may

not be safe to do so, even though the signal displays are cor-

rect. During the signal change from permissive movements

in both directions to a lagging protected movement in one

direction, a yellow trap is presented to the left-turning driver

whose permissive left-turn phase is terminating. Figure 1-2

illustrates how the yellow trap occurs.

As shown in Figure 1-2, the yellow trap occurs when a sig-

nal changes from the permissive left-turn intervals in both

directions to a lagging protected movement in only one direc-

tion. A driver attempting to make a left turn on the permissive

circular green indication becomes trapped in the intersection

when the circular green indication turns yellow for the change

interval (for the through traffic). The left-turn driver who is

attempting to clear the intersection sees the adjacent through

lanes receive the circular yellow indication for their change

interval. The left-turning driver mistakenly believes that the

opposing traffic also has the yellow change interval and so

makes the left turn, in effect becoming a sneaker. The yellow

trap occurs because the opposing traffic does not, in fact,

receive a yellow change interval but instead has a circular

green indication in the through lanes and a protected left-turn

arrow indication. The potential for serious conflict occurs

between the sneaker and the opposing, non-stopping, through

traffic.

To avoid the hazardous yellow trap situation, traffic

engineers use either simultaneous leading (lead-lead) or lag-

ging (lag-lag) left-turn phasing. With lead-lead (protected-

permissive) left-turn phasing, both left-turn phase indications

are initially illuminated together. With most modern signal

controllers, if the left-turn demand diminishes on one side, the

opposing through-lane traffic can proceed concurrently with

the remaining left-turn phase movement traffic. With lag-lag

(permissive-protected) left-turn phasing, both left-turn phase

indications may not be illuminated simultaneously; however

both left-turn phases must always terminate simultaneously.

The lag-lag left-turn phasing results in a potential decrease

in capacity and increase in delay. For both the protected-

permissive (lead-lead) and permissive-protected (lag-lag) sig-

nal phasing operation, the permissive left-turn circular green

indication can be illuminated for the through movement and

the left-turn movement. The driver making the left-turn move-

ment may proceed if there is an acceptable gap in the oppos-

ing traffic stream. The lead-lead and lag-lag signal phasing

operation has been in use for many years.

Since approximately the mid-1980s, some traffic engineers

have implemented an innovative signal phasing operation

known as the “Dallas Display.” The Dallas Display permits

phase overlaps and was designed to eliminate the potentially

unsafe yellow trap situation by allowing a continued permis-

sive left-turn during the opposite approach leading and lag-

ging protected left-turn phase. The Dallas Display advances

traffic engineers’ ability to maximize signal coordination by

allowing protected-permissive and lead-lag operation within

the same signal cycle. Research has shown that the Dallas

Display is operationally efficient and minimizes delay while

maintaining safety (3).

Traffic engineers have long cited the advantages of imple-

menting PPLT control, such as improved operational effi-

ciency and traffic progression, reduced vehicle delay, reduced

fuel consumption, and reduced air pollution. The disadvan-

tage, some argue, is that PPLT control can be deployed in such

a way that the yellow trap is created, thereby causing potential

driver confusion. Nevertheless, many practicing agencies have

found the advantages to outweigh the disadvantages. Addi-

tionally, several agencies in the United States have deployed

various types of signal phasing techniques to avoid the yel-

low trap, resolve the potential driver confusion problem, or

both. Over the past two decades (in some limited cases, three

decades), some agencies have been granted approval from

FHWA to implement unique displays or display arrangements

on an experimental basis. The premise for these implementa-

tions has been the potential for a safer or more efficient traffic

control device.

The National Committee on Uniform Traffic Control

Devices (NCUTCD), which provides guidance to FHWA on

the MUTCD, has expressed concern that the variety of PPLT

controls in use may confuse drivers traveling throughout the

United States and has long proposed a comprehensive national

study that would validate the operational advantages and

safety aspects of the various PPLT control devices and sig-

nal arrangements.

NCHRP 3-54 RESEARCH OBJECTIVE

The objective of NCHRP Project 3-54 was to evaluate

the safety and effectiveness of different signal displays and

phasing for PPLT control through laboratory and field stud-

Maryland

Seattle, WA

Washington

State

Delaware

Cupertino, CA

Michigan

Area

Used

Area

Used

Lens Color

and

Arrangement

Lens Color

and

Arrangement

* Green or Yellow

Bi-modal Lens

Left-Turn Indication Left-Turn Indication

Protected

Mode

Protected

Mode

Permitted

Mode

Permitted

Mode

Sparks, NVReno, NV

LEFT LEFT LEFTLEFT

i:\projfile\2036\cdrfiles\figures\Figure_1-1.cdr

Indicates Flashing

Figure 1-1. Variations in PPLT displays.

ies. Study activities were designed to gather, analyze, and

interpret data that would serve as the basis for recommend-

ing a uniform display for PPLT control. The study con-

sidered all current applications of PPLT control in the

United States, including arrangement, indications, place-

ment, phasing sequence, and safety considerations (e.g., the

yellow trap).

NCHRP Project 3-54 did not develop any guidelines,

warrants, or recommendations for the use of PPLT control.

The underlying assumption was that the traffic engineer had

decided that PPLT control is an appropriate left-turn treat-

ment. The goal of this research project was to identify the

“best” or most appropriate signal display, including arrange-

ment and indications.

Opposing

Through Signal

Opposing

Left-Turn Display Not Shown

All Red

Protected Left Turn

Clearance Interval

(End Protected Left-Turn)

Permissive Phase

Change Interval

(Yellow Trap)

Opposing Through

Phase Indication Still Green

i:\projfile\2036\cdrfiles\figures\Figure_1-2.cdr

Figure 1-2. Yellow trap with MUTCD 5-section PPLT display.

DEVELOPMENT OF RESEARCH PLAN

The NCHRP 3-54 study was conducted in two distinct

phases. Phase 1 was exploratory in nature and focused on the

development of the study plan to be carried out in Phase 2.

Phase 1 study activities included the following:

•

Review relevant literature,

•

Deﬁne appropriate study factors,

•

Identify and recommend study approach, and

•

Document results.

The research team developed the Phase 2 work plan using the

systems engineering approach. The work plan started with a

study of all known PPLT displays and would methodically

reduce the number PPLT displays being studied to a select few

that would receive more concentrated study. The basic ele-

ments of the proposed work plan were (1) a survey of current

PPLT use across the United States and at several international

locations, (2) a study of driver understanding of known PPLT

displays, (3) the use of computer simulation on a select num-

ber of PPLT displays, and (4) ﬁeld studies. Figure 1-3 depicts

how the work plan in Phase 2 would systematically narrow

the number of displays being studied to a point where the

“best” display would be identiﬁed.

Building on the systems engineering approach, a revised

Phase 2 work plan was developed as shown in Figure 1-4.

The project’s initial study activities were designed to help

reﬁne the number of displays studied in the following study

tasks. Subsequent activities focused on a select number of

displays that were studied with laboratory and ﬁeld techniques

that conﬁrmed earlier ﬁndings and answered questions that

would ultimately lead to recommendations.

Throughout the study, the research team solicited advice

and comments from a team of technical advisors. These advi-

sors were selected from the traffic engineering community

because of their hands-on experience with various applica-

tions of PPLT control and unique permissive displays.

Because many questions of a practical application could

not be answered with a laboratory test, the research team

designed an Engineering Assessment study task. This Engi-

neering Assessment task augmented the findings of other

study tasks, including many ﬁeld implementation issues. The

Engineering Assessment task started in the initial stages of

the project and continued through to the ﬁnal stages of the

project. In its ﬁnal form, the NCHRP 3-54 study consisted of

individual study tasks as identiﬁed in Figure 1-5. In October

of 1999, the research team and project panel met to review

the study PPLT displays and to determine which displays

showed the most potential. As part of the meeting, several

decisions were made, including the decision to reduce the

number of displays to those with the most potential for fur-

ther study. Additionally, the project panel directed the

research team to conduct a thorough evaluation of one par-

ticular display through ﬁeld implementation.

REVIEW OF KEY WORK PLAN ACTIVITIES

As previously stated, the NCHRP 3-54 study was structured

into individual study tasks that consisted of data collection

activities, reporting, and meetings with the project panel. Key

individual study activities are discussed in Chapter 3 of this

report. The greatest amount of detail related to each of the key

study activities is found in the project working papers, which

are appendixes to this report and are provided on the accom-

panying CD-ROM. Below are brief descriptions of nine key

study activities; the other ﬁve tasks were panel meetings and

documentation.

Identify Technical Advisors

Before beginning the formal study effort, potential mem-

bers of a Technical Advisory Group were sought. The Tech-

nical Advisory Group was to consist of at least ﬁve knowl-

edgeable professionals who would provide the project panel

with ﬁrst-hand experience and expertise in the evaluation of

PPLT signal displays. Members of the advisory group were

to be well-respected, experienced, traffic engineers knowl-

edgeable about the PPLT issue. The research team also had

Uniform

Display

Field Studies

Computer

Simulation

Focus Groups and

Expert Panels

Study of Driver Understanding

drive through, conflict

analysis, and accident studies

Figure 1-3. Conceptual work ﬂow of the original Phase 2

work plan.

regular interaction with the Signals Technical Committee of

the NCUTCD in order to provide a regular ﬂow of informa-

tion to the full committee.

Agency Survey

To assess the current state of the practice, the research team

conducted an extensive literature review and administered a

survey of transportation agencies. The literature review iden-

tiﬁed the current state of the art in PPLT displays and included

published and unpublished literature (including material from

international sources).

The agency survey solicited information on the number and

type of left-turn traffic control devices in use across the United

States. The objective of the agency survey was to identify and

quantify the different types of PPLT displays in use. The sur-

vey was administered to transportation professionals at the

state and city levels who were directly involved with PPLT

design and installations. Data on PPLT installations in all 50

states and parts of Canada were obtained.

Photographic Driver Studies

Photographic driver studies evaluated drivers’ under-

standing of the circular green, ﬂashing yellow arrow, ﬂash-

ing circular yellow, ﬂashing red arrow, and ﬂashing circular

red permissive displays in association with different arrange-

ments, placements, and traffic/lane conﬁgurations. Computer-

generated PPLT signal displays on static background pho-

tographs were used to represent the dynamic aspects of the

Assessment

Engineering

Analysis

Crash Data

Driver

Observation

Studies

Draft Video

Driver

Surveys

Agency

Driver

Studies

Survey

Photographic

Identify

Technical

Advisors

Provide Input

1st Interim

Report

1st Panel

and

Meeting

Driver

Video

Survey

Identify Displays

for Further

Study

NCUTCD

Recommendations to

Submit

Final

Report

Recommendations

Refine

Meeting

2nd Panel

and

Report

2nd Interim

Recommendations

Findings and

Figure 1-4. Proposed Phase 2 work plan.

Assessment

Engineering

Analysis

Crash Data

Studies

Field Traffic

Conflict

Studies

Operations

Field Traffic

Agency

Driver

Studies

Survey

Photographic

Identify

Technical

Advisors

Provide Input

Shading indicates

modification to work plan

(refer to Figure 1-4

for comparison)

Provide Input

1st Interim

Report

1st Panel

and

Meeting

Refine

Techniques

Research

Identify Displays

for Further

Study

NCUTCD

Recommendations to

Submit

Final

Report

Recommendations

Refine

Meeting

2nd Panel

and

Report

2nd Interim

Studies

Driver

Confirmation

Recommendations

Findings and

Studies

Field

Implementation

Figure 1-5. Final Phase 2 work plan.

PPLT displays. Over 300 drivers participated in each of eight

geographic locations around the United States for a total of

more than 2,400 drivers. All study participants were licensed

drivers.

Field Traffic Operations Studies

Another major focus involved completion of (1) several

traffic operations studies to determine the effects of the PPLT

signal display types and (2) a traffic conﬂict study to deter-

mine the safety effects of each of the PPLT displays. The

observation studies included quantifying the capacity and

delay associated with various PPLT displays by analyzing

saturation ﬂow rates, lost times, response times, and follow-

up headways. Field traffic conﬂict studies focused solely on

the left-turn movement.

The traffic observation studies were conducted in the same

eight geographic locations as the photographic driver studies

in order to provide additional insights into driver behavior

related to alternative PPLT displays. As part of this effort, the

research team worked with local traffic engineers to identify

representative sites with PPLT displays and to then gather

crash, traffic, geometry, and other relevant data for each site.

More than 8 hours of conﬂict data per intersection were col-

lected at three intersections within each geographic study

area for a total of 192 hours.

Crash Data Analysis

The research team conducted a review of crash data to

determine and compare left-turn crash rates associated with

various PPLT displays. In addition, selected components of

a crash database created in 1988 as part of a FHWA study

were examined.

Engineering Assessment

The Engineering Assessment sought to identify the objec-

tive and subjective information needed to evaluate the pro-

posed displays. To assist in the assessment, an evaluation

matrix was developed. The evaluation matrix included con-

siderations addressing safety, operations, implementation,

human factors, and versatility and was updated as each study

activity was completed.

Data Analysis and Report of Preliminary

Findings

Using the data collected from the agency evaluation, pho-

tographic driver studies, conﬂict studies, and crash studies,

the research team prepared a comprehensive analysis of the

results obtained to that point in the research project. This

information was presented to the project panel, which in turn

reviewed the ﬁndings and ultimately reduced the number of

PPLT displays for future consideration. In addition to select-

ing a subset of promising PPLT displays to be further stud-

ied, the panel provided direction to complete the conﬁrma-

tion study and to conduct a ﬁeld implementation study.

Driver Conﬁrmation Studies

Using full-scale dynamic driving simulators located at the

University of Massachusetts (UMass) and the Texas Trans-

portation Institute (TTI), the research team developed a vir-

tual driving environment to further test driver understanding

of select PPLT displays and associated behavior. Study par-

ticipants were required to navigate a virtual world containing

various signal displays, different arrangements and locations,

and opposing traffic, with the intent of evaluating driver per-

formance at pre-selected intersections controlled by PPLT

displays in an actual driving environment. The researchers at

UMass and TTI were tasked with each testing 200 drivers of

various demographic backgrounds for a total of 400 drivers.

Field Implementation Studies

Field implementation of an experimental ﬂashing yellow

arrow display was conducted to ﬁeld test a display that had

shown promise in safety and driver comprehension in previous

task activities and had ranked high in the Engineering Assess-

ment. Volunteer agencies were sought from across the United

States to install and operate the experimental ﬂashing yellow

arrow display. In conjunction with use of the experimental

display (which required permission from FHWA to operate),

“before” and “after” studies were completed at each inter-

section where the ﬂashing yellow arrow display was installed

and at nearby control sites. These “before” and “after” stud-

ies allowed the research team to quantify the impact of the

changeover from the MUCTD circular green indication to

the ﬂashing yellow arrow indication.

REPORT ORGANIZATION

Chapter 2 of this report provides background material

and sets the stage for discussion of the study methodology,

findings, and recommendations as presented throughout the

remainder of the report. Chapter 2 presents a detailed descrip-

tion of PPLT control, key issues related to this type of signal

phasing, and a description of current practices to mitigate

some of the known disadvantages of using PPLT.

Chapter 3 summarizes the results of the individual project

work elements, including the ﬁndings derived from each of

the major project activities and the implications those ﬁndings

had on the direction of the project. Chapter 4 provides a more

in-depth discussion of ﬁndings surrounding one of the exper-

imental PPLT displays, the ﬂashing yellow arrow. Chapter 5

presents the ﬁnal recommendations derived from the research

project.

CHAPTER 2

BACKGROUND

This chapter provides a general overview of left-turn con-

trol, advantages and disadvantages of PPLT signal phasing,

current standards for PPLT control, and some of the previ-

ously documented variations in PPLT displays used around

the world. Much of the material presented herein was gath-

ered through a review of published literature and current

industry practice.

For decades, traffic engineers have relied on the MUTCD

(1) to provide guidance regarding the installation and opera-

tion of all types of traffic control devices, including left-turn

signal display and phasing. To be effective, the MUTCD must

be specific enough to ensure uniformity, while allowing lat-

itude to adapt the traffic control device to speciﬁc needs. For

left-turn control, the MUTCD addresses the design and appli-

cation of traffic control signs, pavement markings, traffic

signal installations, and traffic islands. The MUTCD identi-

ﬁes several possible combinations of left-turn and through-

movement signal lens arrangements (or displays) and provides

some general guidelines for locating signal heads and advi-

sory signing. The choice of left-turn control depends on sev-

eral factors that must be evaluated by the traffic engineer using

design guidelines and engineering judgment.

MODES OF LEFT-TURN CONTROL

The MUTCD deﬁnes four modes of left-turn control: per-

missive, protected, protected/permissive, and variable left-

turn mode as described below.

Permissive left-turn control typically is used at locations

without left-turn signals. Under permissive operation, the

MUTCD does not require an exclusive signal indication or

signal face for left turns. Consequently, one signal display

can be used for all traffic movements on a single approach

and the circular green indication permits left turns to be made

after drivers yield to oncoming traffic and pedestrians.

Protected left-turn control is used where there is an exclu-

sive display for left-turn movements. With this type of traf-

ﬁc control, left turns may be made only when a green arrow

indication is displayed.

Under PPLT control, left-turning traffic is protected from

oncoming traffic during the protected interval, during which

the green arrow indication is displayed. In another part of the

cycle, during which the circular green indication is typically

displayed, left-turn movements may be made after drivers

yield to oncoming traffic and pedestrians.

Variable left-turn mode describes a situation in which the

operating mode changes among the protected-only mode, the

permissive-only mode, and/or the protected/permissive mode

during different periods of the day.

Combinations of signal arrangements used with the vari-

ous left-turn controls cited above (as deﬁned in the MUTCD)

are illustrated in Figure 2-1. The application of these displays

depends on the availability of an exclusive left-turn lane, the

traffic signal phasing, and the mode of signal operation.

CURRENT STANDARDS FOR

PROTECTED/PERMISSIVE CONTROL

The MUTCD is the mandated source for determining traf-

ﬁc control devices for left-turn maneuvers (1). It is not a legal

requirement to install any device identiﬁed in the MUTCD;

however, if a traffic control device is installed, it must com-

ply with the provisions of the MUTCD subject to mandatory,

advisory, or permissive requirements.

The PPLT mode of operation, as currently deﬁned in the

MUTCD, has a protected, left-turn interval indicated by a

green arrow during part of the signal cycle and a permissive

left-turn interval indicated by a circular green indication dur-

ing another part of the cycle where the left turn must yield to

opposing traffic.

To ensure that these basic requirements are met, the

MUTCD identiﬁes ﬁve considerations for the employment

of traffic signal displays and other traffic control devices:

placement, operation, design, maintenance, and uniformity.

These considerations are discussed below.

Signal Display Placement

To understand the placement needs of a PPLT display, it

is ﬁrst important to understand how general traffic signal

placement is governed and how left-turn signal displays are

located for permissive mode only and protected mode only.

General Traffic Signal Display Placement Criteria

A traffic signal display should be placed to ensure that it is

within the driver’s cone of vision so that it will unmistakably

command attention. The display should also be positioned in

relation to the point, object, or situation to which it applies to

help convey the proper meaning. In addition to being suitably

legible, the display must be located so that a driver traveling

at normal speed has adequate time to observe the display,

comprehend its meaning, and make the proper response.

As illustrated in Figure 2-2, the MUTCD indicates that the

traffic signal display should be located not less than 40 ft nor

more than 150 ft beyond the stop line. The signal “shall be

located between two lines intersecting with the center of the

approach lanes at the stop line, one making an angle of approx-

imately 20 degrees to the right of the approach extended and

the other making an angle of approximately 20 degrees to the

left of the center of the approach extended” (1). Previous

research has suggested that this 20-deg “cone of vision” be

reduced to 10 deg to improve conspicuity (4).

Beyond the basic horizontal signal face location require-

ments shown in Figure 2-2, speciﬁc placement criteria are

identiﬁed in the MUTCD; these criteria depend on the type

of left-turn control. These requirements are described below.

b. c. d.

f. g. h.

i. j. k. l.

m. n. o. p.

q. r. s. t. u.

Figure 2-1. MUTCD arrangements of traffic signal displays.

12 m

(40 ft)

Location of signal heads within these areas:

200 mm (8 in) or 300 mm (12 in)

signal lenses

300 mm (12 in) signal lenses, unless

a near-side signal face is used

Minimum distance of signal faces from stop line.*

***

Maximum distance from stop line for 200 mm (8 in)

signal faces, unless a near-side signal face is used.

Maximum distance from stop line for 200 mm (8 in)

signal faces when near-side supplemental signal

face is used, and maximum distance from stop line

for 300 mm (12 in) signal faces, unless a near-side

signal face is used.

35 m

(120 ft)

45 m

(150 ft)

3m (10 ft)

CENTER

OF APPROACH

X/2

20 20

Figure 2-2. MUTCD criteria for horizontal location of signal faces.

Left-Turn Signal Display Placement Criteria

for Permissive Mode Only

The MUTCD does not require an exclusive signal display

for the left-turn movement if the left turn is going to be made

in the permissive mode only. In this mode, the left-turning

vehicle is directed by the through-traffic signal indication (a

circular green) that is terminated with a circular yellow fol-

lowed by a circular red.

Left-Turn Signal Display Placement Criteria

for Protected Mode Only

The protected-only left-turn signal phase requires an exclu-

sive signal face to control the left-turn movement. This sig-

nal face is normally located in line with the center of the left-

turn lane, either overhead on the far side of the intersection

or ground mounted in the median. The MUTCD does not

require dual signal indications and 8 ft of horizontal separa-

tion between signal faces for protected-only left-turn signal

indications and the adjacent signal indications; however, the

MUTCD does require that a signal face mounted on a span

wire or mast arm be located as near as practical to the driver’s

normal view line (1).

Left-Turn Signal Display Placement Criteria

for PPLT Mode

Because a PPLT display controls both the permissive left-

turn movement and the protected left-turn movement, there

is ﬂexibility in the location of the display. Figure 2-3 illus-

trates two potential overhead signal display placement alter-

natives allowed by the MUTCD for PPLT control. In addi-

tion to the alternatives shown in Figure 2-3, post-mounted

median and farside display placements are also used. The

placement of a given PPLT display ultimately depends on

the type of display used. In some instances, PPLT control is

implemented without the beneﬁt of having an exclusive left-

turn lane.

Operation

The traffic signal display should be operated in conjunc-

tion with the appropriate traffic control devices and equip-

ment to meet traffic requirements at a given location (1). Fur-

ther, “the display must be placed and operated in a uniform

and consistent manner to ensure that drivers can respond

properly to the display, given their previous exposure to sim-

ilar traffic control situations.” (1)

Center

of Lane

Optional

Placement

Lane

Line

Figure 2-3. Overhead PPLT display placement options.

Design

Features such as size, contrast, colors, shape, composition,

and lighting or reﬂectorization should be combined to draw

attention to the display (4). The shape, size, and colors of the

display should produce a clear meaning. Legibility and size

should be combined with placement to permit adequate

response time. The display’s uniformity, size, legibility, and

comprehensibility should command respect from drivers

when the display is encountered.

The PPLT signal display must provide for both protected

left turns and permissive left turns during the signal cycle.

The most commonly used signal displays are the ﬁve-section

cluster, vertical, or horizontal, shown in Figures 2-1 -o, -m,

and -n, respectively. As shown in Figures 1-1 and 2-4, the

signal indications for PPLT mode left turns are provided by

either a shared signal face or an exclusive signal face used

only by left-turn traffic.

Shared PPLT Display Arrangement

In a shared application, the signal face indicating a pro-

tected left-turn movement is one of the two required signal

faces for the approach and is usually not directly over the

left-turn lane. It displays a left-turn green arrow signal indi-

cation and the adjacent through movement indication (circu-

lar red or circular green) simultaneously. The MUTCD fur-

ther requires that the signal faces for the through traffic on

the opposing approach simultaneously display circular red

signal indications. During the permissive left-turn movement,

all signal faces on the approach display circular green signal

indications. At any point in the signal cycle, all signal faces

on the approach are required to simultaneously display the

same color of circular indications to both through and left-

turn road users (variations are allowed when louvered signals

are used as explained later in this chapter) (1).

Exclusive PPLT Display Arrangement

In applications using an exclusive signal display arrange-

ment for the left-turn movement, a separate signal face, usu-

ally located directly over the left-turn lane, is provided in addi-

tion to the minimum two required signal faces for the through

movements. The separate face is required to simultaneously

display a left-turn green arrow signal indication and a circular

red signal indication for the protected phase of the signal cycle.

A circular green indication is displayed for the permissive

interval. The MUTCD further requires that the signal faces for

the through traffic on the opposing approach simultaneously

display circular red signal indications. During the permissive

left-turn movement, the left-turn signal face displays a circu-

lar green signal indication (1).

Maintenance

Traffic signal displays should be well maintained to ensure

that legibility is retained and the display is visible. The display

should be removed if it is no longer needed (1). Clean, legible,

and properly mounted displays, in good working condition,

command the respect of every travel mode at the intersection.

In addition to regular maintenance, traffic displays should be

adjusted regularly to address current conditions. The fact that

a display is in good working order should not be a basis for

deferring needed replacement or change. Conversely, poor

maintenance can destroy the value of a group of traffic signal

displays by minimizing the respect commanded by individual

displays.

Uniformity

Uniformity means treating similar traffic situations in the

same way. Uniform traffic signal displays aid the road user by

simplifying recognition and understanding (1). Uniform dis-

plays also help road users, police and enforcement personnel,

and traffic courts interpret appropriate driver behavior. Uni-

form traffic control displays are also economical because con-

sistent manufacturing, installation, maintenance, and admin-

istration processes can be used.

Signal Phasing

Various signal phase sequences can be used for the PPLT

mode such as leading or lagging protected turn intervals, with

adjacent through traffic either moving concurrently with the

left turn or stopped. The MUTCD requires that all same-color

circular indications in all signal faces on an approach be simul-

taneously illuminated when PPLT operation is used with one

exception: when using an exclusive left-turn signal face with

circular green and circular yellow signal indications that are

visibility limited from the adjacent through movement, the

left-turn signal is not required to simultaneously display the

same color of circular indication as the faces for the adjacent

through movement. Further, in this visibility limited arrange-

ment, a circular green signal indication for the permissive left-

turn movement can be displayed while the signal faces for the

adjacent through movement display a circular red indication

and the opposing left-turn displays a left-turn green arrow indi-

cation for a protected left-turn movement (1). This arrange-

ment, commonly referred to as the Dallas Display (or Dallas

Phasing), is used to eliminate the previously acknowledged

yellow trap. Figure 2-5 illustrates a typical Dallas Display.

In situations where an exclusive left-turn signal face is

provided and the left-turn signal face does not simultane-

ously display the same color circular indication as the adja-

cent through movement, the MUTCD requires that a combi-

nation of a LEFT TURN SIGNAL sign (R10-11) and a LEFT

MUTCD

4-Section Horizontal

MUTCD

5-Section Horizontal

MUTCD

5-Section Vertical

Used in Texas, Nebraska,

and others

Used in Texas, Nebraska,

and others

Used in Texas and

most Western States

Area

Used

Lens Color

and

Arrangement

Left-Turn Indication

Protected

Mode

Permitted

Mode

Variation

of 5-Section Cluster

MUTCD

Typical

5-Section Cluste

Figure 2-4. Variations in PPLT displays.

ing circular red indication, a circular yellow or yellow arrow

indication, and a green arrow indication for left-turn move-

ments from the top of the “T” to the stem of the “T.” Because

many of these intersections are freeway on-ramps, the leg on

the top of the “T” in the same direction as the left turn may

not have signal indications. The signal rests in ﬂashing red

for the left-turn driver with green for opposing traffic. After

the left-turning vehicle occupies the left-turn bay for a set

period, the driver receives a protected phase. Michigan uses

PPLT phasing with ﬂashing circular red indications. The left-

turn lane has an exclusive three-section display consisting of

a circular red indication, a circular yellow indication, and a

green arrow indication for left-turn movements. In some cases,

a yellow arrow is used in place of the circular yellow indica-

tion. The left-turn movements are operated in a PPLT (dual

lagging) mode. The circular red indication is ﬂashed during

the permissive interval. A protected left-turn green arrow indi-

cation is provided only if left-turn demand exists at the end

of the permissive phase.

The ﬂashing permissive circular red operation has been

used in Michigan since about 1975. One of the earliest instal-

lations of the ﬂashing red is believed to have been in Ann

Arbor, Michigan. The Michigan Department of Transporta-

tion (MDOT) estimated that the State of Michigan operates

the ﬂashing circular red indication at 100 locations, mostly

in urban areas. Another 200 installations are operated locally

by Wayne and Oakland Counties. In the greater Grand Rapids

area, 72 additional locations use this PPLT indication.

Delaware uses the ﬂashing red arrow indication in a four-

section, left-turn display. The display consists of a red arrow

indication next to a circular red indication, with a yellow

arrow indication and a green arrow indication centered under

the red indications. The permissive interval is indicated by a

ﬂashing red arrow.

The Delaware DOT estimated that approximately 100 loca-

tions in Delaware have the ﬂashing red arrow permissive

phase, with the ﬁrst installations dating to the early 1980s.

Most locations operate with a leading protected phase. After

the red arrow ﬂashing permissive interval, a solid circular red

indication is displayed to the left-turning vehicle, rather than

again displaying a yellow arrow indication (a change inter-

val is not provided). Delaware has also developed controller

TURN YIELD ON GREEN (symbolic circular green indica-

tion) sign (R10-12) be used (1).

In some applications, use of protected, permissive, or PPLT

operations at a given location may be changed by time of day

to reﬂect changes in the traffic conditions. In these instances,

in addition to meeting the previously documented criteria, the

MUTCD stipulates that (1) the circular green and circular yel-

low signal indications shall not be displayed when operating

in protected-only mode and (2) the left-turn green arrow and

left-turn yellow arrow signal indications shall not be displayed

when operating in the permissive-only mode. Although no

speciﬁc signing for a time of day application is identiﬁed, the

MUTCD notes that “additional appropriate signal indications

or changeable message signs may be used to meet the require-

ments for the variable left-turn mode” (1).

ALTERNATIVE DISPLAYS USED

IN THE UNITED STATES

Several innovative displays and phasing arrangements have

been created within the past 20 years. Five variations of the

display indicating the permissive phase are known to exist,

only one of which includes the use of the MUTCD standard

circular green indication. These include use of the ﬂashing cir-

cular red, the flashing red arrow, the flashing circular yel-

low, and the ﬂashing yellow arrow indications. Figures 1-1 and

2-4 illustrate the various PPLT displays used throughout the

United States. There have also been some innovative advances

in signal phase sequence, such as the “Dallas Display,” which

alleviate the yellow trap. The manner of use and frequency

of occurrence for each type of display are discussed in the

following subsections and reﬂect data collected in 1998.

Flashing Red Display

Flashing red displays are used in Maryland (ﬂashing cir-

cular red indication), Michigan (ﬂashing circular red indica-

tion), Delaware (ﬂashing red arrow indication), and Califor-

nia (ﬂashing red arrow indication). The ﬂashing circular red

indication as currently used in Maryland is primarily applied

at two-phase “T” intersections (approximately 13 locations).

This display consists of a three-section display with a ﬂash-

5-SECTION “DALLAS DISPLAY”

PPLT with “Dallas Phasing”

eliminating yellow trap

ADJACENT THRU

HEAD

Louvered

Signal

Lens

Figure 2-5. Illustration of Dallas display.

logic to omit the left-turn phase call until the opposing queue

is dissipated.

The City of Cupertino, California, has installed a ﬂashing

red arrow permissive indication in at least three locations.

The City uses a vertical, four-section display with a circular

red indication, a ﬂashing red arrow indication, a yellow arrow

indication, and a green arrow indication. Typically, one left-

turn display is median mounted with a second left-turn dis-

play post mounted on the far left side of the intersection.

Flashing Yellow Display

The City of Seattle, Washington, uses the ﬂashing circular

yellow indication to communicate the permissive left-turn

interval at approximately 20 installations, with additional

intersections still being added periodically. Seattle is using a

four-section vertical display that has a circular red indication,

a circular yellow indication, a ﬂashing circular yellow indi-

cation, and a dual indication yellow arrow/green arrow sec-

tion. Typically, these locations are low volume and operate

in a PPLT mode throughout the day. The ﬂashing circular

yellow indication has been in use in the City of Seattle since

about 1966 and has also been installed in a few other loca-

tions in the Seattle metro area. The staff at the City of Seat-

tle promotes the use of the ﬂashing circular yellow indication

because this indication provides high contrast during the night-

time hours of operation. It is during the nighttime hours that

the City has identiﬁed a reduction in crash rates over the use

of the circular green display (5). The implementation of the

ﬂashing circular yellow indication was before light-emitting

diode (LED) lens technology met acceptable standards for

widespread deployment.

The City of Reno, Nevada, installed a ﬂashing yellow arrow

permissive display at ﬁve locations around the City. The typ-

ical implementation design used a four-section vertical dis-

play with a red arrow indication, a yellow arrow indication,

a yellow ﬂashing arrow indication, and a green arrow indi-

cation. The display was mounted on a mast arm over the left-

turn lane, with a second left-turn display post-mounted on the

far left side of the intersection. Because of a change in the

City of Reno’s Traffic Engineering staff, the City elected to

remove the ﬂashing yellow arrow displays so that all per-

missive displays were uniform within the city.

In April 1998, the nearby City of Sparks, Nevada, installed

a ﬂashing yellow arrow permissive display at six locations

with approval from FHWA. The City has continued to install

the ﬂashing yellow arrow display at more locations. The exper-

imental design uses an exclusive ﬁve-section cluster display

arrangement, with the ﬂashing yellow arrow in the standard

yellow arrow location. The circular green indication is illu-

minated simultaneously with the ﬂashing yellow arrow indi-

cation during the permissive left-turn interval. By design, the

ﬂashing yellow arrow is a supplemental indication to the cir-

cular green indication; therefore, two indications essentially

convey the same information. To accomplish the ﬂashing yel-

low output, the city installed specially designed logic control

units in the controller cabinet. The display is accompanied

with the supplemental sign R10-1 LEFT TURN YIELD ON

GREEN (circular green indication symbol). The supplemen-

tal sign is consistent with standard practices and consistent

with the MUTCD.

Lead-Lag Displays

A traffic control technique designed to avoid the yellow

trap is known as “Dallas Display” operation. The Dallas Dis-

play, developed in the mid 1980s (1986-1987) by a group of

traffic engineers in Dallas County, Texas, requires the stan-

dard three-section signal display for the through movement

and an exclusive, ﬁve-section left-turn signal display. First

implemented in the cities of Dallas and Richardson, Texas,

the ﬁve-section signal display is typically centered over the

left-turn lane in a vertical, cluster, or, most often, horizontal

conﬁguration. The circular green indication and circular yel-

low indication in the left-turn display are overlapped with

both through movements and are shielded so that they can be

seen only by the left-turning traffic.

An alternative to the Dallas Display was developed in

Arlington, Texas, and is sometimes called the “Arlington Dis-

play.” The Arlington Display uses the same Dallas Display

concept, except that the lagging protected left-turn direction

does not receive a permissive interval during the leading direc-

tion protected interval. This permissive interval is excluded for

two reasons. First, the permissive direction is usually facing

an opposing through queue that is just starting to dissipate,

making it unlikely that the permissive left-turn maneuver can

be made safely. Second, some practitioners believe that the

display of the permissive interval without the same-direction

through traffic receiving a circular green indication is con-

fusing to drivers.

The Dallas or Arlington PPLT displays are used at more

than 600 locations in Texas. The Dallas Display operation has

recently been installed at several locations in Las Vegas and

Carson City, Nevada. Las Vegas has programmed several other

locations for conversion to PPLT phasing in the near future.

Most recently, the cities of Los Angeles and Upland, Califor-

nia, have installed the Dallas Display at several intersections.

As noted within the previous discussion of display applica-

tions, the 2000 MUTCD allows for the use of visibility-limited,

left-turn signal faces that implement the Dallas Display.

PRACTICES OUTSIDE THE UNITED STATES

The United States is not the only country trying to increase

drivers’ understanding of the permissive left-turn display.

Practices from other countries that were identiﬁed for this

report are discussed below.

Canada

Within Canada, there has been a concerted effort among

many of the provinces to gain uniformity in the displays used

for the protected, permissive, and PPLT control. Many areas

of Canada use a ﬂashing indication, but for different reasons

than the United States. For example, the provinces of British

Columbia, Alberta, and Saskatchewan use the ﬂashing green

left-turn arrow indication for the protected left-turn phase,

whether it is protected only or protected/permissive. Ontario

communicates a protected left-turn movement using the ﬂash-

ing circular green indication at some intersections and a solid

green arrow indication (same as the United States) at other

locations.

In 1997, the Transportation Association of Canada adopted

new traffic signal standards aimed at increasing driver safety

and signal operating efficiency. The most signiﬁcant changes

involve the introduction of one arrow display (ﬂashing green

arrow indication) for all protected left- and right-turn signal

displays, removal of the nonstandard ﬂashing circular green

indication for protected turn indications, introduction of the

steady amber arrow change interval, provision of a minimum

of two traffic signal heads for through and left-turn move-

ments, and provision of a ﬂashing DON’T WALK change

interval for pedestrian displays. The steady circular green

indication is used to display the permissive interval.

Based on telephone interviews conducted by the research

team in 1995 and a 2002 follow-up conversation, support for

the ﬂashing display centered on the following issues: (1) the

ﬂashing indication provides a more visible message to the

driver; (2) the ﬂashing indication assists drivers with color

vision problems (color anomalous/color deﬁcient); and (3) the

indication has been associated with increased saturation ﬂow

rates for the left-turn movement. Proponents of the ﬂashing

display also acknowledge that there are disadvantages,

including the following: (1) the display is not consistent with

the steady green arrow indication currently used in many

areas throughout Canada and the United States and (2) there

is no uniform meaning for a ﬂashing indication.

To support the hypothesis that the ﬂashing display increases

saturation ﬂow, there were limited studies of the left-turn sat-

uration ﬂow for the ﬂashing display versus the non-ﬂashing

display (6). In a 1991 study, the saturation ﬂow for the two

displays was studied in Edmonton, Alberta, and Victoria,

British Columbia. The data conﬁrmed the hypothesis that the

ﬂashing display increases the saturation ﬂow. It is believed

that the ﬂashing indications demand more attention from

drivers. It should be noted that the NCHRP 3-54 study also

studied saturation ﬂow rate for different signal displays and

indications and did not reach similar ﬁndings.

In many provinces, left-turn indications are being displayed

through a variable (ﬁber-optic or LED) lens. The green arrow

indication will terminate to the yellow change interval using the

same lens. In response to concern that drivers with color vision

problems would have difficulty in distinguishing between the

two indications, the ﬂashing indication was proposed. The

Canadian Association of Optometrists supports use of a ﬂash-

ing indication (7). It was believed that a ﬂashing green dis-

play would allow all drivers, regardless of their vision con-

ditions, to better discriminate between the different displays.

There is some resistance in Canada to the use of the ﬂashing

protected left-turn indication. For example, Ontario already

uses a steady green for simultaneous leading PPLT phasing.

Ontario has thousands of signalized left turns and the cost to

convert has been estimated to be high.

Standard speciﬁcations were approved by the Council for

Traffic Control Signals in 1997. Part B of the Manual of Uni-

form Traffic Control Devices for Canada was subsequently

completed and adoption has been underway across the coun-

try. The British Columbia Ministry of Transportation standard

for protected only left-turn movements was a display consist-

ing of a three-section head (steady green arrow, steady yellow

arrow, and a solid circular red indication). The Ministry’s stan-

dard for PPLT displays incorporated a ﬂashing green arrow

and then a steady yellow arrow, resulting in a four-section

signal. The administrative staff of Highway Safety at the

British Columbia Ministry of Transportation estimates imple-

mentation of the new standards should be completed within

10 years. It has been estimated that the cost of implementa-

tion is $35 million (1992 Canadian dollars) (8).

Europe

Several European countries are experimenting with the

use of the ﬂashing yellow indication for the permissive left

turn. In Heidelberg, Germany, the ﬂashing yellow arrow indi-

cation is used on the Neckarstaden at the Congress House

and Stadhalle for the permissive interval of PPLT phasing.

The ﬂashing yellow arrow indication is also used for the per-

missive interval in Bern, Switzerland. In Strasbourg, France,

the ﬂashing yellow indication is used for right-turning traffic

to indicate “yield to pedestrian traffic.” In Spain, there is some

use of the ﬂashing yellow arrow indication for the permissive

left-turn interval.

There has been a concerted effort by practicing European

traffic engineers to develop a European standard for the use

of the ﬂashing yellow arrow indication, though efforts to

implement nationwide across Switzerland were unsuccessful

because of concerns expressed by the Swiss federal police

department.

A European research study concluded that (1) turning vehi-

cles cause accidents, because they periodically misinterpret

the prevailing signals’ “full green” and “green” indication with

ﬂashing light (In Switzerland, the Ordinance for Road Signals

stipulates that a ﬂashing warning signal be positioned beside

the green signal to caution drivers about on-coming vehicles.)

as a “go ahead”; and (2) the “flashing yellow arrow” proves

to be a simpler and more uniform signal—drivers of turning

vehicles understand it better and at least some accidents are

prevented (9, 10).

A more in-depth study of the ﬂashing yellow arrow indica-

tion in 1990 compared the accident statistics at 35 signal instal-

lations with ﬂashing yellow arrows in Zurich, Switzerland, and

St. Gallen, Germany, with a control study of 22 intersections

without ﬂashing yellow arrows in Zurich, Switzerland, and

Winterthur, Germany (11). Accident data were analyzed for

a 2.5-year period before the ﬂashing yellow arrow indication

was installed and for a 1.5-year period afterwards. Accord-

ing to this study, the traffic signals with the ﬂashing yellow

arrows led to a signiﬁcant accident reduction at the 35 survey

installations.

Australia

Researchers at the University of Adelaide, Australia, con-

ducted a three-part study of six traffic signal displays, includ-

ing a ﬂashing yellow arrow indication. Their efforts focused

on analysis of crash data at intersections with the study dis-

plays, a driver survey, and a reaction time experiment com-

paring speed and accuracy of responses to computer-based

animation of the traffic light displays. Overall, the study deter-

mined that there were no conclusive ﬁndings demonstrating

superiority of the ﬂashing indication as compared with other

traditional display applications. The research report conclu-

sion section identiﬁes that the ﬂashing yellow arrow indica-

tion shows promise and offers a recommendation that the

use of a ﬂashing yellow arrow display be further investigated,

potentially through dialog with researchers from the NCHRP

3-54 project being conducted in the United States (12).

THE LEFT-TURN PROBLEM

As evidenced by the multiple phasing and display options

identiﬁed in the previous sections, accommodating left-

turning vehicles at signalized intersections has been an ongo-

ing concern for transportation engineers as they seek a balance

between intersection capacity and safety through signal phas-

ing techniques. After deciding that a left-turn phase is required,

one of the major decisions faced in timing traffic signals is to

appropriately assign time for left-turn movements. As the

number of left-turning vehicles increases, average delay and

accident potential for both through and left-turning vehicles

also increases. Exclusive left-turn lanes and protected left-

turn phases are commonly used to minimize the impact of

left-turning vehicles. When a protected left-turn phase is used,

however, the time to provide that phase must be taken from

the through phases, or the cycle length must be extended.

Other decisions the engineer must make concern the type of

left-turn phasing that best satisﬁes the left-turn demand and

the left-turn phase sequence that maximizes progression, par-

ticularly if the intersection is located on an arterial street.

ADVANTAGES OF PROTECTED/PERMISSIVE

LEFT-TURN PHASING

Transportation professionals have chosen PPLT phasing

for many reasons, including minimizing delay, improving

progression, and reducing fuel consumption and air pollu-

tion. From an operational standpoint, an agency might con-

sider using PPLT phasing to increase the operational effi-

ciency of an intersection. PPLT phasing can reduce delay for

left-turning vehicles under low-to-moderate traffic volume

conditions. PPLT phasing is especially effective in reducing

left-turn vehicle delay when it is operated with a coordinated

signal system.

Many jurisdictions use the lead-lag left-turn phase sequence

at intersections within a signal system to improve progres-

sion. The beneﬁts of the lead-lag left-turn phasing are further

enhanced with protected/permissive lead-lag phasing. By

allowing vehicles to turn left during the permissive interval,

required left-turn green phase time can be reduced, allowing

more green time for the coordinated movements. This tech-

nique is especially effective for coordinated arterial signals

where the progressed platoons in each direction do not pass

through the signal at exactly the same time. Several studies

(3, 13, 14) on lead-lag PPLT operation have found intersec-

tion delay and crashes are reduced over traditional PPLT

operation. A comprehensive evaluation of the impacts of

PPLT phasing associated with coordinated signal timing was

completed in western San Bernardino County, California (15).

The City of Upland, California, where much of this study

was completed, was using the Dallas Display. The researcher

documented a 30 to 50% reduction in vehicle delay when

comparing protected-only to PPLT phasing (15). PPLT phas-

ing can also improve the air quality. The California study

showed that increasing average speed, reducing overall travel

time, and reducing the number of stops resulted in a signiﬁcant

reduction in mobile source emissions. The study documented

a reduction in reactive organic compounds (ROC), carbon

monoxide (CO), and nitrogen oxide (NO

) emissions by 9 to

12% per day when comparing protected (lead-lag) with

protected/permissive (lead-lag) left-turn phasing (Dallas

Display).

DISADVANTAGES OF THE MUTCD CIRCULAR

GREEN DISPLAY

Although there are many reasons to use PPLT control,

some disadvantages have not been resolved. First, many traf-

ﬁc and safety engineers question whether drivers fully under-

stand the meaning of the permissive indication. Their obser-

vations suggest that drivers may be confused about how to

make the left turn safely, particularly when PPLT phasing is

used. They have also observed that some left-turning traffic

at intersections with PPLT control does not use the permis-

sive phase (even when adequate gaps in the opposing traffic

are available) and instead turn left only when given the pro-

tected left-turn arrow indication. In other locations, left-turn

drivers may interpret the circular green as a protected display.

A second disadvantage of PPLT lead-lag phasing is the

“yellow trap.” The yellow trap can occur in a number of other

situations, such as during signal preemption for emergency

vehicles, during signal preemption for railroad grade cross-

ings, during re-service of lead left-turn phases, or during an

overlapping green extension for “slot” (or inside) clearance

at an offset intersection.

SUMMARY

Through a review of published literature and industry

practice, this chapter has provided an overview of the current

use of left-turn control, speciﬁcally focusing on PPLT control.

Basic advantages and disadvantages of PPLT signal phasing

and some of the previously documented variations in PPLT

displays used around the world were identified. Chapter 3

builds on this basic information as detailed elements of the

NCHRP 3-54 project are presented.

CHAPTER 3

RESEARCH ACTIVITIES AND FINDINGS

The research project encompassed 14 individual tasks; 8

study tasks and 6 meeting/report tasks. The eight study tasks

are discussed below:

•

Agency Survey,

•

Photographic Driver Study,

•

Field Traffic Operations Study,

•

Field Traffic Conﬂict Study,

•

Crash Data Analysis,

•

Driver Conﬁrmation Study,

•

Field Implementation Study, and

•

Engineering Assessment.

Each of these work activities is discussed in general detail

below. The work activity objective, methodology, results,

and ﬁndings are identiﬁed. A complete description and ﬁnd-

ings for each of these work activities can be found in the

working papers included in the accompanying CD-ROM.

AGENCY SURVEY

The discussion in Chapter 2 demonstrated that there is a

wide variation in the use and application of the PPLT control

throughout the United States and abroad. To identify and quan-

tify the different types of PPLT control displays adequately,

the research team administered an agency survey. The survey

was mailed to all 50 state traffic engineers for DOTs and to

traffic engineers in 275 of the largest city and county trans-

portation agencies in the United States and Canada.

Objective

The objective of the agency evaluation effort was to iden-

tify and quantify the different PPLT displays, design, and phase

sequencing used in the United States.

Methodology

A survey was developed to collect information about the

use of PPLT signal displays in the United States. The survey

consisted of three sections having a total of 15 questions.

The first section of the survey, General Information,

included two questions to determine how many signalized

intersections were located in the jurisdiction and to what

extent the PPLT signal display was used.

The second section of the survey included nine questions

about the use of the PPLT signal display, including inquiries

about the display arrangement, mounting type and location,

use of secondary PPLT signal displays, and the type of sig-

nal indication used for the permissive phase of the PPLT.

The third section of the survey included questions about

how the agency uses PPLT signal displays for different geo-

metric roadway conditions, if the agency uses any special

technique(s) to avoid the yellow trap, and if there are any local

laws or ordinances governing the use of PPLT signal displays.

The survey responses were summarized and analyzed to

determine to what extent PPLT signal displays are being used

in the United States.

Results

One hundred and eighty (55%) of the surveys were returned,

including six from Canada and six from agencies not using

PPLT signal displays. Excluding the Canadian surveys (Cana-

dian use of PPLT signal displays is discussed in Chapter 2 of

this report) and the jurisdictions not using PPLT displays, the

remaining 168 returned surveys were used to develop the

ﬁndings summarized below.

Findings

•

Collectively, the 168 agencies identiﬁed 107,219 signal-

ized intersections. Of those, 30,870, or 29%, used PPLT

signal phasing.

•

The ﬁve-section cluster display was the predominant

arrangement used in 34 states and represented 63% of all

reported PPLT signal displays. The next most commonly

used PPLT signal display arrangement was the five-

section vertical, which accounted for approximately 19%

of all reported PPLT signal displays.

•

The ﬁve-section horizontal display was used predomi-

nantly in two states and accounted for approximately

9% of the reported PPLT signal displays. Use of the

four-section and three-section PPLT signal displays were

less common. They were found to be in use in a few states

and accounted for less than 10% of all reported PPLT sig-

nal displays.

•

A total of 41% of the agencies reported use of a single

(consistent) PPLT signal display arrangement through-

out their jurisdictions. Most agencies used a combination

of post-mounted ﬁve-section vertical displays in median

applications and ﬁve-section vertical or ﬁve-section clus-

ter displays for mast arm or span wire mounting.

•

Thirty-three agencies identiﬁed a total of 1,650 PPLT

signal displays with bimodal arrow indications. These

agencies explained that the bimodal indications reduced

the space requirement of the signal display (by elimi-

nating one signal section/lens), resulting in energy sav-

ings. The remaining 135 agencies did not use bimodal

indications.

•

Mast arm and span wire mounting was used predomi-

nantly in combination with the cluster PPLT signal dis-

play. Use of pole mountings was primarily reported with

vertical PPLT signal displays for median applications.

•

A total of 52% of the responding agencies mounted the

overhead PPLT signal display on the lane line between

the left-turn lane and the adjacent through lane, while

40% mounted the PPLT signal display over the center

of the left-turn lane.

•

The use of secondary or additional PPLT signal displays

was nearly evenly split—49% of the agencies respond-

ing to the survey used a secondary PPLT signal display

(which was usually pole mounted on the far side of the

intersection), while 51% did not.

•

Of the responding agencies, 40% always used the

PPLT signal display as one of the two MUTCD-required

through movement displays, while 37% sometimes did

and 23% never did.

•

The circular green permissive indication was used by

165 of the 168 agencies for left turns. Other permissive

indications in use included the ﬂashing circular yellow

indication, ﬂashing yellow arrow indication, ﬂashing cir-

cular red indication, and ﬂashing red arrow indication.

•

PPLT signal displays were generally used with exclu-

sive left-turn lanes—89% of left-turn lanes were exclu-

sive, 8% were shared, and 3% were a combination of

exclusive and shared lanes.

•

There was little consistency in the use of supplemental

signs among the responding agencies—49% of the agen-

cies always use signs, 34% use signs only in certain con-

ditions, and 17% do not use signs. When signs are used,

more than 75% of the signs are the MUTCD R10-12 sign.

•

A total of 83% of the signalized intersections using

PPLT signal phasing employed a leading sequence, while

11% used a lagging sequence and 6% used a lead/lag

sequence.

•

A total of 53% of the responding agencies indicated that

they did not use any special phasing or techniques to

avoid the yellow trap, while 5% of the agencies indi-

cated that they use Dallas or Arlington Display, 18% use

exclusive left lead with PPLT lag, and 25% indicated

that they used some other method.

•

All off the responding agencies referred to state statutes

or local ordinances that require either specific display

types, display indications, or compliance with state man-

uals; however, only 7% of the agencies indicated that

there was a law or ordinance that required certain prac-

tices in the use of PPLT phasing.

PHOTOGRAPHIC DRIVER STUDY

This section of the report summarizes a photographic

driver survey that was undertaken as part of the research

study. Several human perceptions affect drivers’ under-

standing of the traffic control signal. The visual search

processes, perception and reaction, and recognition and

comprehension that drivers exhibit can be investigated and

used to evaluate PPLT signal displays. Drivers’ expectation

and the complexity of the traffic control signal also influ-

ence drivers’ understanding and can be incorporated into the

evaluation.

Objective

The objective of the photographic driver study was to eval-

uate the different PPLT signal displays used in the United

States (as identiﬁed through the literature review and agency

survey). The evaluation explored driver understanding of the

signal indication under various conditions. The conditions

varied through the use of protected left-turn indications, per-

missive left-turn indications, through-movement indications,

and PPLT signal display arrangements.

Methodology

A computer-based study tool using both dynamic and

static elements was developed and used to collect data about

drivers’ understanding of the different PPLT signal indica-

tions and displays. To create a reasonable simulation of the

driver’s view, photographs of existing signalized intersec-

tions were incorporated into the computer software. These

photographs acted as the background scenes for different

PPLT signal displays. Over 100 photographs were catego-

rized into three groups based on the mounting type, location

of the PPLT signal display, and intersection geometry. Six

photographs (two from each group) were selected to be used

in the study and were incorporated into the software. Five of

the six selected photos contained a vehicle in the opposing

through lane(s). Because of the static nature of the photo, it

was impossible to determine if the opposing vehicle was sta-

tionary or proceeding through the intersection. The remain-

ing photo, without a vehicle in the opposing through lanes,

was used as a control photo providing a means of analyzing

the effect that the presence of a vehicle in the opposing

through lane had on survey responses. Figures 3-1 through

3-6 illustrate each of the six background photos used in the

evaluation.

Two hundred unique scenarios were produced using the

six selected photographs and combinations of protected left-

turn indications, permissive left-turn indications, through-

movement indications, and PPLT signal display arrangements.

An all-red scenario was also created for each of the PPLT sig-

nal display arrangements. A randomizer function was added to

Figure 3-1. Photographic driver survey—background picture 1.

Figure 3-2. Photographic driver survey—background picture 2.

the software, which allowed a subset of 30 of the 200 sce-

narios to be randomly presented to each driver participating

in the study.

Survey Tool

One of the objectives in the development of the photo-

graphic driver survey was to make it as self-explanatory and

self-administering as possible, requiring little input from the

survey administrator as each driver completed the survey. To

meet this objective, all survey instructions were included as

a sound track within the survey software. Along with the gen-

eral survey instructions, the computer operations necessary

to complete the survey were demonstrated through an exam-

ple survey question.

The use of laptop computers in the photographic driver

study provided the opportunity to include three additional fea-

Figure 3-4. Photographic driver survey—background picture 4.

Figure 3-3. Photographic driver survey—background picture 3.

tures in the software design. First, the computer clock time was

recorded for each survey response, measured from the time the

scenario was presented on the computer screen to the time a

response was selected. Response time data were used as a sur-

rogate measure of driver understanding; longer response times

were correlated to lower levels of driver understanding. Aver-

aging the response times for each scenario minimized bias

related to distractions, guessing, and other time variances not

related to understanding.

The second feature added to the software was an initial

screen that allowed the survey administrator to enter the com-

puter number, the location of the survey, and the number of

scenarios to be randomly selected and presented to each

driver. This screen was only active when the software was

initially loaded at the start of each survey session. The final

feature added to the software was a file writing procedure

for processing the survey data. As each survey response was

recorded, the location, date, computer number, demographic

Figure 3-6. Photographic driver survey—background picture 6.

Figure 3-5. Photographic driver survey—background picture 5.

data, scenario number, response time, and response were

written as a single row of data in a text ﬁle.

With respect to the actual operation of the survey instru-

ment, the driver was asked “If you want to turn left, and you

see the traffic signals shown, you would . . .” as each study

scenario was presented. The driver used the computer key-

board to select what he or she believed to be the appropriate

response. Options were as follows:

•

Go.

•

Yield. Wait for gap.

•

Stop, then wait for gap.

•

Stop.

The driver also could choose not to respond.

In addition, drivers were asked additional demographic

questions, including their sexes, ages, and educational levels;

whether they live in a city, suburb, or rural area; and how many

miles they drove the previous year.

Survey Locations

The study was administered to respondents in Dallas, Texas;

Dover, Delaware; Oakland County, Michigan; College Sta-

tion, Texas; Seattle, Washington; Portland, Oregon; Cuper-

tino, California; and Orlando, Florida. Through the study

methodology, a sample size of 2,400 participants was sought.

The study was conducted primarily at local drivers licensing

facilities. Photographs of the driver study being administered

at various locations are provided in Figures 3-7 and 3-8. Data

were compiled into a single spreadsheet database and were

then analyzed to determine the drivers’ understanding of the

different PPLT signal displays and indications, using both

descriptive and statistical procedures.

Results

A total of 2,465 drivers participated in the study, exceed-

ing the target of 2,400 participants. At least 300 drivers com-

pleted the study at all but one of the eight locations. Because

each study respondent was presented with 30 scenarios, a

total of 73,950 responses were recorded.

Of the 2,465 drivers, 58% were male, 41% were female,

and the balance (1%) did not respond to the gender question.

A total of 27% were less than 24 years of age, 44% were

between 25 and 44, 21% were between 45 and 65, and 7%

were over 65. The remaining respondents (1%) did not pro-

vide their ages. In total, 58% of the participating drivers lived

in a city, 30% in a suburban location, 11% in a rural location,

and 1% did not respond. In total, 5% of the respondents indi-

cated that they did not drive at all last year; 31% drove fewer

than 10,000 miles; 44% drove between 10,000 and 20,000

miles; 19% drove more than 20,000 miles; and 1% did not

respond. Education levels among the participants were rela-

tively uniform—29% of the drivers had a high school or

equivalent education, 35% had some college education, and

35% had a college degree, with 1% not responding.

Figure 3-7. Example of driver evaluation study workstation at Department of

Motor Vehicles.

Findings

Analyses were performed to evaluate the drivers’ under-

standing of three signal indications: all-red, protected, and

permissive. (For further information on these analyses, con-

sult Working Paper No. 3, which is included in the appen-

dixes on the accompanying CD-ROM.) Understanding was

measured by the percentage of correct responses to the study

scenarios. The results of statistical analyses presented in the

following sections were all conducted using analysis of vari-

ance (ANOVA) procedures at a 95% level of conﬁdence.

An overall analysis of the demographic characteristics of

the drivers in relation to the total number of survey scenarios

evaluated and percentage of correct responses is presented in

Table 3-1. Note that these results contain responses to all dis-

play types.

All-Red Indication Findings

•

Figure 3-9 summarizes the percentage of correct

responses to all-red indications. Statistical analysis of

Figure 3-8. Example of driver evaluation study workstations at shopping center.

PPLT Signal Displays

Percent Correct Response

93.6

91.2

96.6

94.0

94.3

100

5-Section

Horizontal

5-Section

Vertical

5-Section

Cluster

4-Section

Cluster

3-Section

Vertical

Figure 3-9. Driver understanding of all-red indications in PPLT

signal displays.

the data determined that PPLT signal display type did

not affect driver understanding with all-red indications.

•

There was a significant difference in response times

among gender and age groups. Female drivers (93.5%)

had a slightly higher correct response rate than male

drivers (92.8%). Drivers over the age of 65 had a 90.0%

correct response rate compared with a 93.8% correct

response rate for the under-24 age group.

Protected Indication Findings

•

Figure 3-10 summarizes the percentage of correct

responses to protected indications in the PPLT displays.

•

As suggested by Figure 3-10, a statistically significant

difference was found between display arrangements

when considering protected indications, particularly

when simultaneous displays were used in the same dis-

play arrangement. Simultaneous indications were not

presented with the three and four-section displays.

•

Driver understanding was significantly lower when a

green arrow and a circular red indication were pre-

sented simultaneously within all ﬁve-section PPLT sig-

nal displays.

•

When the green arrow and circular red indications were

shown simultaneously in a ﬁve-section signal display,

driver understanding was lowest with the horizontal dis-

play. Locating the green arrow to the right of the circu-

lar red indication in a ﬁve-section horizontal display

arrangement as required by the MUTCD appeared to

increase confusion.

•

When the green arrow and circular green indications are

shown simultaneously, the ﬁve-section horizontal dis-

play has the lowest level of driver understanding.

•

When only a green arrow indication is provided in a left-

turn signal display, driver understanding of the protected

indications was not affected by the through movement

indication.

•

Age was found to be statistically signiﬁcant because

drivers over the age of 65 had an 82.5% correct response

rate compared with an 88.6% correct response rate for

drivers aged 24 to 44. In general, the percentage of cor-

rect responses decreased as driver age increased.

TABLE 3-1 Summary of survey scenario demographics (as a function of

total number of surveys evaluated for all display types)

Demographic Level

Number of

Responses

Percentage of

Correct Responses

Dallas

9,299

Dover 75.6

68.5

Oakland Count

9,722

70.7

Colle

e Station 9,034 74.6

Seattle

78.5

Portland

71.5

ertino 69.7

Location

Orlando

66.7

Male

72.7

Female

71.3

Gender

Not Provided 456

68.9

< 24 19,942

72.2

24 – 44 32,191

73.1

45 – 65 15,171

71.1

> 65 4,958

67.3

Age

Not Provided 508

74.2

Cit

42,063 71.8

Suburb

72.8

Rural

71.8

Residence

Not Provided 299

67.9

None

62.7

< 10,000 22,523

70.2

10,000 to 20,000 32,746

74.1

> 20,000 13,916

72.4

Miles Driven

Not Provided 241

77.2

72.3

Yes

66.1

Color Vision

(Have trouble seeing

red or green?)

Not Provided 181

80.7

h School 20,738 67.4

Some College

25,849

College Degree 25,891

74.6

Education

Not Provided 292

66.8

9,658

8,869

8,923

8,226

42,189

30,125

21,880

8,528

3,344

69,217

3,372

73.3

•

The average response time associated with the five-

section PPLT signal display arrangement was greatest and

was longer than the average of all response times for all

PPLT signal display arrangements. The longest response

time was associated with the ﬁve-section horizontal PPLT

signal display. This higher average response time corre-

lates with the lower correct response rate attributed to the

simultaneous illumination of the green arrow and the solid

circular red indication.

•

When the protected left-turn indication was exclusive

(from the through-movement indication), the through-

movement indication did not significantly affect driv-

ers’ understanding of the left-turn indication. The

highest correct response rates were associated with

the four-section and three-section PPLT signal display

arrangements. These results support the use of an exclu-

sive head.

Permissive Indication Findings

•

Figures 3-11 and 3-12 summarize the percentage of

correct responses to permissive indications in the PPLT

displays. As shown in Figure 3-11, the circular green

100

93.3

92.4

92.0

93.3

92.9

91.2

89.0

61.6

91.5

79.2

73.1

90.2

PPLT Signal Displays

Percent Correct Response

5-Section

Horizontal

5-Section

Vertical

5-Section

Cluster

4-Section

Vertical

3-Section

Vertical

4-Section

Cluster

Through Movement - Red Through Movement - Green

Figure 3-10. Driver understanding of protected indications in PPLT signal displays.

63.8

Flashing Red Ball

56.6

Flashing Yellow

Arrow

61.7

Flashing Yellow

Ball

50.4

Green Ball

55.6

Flashing Red Arrow

Permitted Indication

Percent Correct Response

Figure 3-11. Percentage of correct responses for the permitted indications in PPLT

signal displays.

indication had the lowest level of driver comprehen-

sion, at 50%.

•

Figure 3-12 provides more detailed information on the

percentage of correct responses to permissive indica-

tions in PPLT signal displays by display type. The dif-

ferences among the arrangements were determined to be

directly related to the permissive indications used within

each signal display.

•

The correct response rate for male drivers was 57.7% as

compared with a 54.8% correct response rate for female

drivers. This difference was statistically signiﬁcant. Age

was found to be statistically signiﬁcant because drivers

over the age of 65 had a 51.4% correct response rate

compared with a 57.5% correct response rate for drivers

aged 24 to 44.

•

Drivers over the age of 65 had extremely low correct

response rates with the permissive circular green indi-

cations. When the permissive circular green indication

and the circular red through-movement indication were

shown, less than 29% of older drivers correctly responded.

•

Drivers over the age of 65 had a higher correct response

rate with the ﬂashing circular red indication and ﬂash-

ing yellow permissive indications than all other age

groups. A total of 70% of drivers over the age of 65 cor-

rectly understood the ﬂashing circular red indication.

•

Average response time was generally lower for the ﬂash-

ing permissive indications. A trend in average response

time by age was very evident as drivers over the age of

65 took between 2 and 4 sec of additional time to respond

when compared with drivers under the age of 24.

•

The analysis of the permissive left-turn indication reveals

that a ﬂashing indication is better understood than a solid

indication, and a circular indication is better understood

than an arrow indication.

General Findings

•

Drivers’ understanding of the PPLT signal displays

was evaluated by combining the results from the all-

red, protected left-turn, and permissive left-turn sce-

narios. The cluster arrangement was associated with

the largest number of correct responses and the fewest

fail critical responses of the five-section PPLT signal

display arrangements. (For further information on these

analyses, consult Working Paper No. 3, which is included

in the appendixes on the accompanying CD-ROM.) The

overall highest correct response rate was associated with

the three-section vertical PPLT arrangement that uses a

ﬂashing circular red indication as the left-turn permissive

indication.

•

There was no evidence to suggest that the placement of

the PPLT signal display inﬂuenced drivers’ understand-

ing of the PPLT signal display.

•

The inﬂuence of the intersection geometry could not be

measured because the study simulated only exclusive

left-turn lane conﬁgurations.

64.5

61.1

55.3

54.3

59.9

56.0

51.4

48.1

54.8

48.5

53.5

44.5

PPLT Signal Displays

Percent Correct Response

5-Section

Horizontal

5-Section

Vertical

5-Section

Cluster

4-Section

Vertical

3-Section

Vertical

4-Section

Cluster

Through Movement - Red

Through Movement - Green

Figure 3-12. Percentage of correct responses to permissive indications by PPLT

display type.

FIELD TRAFFIC OPERATIONS STUDY

Capacity and delay are two common measures of effec-

tiveness in evaluating signalized intersection operations (16).

Capacity is deﬁned as the maximum rate of ﬂow at which

vehicles can be reasonably expected to traverse a point, a uni-

form segment of a lane, or a roadway during a speciﬁc period

under prevailing roadway, traffic, and control conditions.

Delay is the additional travel time experienced by a driver

beyond what would reasonably be desired for a given trip.

Left-turn capacity at a signalized intersection is calculated

using the saturation ﬂow rate. Saturation ﬂow rate is deﬁned

as the maximum rate of ﬂow that can pass through a given

lane group under prevailing traffic and roadway conditions,

assuming that the lane group has 100% of green time avail-

able (16). Saturation ﬂow is usually reached after the fourth

vehicle in a queue has entered the intersection.

At the beginning of each protected left-turn movement, the

ﬁrst several vehicles in the queue experience start-up time

losses that result in their movement at less than the saturation

ﬂow rate (16). This time loss is referred to as the start-up lost

time, which is made up of the perception and reaction time

(response time) to the change in signal indication along with

the vehicle acceleration time to free-ﬂow speed. Start-up lost

time is important in evaluating left-turn lane capacity and

driver reaction to the traffic signal indication. Reaction time

was considered to include the perception time in addition to

the reaction time to the onset of the green arrow indication.

Gap acceptance and follow-up headway also affect left-

turn capacity and delay. Gap acceptance refers to the time

headway in the opposing traffic stream that left-turn drivers

are willing to turn through during the permissive left-turn

phase. The median time headway between two successive

vehicles in opposing traffic streams accepted by left-turn

drivers during the permissive left-turn phase is referred to as

the critical gap. Follow-up headway is the time between the

departure of a permissive left-turn vehicle and the departure

of the next vehicle using the same gap under a condition of

continuous queuing. Left-turn delay can be evaluated by

quantifying each of the operational variables and applying

the procedures included in Chapter 16 of the Highway Capac-

ity Manual (16).

Objective

The objective of the traffic operations study was to quan-

tify the capacity and delay effects of PPLT signal displays

and indications currently used in the United States. This

objective was achieved through analysis of saturation flow

rates, lost times, response times, and follow-up headways.

This analysis was used to describe the driver understanding

associated with the PPLT signal displays and indications.

Methodology

Field traffic operations data were collected at eight study

locations: Dallas, Texas; Dover, Delaware; Oakland County,

Michigan; College Station, Texas; Seattle, Washington; Port-

land, Oregon; Cupertino, California; and Orlando, Florida.

These locations provided a range of PPLT signal displays

and left-turn permissive indications and were the same sites

as the photographic driver study.

With the assistance of local transportation officials, a total

of 26 study intersections were identiﬁed. Intersection selec-

tions were based on left-turn lane geometry, PPLT display

arrangement, and left-turn phasing. At each intersection, a

video camera was installed to record the left-turn traffic ﬂow.

In addition to video taping the left-turn traffic, a researcher

at the intersection simultaneously operated a portable com-

puter and the software program HEADWAY to collect data.

The data collected using the HEADWAY program were then

used to compute left-turn saturation ﬂow rate, start-up lost

time, response time, and follow-up headway.

Results

Saturation ﬂow rate, lost time, response time, and follow-

up headway data were collected for 26 intersections located

throughout the eight study locations. Two additional site loca-

tions were studied in Texas to evaluate the green indication

used in the Dallas Display in a lead-lead and lag-lag phase

sequence. The PPLT signal indications investigated included

the circular green, ﬂashing circular yellow, ﬂashing red arrow,

and flashing circular red indications. At the time the field

observation studies were prepared, no ﬂashing yellow arrow

display was available for study (the ﬁeld studies were com-

pleted before implementation of the Sparks display in April

of 1998).

Findings

The detailed findings of the field traffic operations stud-

ies are documented in Working Paper #4. General findings

include the following:

•

As shown in Table 3-2, analysis of the observed satura-

tion ﬂow rate data determined that the geographic loca-

tion was a signiﬁcant contributor to the variance in the

average saturation ﬂow rates. The PPLT signal display

arrangement and phasing were not statistically signiﬁcant.

•

Analysis of the start-up lost time data showed that the

differences in average start-up lost time were signiﬁ-

cantly inﬂuenced by the PPLT signal phasing and were

not signiﬁcantly inﬂuenced by the PPLT signal display

arrangement or study location.

TABLE 3-2 Saturation ﬂow rate data

Saturation Flow

Rate (vphgpl)

City ID

PPLT

Display

Left-Turn

Phasing

Average SD

Rank

Order

1 5-Vert. GB Lead 2,201 36.55 6

2 5-Horz. GB Dallas-Lead 2,211 36.09

3 5-Horz. GB Dallas-Lag 2,320 97.50

4 5-Horz. GB Dallas-Lead 2,091 98.38

Dallas,

5 5-Horz. GB Dallas-Lag 2,063 58.97 1

6 4-Cluster FRA Lead 2,214 151.91

7 4-Cluster FRA Lead 1,977 19.75

16.5

Dover,

8 4-Cluster FRA Lead 1,999 118.21

9 3-Vert. FRB Lag 2,168 111.09

10 3-Vert. FRB Lag 2,399 86.51

Oakland

County, MI

11 3-Vert. FRB Lag 2,402 52.54

12 5-Horz. GB Lead 1,973 60.65

13 5-Horz. GB Lead 2,025 78.42 1

College Station,

14 5-Cluster GB Lag 2,045 57.13

15 4-Vert. FYB Lead 1,773 3.00

16 4-Vert. FYB Lead --- ---

24.5Seattle, WA

17 4-Vert. FYB Lead --- ---

24.5

18 5-Cluster GB Lead --- ---

24.5

19 5-Cluster GB Lead 1,871 40.13

Portland, OR

20 5-Cluster GB Lead 1,977 7.00

16.5

21 4-Vert. FRA Lead 2,065 48.26 1

22 4-Vert. FRA Lead 1,944 146.93

Cupertino, CA

23 4-Vert. FRA Lead --- ---

24.5

24 5-Cluster GB Lead 2,067 42.58

25 5-Cluster GB Lead 2,072 61.8

Orlando, FL

26 5-Cluster GB Lead 1,963 102.93

1. Intersection Identification Number

2. Number of signal display sections (3, 4, or 5) - arrangement (Horizontal, Vertical, or Cluster)

3. Permitted Indication - G = Green; Y = Yellow; R = Red; B = Ball (circular indication); A = Arrow; F = Flashing

4. Standard Deviation

•

Analysis of the response time data indicated differ-

ences in average response time to be signiﬁcantly inﬂu-

enced by the PPLT signal arrangement and phasing and

by the study location, although most of the variability

was explained by the effect of the PPLT signal phasing.

•

Analysis of the follow-up headway data considered the

effect of the PPLT signal display and the left-turn per-

missive indication. The results of this analysis are shown

in Table 3-3.

As shown in Table 3-3, the follow-up headway associated

with the ﬂashing red PPLT indication used in Dover was longer

than all other observed follow-up headways. The follow-up

headway associated with the flashing red PPLT indication

used in Michigan was comparable to the follow-up headway

associated with the circular green, ﬂashing circular yellow,

and ﬂashing red arrow indications.

FIELD TRAFFIC CONFLICT STUDY

Traffic conﬂicts involve the interaction of two or more

drivers where one or more drivers take evasive action to

avoid a collision (17, 18, 19). Traffic conﬂict studies provide

one of the most effective ways to supplement crash studies

in estimating the crash potential of various PPLT signal dis-

plays. In addition, traffic conﬂict studies can be used to esti-

mate traffic safety when crash rates are not available.

Collecting traffic conflict data can also be valuable in iden-

tifying whether unsafe vehicle maneuvers are prevalent at

an intersection. Conﬂict studies also provide an effective way

to study speciﬁc geometric applications at PPLT intersections.

Conﬂicts can be considered vehicle interactions that may

lead to crashes. For a conﬂict to occur, the road users must

be on a collision course (17, 18, 19). The primary require-

ment of a traffic conﬂict is that the action of one user places

another user on a collision path unless evasive action is taken.

Collisions and near-miss situations that occur without evasive

maneuvers, or when the evasive action is inappropriate or

inadequate for conditions, are also recorded as conﬂicts.

Conflict studies are used to evaluate safety, as well as to

select signal phasing. An Institute of Transportation Engineers

(ITE) study found that 33% of reporting agencies used a left-

turn conﬂict rate of four conﬂicts per 100 left-turning vehicles

as a warrant for implementing PPLT signal phasing (20).

Objective

The objective of the traffic conflict study was to quantify

left-turn conflict rates and event rates for different PPLT

signal displays and indications. Traffic events are unusual,

dangerous, or illegal non-conﬂict maneuvers such as red indi-

cation violations, backing, hesitation on signal change, and

slowing considerably in a traffic lane.

Methodology

Conﬂict studies were conducted in Dallas, Texas; Dover,

Delaware; Oakland County, Michigan; College Station, Texas;

Seattle, Washington; Portland, Oregon; Cupertino, California;

and Orlando, Florida. These locations provided a range of

PPLT signal displays and left-turn permissive indications and

were the same locations where the photographic driver study

and operations studies were conducted.

With the assistance of local transportation officials, a total

of 26 study intersections were identiﬁed. Speciﬁc intersec-

tions were selected based on the left-turn lane geometry,

PPLT display arrangement, and left-turn phasing.

At each intersection, conﬂict data were recorded on a data

collection form. Each intersection was observed for 8 hr on

a weekday between 7:00 a.m. and 6:00 p.m. To back up real-

time observations, a video camera was also installed at each

intersection. The videotape was reviewed to clarify any dis-

crepancy in the manually recorded ﬁeld data.

Observed conﬂicts were categorized as follows:

•

Type 1—opposing left-turn conﬂicts;

•

Type 2—left-turn/same direction conﬂicts;

•

Type 3—left-turn/lane change conﬂicts; and

•

Type 4—secondary conﬂicts, such as those involving a

pedestrian or bicyclist or resulting from a lane overﬂow.

Figure 3-13 illustrates the conﬂict types.

The observed traffic events were also categorized by type

as follows:

•

Type 1—driver hesitating on the left-turn protected

indication

•

Type 2—driver hesitating on the left-turn permissive

indication

•

Type 3—driver going through the circular red indication

•

Type 4—driver backing a vehicle out of the intersec-

tion, back into the left-turn lane

TABLE 3-3 Rank order of PPLT signal display by average

follow-up headway

PPLT Signal

Display

Location Permissive Indication

Rank

Orde

5-Section Vertical Dallas, TX Circular Green

4-Section Vertical Seattle, WA Flashing Circular

Yellow

4-Section Vertical Cupertino, CA Flashing Red Arrow

3-Section Vertical

Oakland County,

Flashing Circular Red

5-Section

Horizontal

College Station,

Circular Green

5-Section Cluster Portland, OR Circular Green

4-Section Cluster Dover, DE Flashing Red Arrow

Type 1: Opposing left-turn conflict

Type 3: Lane-change conflict

Type 4: Left-turn ped/bicycle far/near side conflict

Type 2: Left-turn, same direction conflict

Type 4: Opposing right-turn-on-red conflict

Type 4: Left-turn

overflowlane

Figure 3-13. Illustration of conﬂict types.

Results

A total of 11 hr of data were collected at each of the 24

study intersections for a total of 264 hr of observation time.

During the observation period, the research team observed

approximately 22,000 vehicles; of which 5,000 were left-

turn vehicles; and 17,000 were through vehicles.

A total of 166 left-turn conflicts were observed: 155

Type 1, 9 Type 2, and 2 Type 3 conﬂicts. No Type 4 conﬂicts

were observed.

Of a total of 242 traffic events observed, 147 were Type 1,

53 were Type 2, 5 were Type 3, and 37 were Type 4.

Findings

•

Overall, the left-turn conﬂict rates were low for all PPLT

displays evaluated.

•

Few left-turn conﬂicts were associated with the PPLT

display.

•

Most left-turn events were related to hesitation at the

onset of the green indication.

•

Aggressive driving appeared to be the cause of 146 of

the 155 Type 1 conﬂicts. Drivers continued to make left-

turn maneuvers during the yellow and all-red phase fol-

lowing the protected left-turn phase and were in conﬂict

with the opposing through traffic.

•

Eight Type 1 conﬂicts appeared to be the result of the

driver’s assuming the right-of-way when the left-turn

permissive circular green indication was illuminated.

Two of these conﬂicts occurred at intersections with the

five-section horizontal PPLT arrangement and the

remainder occurred at intersections with the five-

section cluster PPLT arrangement.

•

One Type 1 conﬂict appeared to be the result of the

driver’s assuming the right-of-way when the left-turn

permissive ﬂashing red arrow indication was illumi-

nated on a four-section cluster arrangement.

•

The nine Type 2 conﬂicts were primarily the result of a

driver’s hesitating to turn left on the left-turn permissive

indication. The sudden hesitation would cause a con-

flict with the following vehicles. There appeared to be

a relationship between the driver’s understanding of the

permissive circular green indication and the observed

Type 2 conﬂicts.

•

The two Type 3 conﬂicts were a result of driver error

and not the lack of understanding of the PPLT signal

display.

•

The largest occurrence of Type 1 traffic events involved

the ﬁve-section horizontal PPLT signal display arrange-

ment. The simultaneous illumination of the green arrow

and the circular red indications appeared to increase the

workload of the driver, resulting in an increase in driver

uncertainty.

•

Type 2 traffic events were observed at each of the study

intersections. The occurrence did not appear to be related

to the PPLT signal arrangement or phasing or indication.

•

Numerous drivers were observed proceeding through

the all-red indication (i.e., red light runners). Therefore,

Type 3 events were recorded only when the action was

clearly a function of driver misunderstanding. For this

reason, only ﬁve Type 3 events were recorded, and the

occurrences showed no pattern to suggest an inﬂuence

of the PPLT signal display, indication, or phasing.

•

Of the 37 observed Type 4 events, 33 were associated

with a ﬂashing permissive indication. The driver would

enter the intersection during the permissive phase and

not have the opportunity to make the left-turn maneu-

ver. The driver would then choose to back up.

CRASH DATA ANALYSIS

Given the the extensive use of crash data in many past

studies in the evaluation of left-turn control, the research

team conducted a limited study of crash history related to the

unique displays used in PPLT control. The project panel and

research team recognized that current crash reporting tech-

niques do not adequately document causes of a crash as they

relate to traffic signal operation, much less the particular sig-

nal display. To overcome this short coming, the research

team identiﬁed a work plan that would compare the various

PPLT control displays against one another and make judg-

ments as to whether a particular display was more or less

prone to increased crash occurrence.

Objective

The objective of the crash analysis was to determine and

compare left-turn crash rates associated with various PPLT

signal displays. The details of the crash data analysis are

described in the following sections.

Methodology

To accomplish the objective of the crash data analysis,

three tasks were conducted. Those tasks are listed below:

•

Perform a literature review to obtain background infor-

mation on crash data analysis procedures and to review

the results of previous studies;

•

Obtain traffic volumes, geometric design, signal display

information, and 3 years’ of crash data for the study

intersections (the same intersections studied in the oper-

ational study and the conﬂict study were evaluated in the

crash analysis). In addition, crash data were sought from

agencies around the United States;

•

Reduce the data and analyze the results.

Results

Since the mid-1970s, more than 40 reports have been pub-

lished on the use and operational impacts of the PPLT signal

control treatment. Several of these reports were spurred by

the increase in the use of PPLT phasing. In general, many of

these reports identiﬁed trends in vehicle delay, vehicle con-

ﬂicts, and vehicle crashes. Further, these studies related the

effects of PPLT phasing to geometric and physical condi-

tions, such as traffic volume, number of opposing lanes, sig-

nal displays, and the use of supplemental signing. Over 30%

of the reports are focused on warrants or guidelines related

to the use of PPLT phasing.

The purpose of the literature review was to identify how

crash data were evaluated and what crash statistic should be

used for this study. The literature review was selective in that

not all reports related to PPLT phasing were studied. Certainly,

many good reports are available as a resource. The literature

review identiﬁed reports that were key to this study because of

the signiﬁcance related to their ﬁndings and/or the strict focus

related to PPLT phasing. Specifically some reports focused

on the PPLT indications, display arrangement, and/or display

location—all of which are study factors for this NCHRP study.

The select literature review identiﬁed several key ﬁndings

related to PPLT phasing:

•

Hummer determined that the crash frequency is higher

for PPLT intersections with leading left-turns compared

with lagging left-turns (21). However, Upchurch deter-

mined that this was true for intersections with three

opposing lanes of traffic (22). Upchurch determined that

the lagging PPLT had the worst crash record when there

were two opposing lanes of traffic.

•

Almost all literature shows that the leading protected

left-turn phasing has the lowest crash rate (22, 23).

•

Upchurch (22) and Hauer et al. (24) both determined

that crash frequency relationship to traffic demand is

nonlinear.

•

Bonneson and McCoy determined there was no statisti-

cal difference in crash frequency among the most com-

mon PPLT display arrangements (25).

•

The Washington Section of ITE determined the ﬂashing

circular yellow display was safer than the circular green

permissive display (26).

•

Agent (27, 28, 29) determined that the use of PPLT phas-

ing can reduce left-turn delay by 50% and total delay by

24%, compared with protected-only phasing.

Although not all reports agree with one another, the gen-

eral ﬁndings related to the use in PPLT phasing, in aggregate,

suggest the following:

•

Vehicle delay decreases,

•

Fuel usage decreases,

•

Vehicle progression is improved, and

•

Vehicle crashes increase.

The use in PPLT phasing should be applied on a case-by-

case basis, because not all intersection approaches are candi-

dates for PPLT phasing.

Data Analysis

For the purpose of this study, the research team identiﬁed

three sources of crash data for analysis. First, crash data were

obtained from the same 24 study intersections examined in the

operational and conﬂict studies. Speciﬁcally, crash data were

obtained for each of the following eight geographic regions

of the United States (three intersections per geographic loca-

tion): College Station and Dallas, Texas; Portland, Oregon;

Seattle, Washington; Detroit, Michigan; Cupertino, Califor-

nia; Dover, Delaware; and Orlando, Florida. The crash data

were requested for the most recent 3 years.

The second source of crash data was from the network of

volunteers identiﬁed through the Agency Survey. Without

exception, these data represented the circular green permis-

sive display (as allowed in the MUTCD).

For the third source of crash data, the research team evalu-

ated the database created in 1988 as part of a previous FHWA

study of PPLT displays (23). This third source of data was

identiﬁed as a potential source of historical data, in that some

of those intersections were thought to still be in operation as

they were in 1988. The research team intended to use the

information in the database as additional analysis locations

and thus as new data points or for comparative analysis. It

was later determined that none of the intersections in the data-

base were using the same PPLT display and the same inter-

section geometry and, as a result, no further analysis of the

database was prepared.

Findings

Crash Analysis of the 24 Study Intersections

The research team calculated several different crash sta-

tistics as follows: (1) average number of crashes per year per

intersection, (2) the average number of crashes per year per

100 left-turning vehicles, (3) the average number of crashes

per year per 100,000 left-turning times opposing through

vehicles, and (4) the average rate for the intersection based

only on left-turn crashes. The ﬁndings of these analyses are

presented below.

•

The analysis of the average crash rate per year indicated

that the four-section (dual indication) ﬂashing circular

yellow indication used in Seattle experienced the fewest

number of crashes (per year) at 0.75. The highest aver-

age crash rate per year occurred in Oakland County,

Michigan. Ranking by city is summarized in Table 3-4.

•

An analysis of average crashes per 100 left-turning

vehicles indicated again that the ﬂashing circular yellow

indication used in Seattle was the lowest. The highest

average statistic was 2.29 in College Station, Texas,

which had a circular green indication. Ranking by city

is summarized in Table 3-5.

•

An analysis of the average crashes per [100,000 left-

turning vehicles multiplied by opposing through vehi-

cles] indicated yet again that the ﬂashing circular yellow

indication used in Seattle had the lowest statistic and

that the circular green indication in College Station had

the highest statistic. Ranking by city is summarized in

Table 3-6.

•

An analysis of the average left-turn crash rate by inter-

section indicated that Cupertino, California, had the low-

est average at 0.28. The highest average crash rate by

intersection was College Station, Texas, at 0.70. Rank-

ing by city is summarized in Table 3-7.

In the aggregate, the crash analysis ﬁndings show that the

PPLT displays studied did not perform consistently within a

selection of four crash statistics. The ranking of one crash

statistic did not match that of another crash statistic. Two par-

ticular displays, the ﬂashing circular yellow indication used

in Seattle and the circular green indication used in College

Station, did rank similarly among most of the crash statistics.

Further, there was no correlation between the crash rate analy-

sis and the conﬂict study results in rank ordering.

Crash Analysis of the Intersections Identiﬁed

Through Volunteer Agencies

Six volunteer agencies responded to a request to supply

crash data: the City of Modesto, California; the Mississippi

DOT; the North Carolina DOT; the Texas DOT; the Wash-

ington DOT; and the Wisconsin DOT. A total of 135 inter-

sections of crash data were received, of which 120 intersec-

tions were used in the analysis. These crash data represented

a total of 284 intersection approaches using PPLT control.

All of these intersections used the ﬁve-section cluster or ﬁve-

section vertical display arrangement. The crash analysis con-

sisted of computing the average crash rate for the intersec-

tion and resulted in a rate that represents the overall average

accident rate for the PPLT design type, calculated as the aver-

age of the intersection accident rates.

In general, most of the intersections yielded results similar

to those found in Portland, Orlando, Dallas, and College Sta-

tion (locations that also use the circular green indication).

Several intersections yielded average crash rates of 1.5 and

higher (which may indicate a safety problem). No meaningful

trends were identiﬁed.

DRIVER CONFIRMATION STUDY

The driver conﬁrmation study was conducted using fully-

interactive dynamic full-scale driving simulators located in

the Human Performance Laboratory on the University of

Massachusetts—Amherst (UMass) campus and at the Texas

Transportation Institute (TTI) at Texas A&M University. An

evaluation of the same PPLT signal displays in a static envi-

ronment was also completed at both locations to provide

TABLE 3-4 Ranking of PPLT performance based on

crashes per year

City

PPLT Indication

Crash Rate

Seattle Circular Flashing Yellow

0.75

Cupertino Flashing Red Arrow

0.83

Dover Flashing Red Arrow

0.85

Portland Circular Green

1.04

Orlando Circular Green

1.48

Dallas Circular Green

2.06

College Station Circular Green

2.53

Oakland County

Flashing Circular Red

2.92

TABLE 3-5 Ranking of PPLT performance based on

crashes per 100 left-turning vehicles

City

PPLT Indication

Crash Rate

Seattle Circular Flashing Yellow

0.47

Portland Circular Green

0.71

Orlando Circular Green

0.73

Cupertino Flashing Red Arrow

0.87

Dover Flashing Red Arrow

0.96

Dallas Circular Green

1.10

Oakland County Flashing Circular Red

1.23

College Station Circular Green

2.29

TABLE 3-6 Ranking of PPLT performance based on

crashes per [100,000 left turn  opposing through vehicles]

City

PPLT Indication

Crash Rate

Seattle Circular Flashing Yellow

0.87

Cupertino Flashing Red Arrow

0.91

Orlando Circular Green

0.92

Oakland County Flashing Circular Red

1.18

Dover Flashing Red Arrow

1.85

Portland Circular Green 2.2

Dallas Circular Green

4.56

College Station Circular Green

6.75

comparison data to the simulator experiment as well as to the

photographic driver study.

Objective

The objective of the driver conﬁrmation study was to eval-

uate drivers’ comprehension of the most promising types of

PPLT signal displays using full-scale driving simulators. The

following sections summarize the development and adminis-

tration of the driving simulation experiment and the follow-

up static evaluation completed at both universities.

Signal Displays Studied

The previous research activities were used collectively to

identify the displays studied in this research task. Following

lengthy and detailed discussions of study results, the research

team and project panel identiﬁed 12 PPLT signal displays for

further evaluation. The selected displays differed in permis-

sive indication, display face, location, and through movement

indication. Each of the PPLT signal displays included only

the circular green indication and/or flashing yellow arrow

permissive indications.

The ﬂashing red permissive indications and the circular

yellow permissive indication were not evaluated in this effort.

The ﬂashing red arrow and circular red indications were

eliminated by the project panel from future consideration

after much discussion with the research team because of the

message that these indications presented (i.e.,“stop, then pro-

ceed if a gap is available” rather than the more typical PPLT

“proceed if a gap is available” message). Additionally, the

ﬂashing permissive red indications may dilute the meaning

of other red stop indications. The circular yellow indication

was eliminated because the ﬂashing yellow arrow display

had nearly the same level of driver comprehension and the

ﬂashing yellow arrow could be applied to either left- or right-

turn treatments.

The circular green permissive indication represented the

current state of the practice, and the flashing yellow arrow

permissive indication represented the most promising alter-

native, based on study ﬁndings to date and panel recommen-

dations. Figure 3-14 illustrates the PPLT displays evaluated

in the driving simulation experiment.

As shown in Figure 3-14, the research team included a

unique PPLT display that was currently in operation in Sparks,

Nevada. This “Sparks Display” uses an exclusive left-turn

signal display (ﬁve-section cluster) and simultaneously illu-

minates the circular green indication and the ﬂashing yellow

arrow indication during the permissive left-turn interval. In

this application, the ﬂashing yellow arrow indication is sup-

plemental information to the circular green indication.

Simulator Environment

Similar driving simulators at UMass and TTI were used

to complete the experiment. The two simulators are brieﬂy

described below.

A ﬁxed-base, fully interactive dynamic driving simulator,

housed in the Human Performance Laboratory on the UMass

campus, was used to complete the driving simulation exper-

iment. The vehicle base of the driving simulator is a 1995,

four-door Saturn sedan. Drivers can control the steering, brak-

ing, and accelerating similar to the actual driving process; the

visual roadway adjusts according to the driver’s actions. Three

separate images were projected to create the “visual world”

on a large semi-circular projection screen creating a ﬁeld-of-

view that subtends approximately 150 deg. The simulator

also featured a surround audio system, a 60-Hz refresh rate,

and a resolution of 1024 x 768 dots per inch. The UMass

driving simulator is pictured in Figure 3-15.

At TTI, the apparatus used for the experiment was a driving

environment simulator (DESi). Almost identical to the UMass

simulator, DESi consisted of three white screens 2.28 m (90 in)

in height and width, a 1995 Saturn SC2 complete vehicle,

three image-generation personal computers, one data collec-

tion personal computer, and three liquid crystal display Prox-

ima 6810 projectors. The three separate images projected

onto the screens were aligned so they appeared to the driver

as one single image covering a 150-deg ﬁeld of view hori-

zontally and a 50-deg ﬁeld of view vertically. Consistent with

the UMass simulator, participants sat in the center of the DESi

in the driver’s seat of the Saturn, from which they could con-

trol the steering, braking, and accelerating similar to the

actual driving process. The TTI driving simulator is pictured

in Figure 3-16. Figure 3-17 depicts a typical PPLT intersec-

tion in the UMass driving simulator experiment.

Simulator Study Methodology

Each driver was presented with a practice course before

the execution of the actual experimental modules. The prac-

tice course was provided to familiarize the driver with the

simulator environment. Oral communication between the

TABLE 3-7 Ranking of PPLT performance based on

average left-turn crash rate

City

PPLT Indication

Crash Rate

Cupertino Flashing Red Arrow

0.28

Dover Flashing Red Arrow

0.29

Dallas

Seattle

Circular Green

Circular Flashing

Yellow

0.34

0.3

Oakland County Flashing Circular Red

0.44

Orlando Circular Green

0.49

Portland Circular Green 0.52

College Station Circular Green

0.70

Scenario

1, 2

3, 4

5, 6

7, 8

9, 10

11, 12

Lens Color

and

Arrangement

Left-Turn Indication

Protected

Mode

Permitted

Mode

i:\projfile\2

036\cdrfiles\figures\Figu

re_3-5.cdr

GGG

or or or

R = RED Y = YELLOW G = GREEN = FLASHING YELLOWY

1, 3, 5, 7, 9, 11 -GB through indication; 2, 4, 6, 8, 10, 12 - RB through indication

The indication illuminated for the

iven mode is identified b

the color letter

Figure 3-14. PPLT displays evaluated in driver simulator experiment.

research team and the driver was avoided. Therefore, drivers

navigated through the modules by using guide signs provided

on each intersection approach. In addition, drivers were asked

to observe speed limit signs (30 mph), providing a higher

level of realism and speed control during the experiment. The

driving portion of the experiment, including the practice

module, required between 15 and 20 min to complete.

Drivers’ response to each PPLT signal display scenario

presented was recorded manually as correct or incorrect by

two members of the research team. Incorrect responses were

further classiﬁed as being fail-safe or fail-critical. A fail-safe

response was one in which the driver did not correctly

respond to PPLT signal display but did not infringe on the

right-of-way of the opposing traffic. A fail-critical response

was an incorrect response in which the driver incorrectly

responded to PPLT signal display and impeded the right-of-

way of opposing traffic, creating the potential for a crash.

Throughout the study, drivers were asked to express their

thoughts out loud about anything they observed. Research

team members were present to record the results of the sim-

ulation, including responses at each intersection and other

driving-related factors such as indecision, unnecessary brak-

ing, or pertinent verbal comments. Each experiment was

recorded on videotape allowing the researchers to verify

and review the manually collected data.

Opposing Traffic

Each of the PPLT signal displays was evaluated with oppos-

ing traffic at the intersection. The introduction of opposing

traffic required drivers to evaluate simultaneously the PPLT

signal display, traffic movement, and opposing gaps to com-

plete a safe permissive left-turn maneuver. This methodology

was used to replicate the decision process required during

actual operation of a motorized vehicle within the roadway

system.

All gaps in opposing traffic were consistently applied at

intersections where drivers were required to make a permit-

ted left-turn maneuver. Six opposing vehicles were used.

Two vehicles were always positioned at the stop bar in the

two through lanes opposing the left-turn driver. The remain-

ing four were positioned further upstream in a speciﬁed gap

sequence. Gaps were set at 3 and 7 sec in a series of 7-3-7-7;

therefore, opposing vehicles crossed the intersection 7, 10,

17, and 24 sec behind the two initially queued opposing vehi-

Figure 3-15. UMass Human Performance Laboratory driving simulator.

cles. The critical gap concept was used to select the gap sizes.

The Highway Capacity Manual indicates that a critical gap

value of approximately 5.5 sec for permitted left-turn maneu-

vers in the design of a four-lane roadway is acceptable (16).

Therefore, a 3-sec gap was selected because it was small

enough that most drivers will not accept it, and a 7-sec gap

was selected because it was generally acceptable to most driv-

ers. Providing a consistent sequence of 3- and 7-sec gaps

prevented gap selection from being a signiﬁcant variable in

the PPLT analysis.

The only difference in the driving simulator experiments

at UMass and TTI was the initial method of introducing

the opposing traffic. For simpliﬁcation, the methods will be

referred to as the Release Method of Opposing Traffic (RMOT)

and the Continuous Method of Opposing Traffic (CMOT).

The RMOT is the methodology described above. A trigger

in the simulator model, similar to that used to change the sig-

nal indications, was placed near the left-turn stop bar at each

PPLT intersection to release the opposing traffic. By placing

the opposing traffic release trigger approximately 5 ft from

the stop bar, left-turn drivers were required to make a decision

as to the meaning of the PPLT signal indication and desired

action before knowing the actions of the opposing traffic.

At TTI, the research team explored the effects of a slightly

different opposing vehicle method. A total of 116 drivers

completed the experiment, which used the CMOT traffic. The

CMOT method of opposing traffic had the opposing traffic

moving as the driver approached the intersection. All gaps in

opposing traffic were consistent with the RMOT method and

applied at each intersection where drivers were required to

make a permitted left-turn maneuver.

The opposing traffic consisted of three vehicles. As the

driver approached the intersection, a trigger in the simulation

located approximately 400 ft upstream of the left-turn stop bar

released the opposing traffic. At this time, the ﬁrst opposing

vehicle was located approximately 950 ft downstream of the

driver. The opposing vehicle was set to match the speed of

the driver. In this setup, the ﬁrst opposing vehicle approached

the intersection, almost mirroring the driver so that they

reached the intersection at approximately the same time. The

next two vehicles followed the initial opposing vehicle 3 and

10 sec after the ﬁrst vehicle; therefore, the driver observed a 3-

and a 7-sec gap after the initial opposing vehicle had passed.

Using two methods of opposing traffic allowed for an eval-

uation of opposing traffic impacts on driver comprehension

of PPLT signal displays. To determine the geographical

Figure 3-16. TTI driving environment simulator (DESi).

effects of drivers in the simulated environment, the last 93 driv-

ers at TTI completed the experiment observing the RMOT

opposing traffic. Therefore, experimental methodology and

the information observed by drivers in the simulation were

identical.

Video-Based Static Evaluation

After completing the driving portion of the study, drivers

were asked to participate in a static evaluation of PPLT

signal displays. The static evaluation was administered

using videocassette recordings of the screen captures for

the 12 PPLT displays. The driver was shown each display

for 30 sec and asked to choose one of four responses to the

traffic signal displays. Similar to the earlier Photographic

Driver Study research effort, the four potential responses

were as follows:

•

Go, you have the right-of-way.

•

Yield, then go if a gap in the opposing traffic exists.

•

Stop ﬁrst, then go if a gap in the opposing traffic exists.

•

Stop and wait for the appropriate signal.

Once drivers responded with one of the four possible

choices, they were asked to indicate their confidence in the

answer. Additionally, any comments made by the drivers

regarding the displays were manually recorded.

Conﬁrmation Study Sample Size

The approved research plan stipulated that 400 drivers

complete the driving simulator experiment, of which 200

subjects would be tested at UMass and 200 subjects would

be tested at TTI. To represent the general driving population,

four age groups of drivers were identiﬁed. In addition, an

attempt was made to include an equal number of male and

female drivers and a range of educational and ethnic back-

grounds. A total of 464 drivers participated in the study, of

which 432 completed all experimental elements, yielding

4,613 individual evaluated PPLT scenarios (5,230 PPLT sce-

narios with the static evaluation).

Of the 432 participants who completed the study, the data

for 316 became the focus of the report ﬁndings. The 316

drivers were a combination of the UMass drivers along with

the TTI drivers who completed the experiment using the pre-

viously described RMOT method.

Conﬁrmation Study Findings

Driver comprehension was determined from the distribution

of correct and incorrect responses for each of the selected

PPLT signal displays. Several categories of incorrect responses

were used to further evaluate these data. Analysis of variance

(ANOVA) statistical techniques were used to analyze all col-

lected data to evaluate drivers’ comprehension related to the 12

selected PPLT signal displays.

Left Thru Thru

Driver Movement: Permitted Left

Figure 3-17. Screen capture of typical intersection in simulator at UMass.

Driving Simulator Findings

Based on analysis of the 316 (223 at UMass and 93 at TTI)

driver evaluations (3,402 displays) that were obtained in a

procedurally equivalent manner, several key ﬁndings were

made. These ﬁndings are presented below as they relate to

geographical effects comparing drivers from Massachusetts

and Texas, effects of display type, driver demographics, ﬁrst-

time reactions, and learned behavior.

Statistical comparisons were made to evaluate whether

there were detectable geographic effects between drivers in

Texas and Massachusetts. When considering the 12 experi-

mental PPLT signal displays in each geographic location,

there were no statistically signiﬁcant differences between

locations. Because no statistically signiﬁcant differences were

found between drivers from Texas and Massachusetts, the

two databases were combined for further analysis.

With respect to display type, the following ﬁndings were

then made:

•

In the aggregate, the data showed a high level of com-

prehension with no variation between the different

PPLT displays tested. Drivers responded correctly 91%

of the time with no statistical difference between the 12

PPLT displays

•

The percentage of correct responses showed no statistical

difference in driver comprehension when the data were

cross-analyzed by permissive indication, display arrange-

ment, through indication, and location of the display.

•

There was no statistical difference in the percentage of

correct responses between permissive indication (circular

green indication, ﬂashing yellow arrow, circular green

indication/ﬂashing yellow arrow), signal display arrange-

ment (ﬁve-section cluster, four-section vertical, or ﬁve-

section vertical), PPLT display location (shared or exclu-

sive), or adjacent through indication (circular green

indication or circular red indication). Additionally, there

were no signiﬁcant differences by the various PPLT dis-

play components in terms of the percentage of fail criti-

cal responses.

Demographic factors such as sex, age, driver experience,

and education were all expected to inﬂuence driver under-

standing of the PPLT displays. When evaluated on the basis

of demographic effects, the following ﬁndings were made:

•

The data showed that the overall level of correct responses

to each permissive indication was not signiﬁcant when

analyzed by sex, age, number of miles driven annually,

or education.

•

Combined analysis of the data showed an interaction

effect between sex and age. In this analysis, there was a

statistically signiﬁcant difference between the three age

groups (ages under 24, ages 24–45, ages over 45) within

the female drivers and the percentage of correct responses.

•

There were no statistically signiﬁcant differences in the

percentage of correct or fail critical responses for the sex

demographic across the 12 PPLT signal displays evalu-

ated. Males and females had statistically equivalent lev-

els of comprehension.

•

Considering failure responses, the age demographic

resulted in statistically signiﬁcant differences. Drivers

over the age of 45 had signiﬁcantly fewer fail critical

responses. Overall, older drivers were more cautious in

the driving simulator experiment, often opting to wait for

all opposing vehicles to pass before completing the per-

missive left-turn maneuver.

•

Drivers who had driven between 10,000 and 20,000

miles in the previous year had signiﬁcantly more correct

responses and signiﬁcantly fewer fail critical responses

than those who drove fewer than 10,000 miles in the

previous year.

•

Education level of the drivers was not statistically sig-

nificant in determining comprehension levels in terms

of the percentage of correct responses. However,

PPLT Scenario 3 (five-section cluster in a shared loca-

tion with a flashing yellow arrow permissive indica-

tion and circular green through indication) was com-

prehended significantly more by drivers with only a

high school diploma than drivers with a higher educa-

tion level.

Another area of interest in evaluating drivers’ responses to

the PPLT displays presented in the simulator environment

was the manner in which drivers responded the ﬁrst time they

were presented a given PPLT display. Such an evaluation

was expected to provide insight as to the intuitive nature of

the PPLT display(s). In addition, comparison of results for

the same driver reacting to each permissive indication were

reviewed to assess whether any detectable learning occurred

as they proceeded through the virtual world. Each of the

PPLT signal displays was equally likely to be the ﬁrst PPLT

display evaluated because of the balanced design of the

experiment. Consequently, each of the 316 drivers in this

dataset had an equally likely chance of observing any of the

12 PPLT signal displays first. Findings of the evaluation

include the following:

•

Analysis of the ﬁrst observed PPLT display encountered

by each of the 316 drivers determined that the number

of correct responses was not signiﬁcantly different across

the 12 PPLT signal displays.

•

When reviewing driver responses to the ﬁrst observed

PPLT signal display, there were signiﬁcantly more fail

critical responses when using the ﬁve-section cluster in

a shared location with a ﬂashing yellow arrow permis-

sive indication and circular green through indication

than when using a ﬁve-section cluster in a shared loca-

tion with a circular green permissive indication and cir-

cular green through indication.

•

It was noted that the ﬁve-section cluster with circular

green permissive indication and circular green through

indication has been commonly used in both Massachu-

setts and Texas, and it was deemed reasonable to assume

that drivers had encountered this display prior to partic-

ipating in the experiment.

Static Evaluation Findings

Based on analysis of 436 driver evaluations (5,230 dis-

plays), the ﬁndings of the video-based static evaluation exper-

iment included the following:

•

Overall driver comprehension was high—83% of 5,230

scenarios were evaluated correctly.

•

The permissive indication resulted in statistically sig-

niﬁcant differences of correct and fail critical responses.

In contrast to the previously described driver simulator

findings, displays with the ﬂashing yellow arrow permis-

sive indication and the circular green/ﬂashing yellow

arrow simultaneous permissive indication had signiﬁ-

cantly more correct responses than displays with the cir-

cular green permissive indication. Displays with the cir-

cular green permissive indication were associated with

signiﬁcantly more fail critical responses than displays

with either the ﬂashing yellow arrow or circular green/

ﬂashing yellow arrow permissive indications.

•

PPLT displays with the four-section vertical display face

had a signiﬁcantly greater number of correct responses

compared with the ﬁve-section vertical and ﬁve-section

cluster displays. However, only the ﬂashing yellow arrow

permissive indication was evaluated in this display face,

and it is likely this combination that accounts for the

increased percentage of correct responses.

•

Scenarios that contained circular red indications in the

through movement signal heads resulted in a signiﬁ-

cantly lower percentage of correct responses to the PPLT

displays than when the circular green through indica-

tions were presented. PPLT displays associated with the

circular red through indication also resulted in signiﬁ-

cantly more fail critical responses. The research team

considered the lower correct response rate to result from

unfamiliarity. It is not common to use a permissive

green indication for the left-turn lane and a red indica-

tion for the through movement. This demonstrated again

that simultaneously displaying conﬂicting indications

causes confusion.

•

The location of the PPLT signal display (whether it was

a shared or an exclusive display) did not result in statis-

tically signiﬁcant differences in correct responses.

•

In contrast to the previously described driver simula-

tor findings, statistically significant differences were

observed within age, education, and driving experience

demographics. Drivers over the age of 45 had a signiﬁ-

cantly lower comprehension of the PPLT signal displays.

Drivers with only a high school diploma had a signiﬁ-

cantly lower comprehension than drivers with a higher

education level. Interestingly, those who drove between

10,000 and 20,000 miles in the previous year had sig-

niﬁcantly more correct responses than those who drove

fewer than 10,000 miles and those who drove over 20,000

miles the previous year.

Driving Simulator and Static Evaluation

Comparison Findings

The results of both the driving simulator and static eval-

uation were compared for each driver and for all drivers

combined. This analysis was completed to see how basic

driver comprehension (as demonstrated by the static evalua-

tion) compared with driver actions when presented the same

signal display scenario, but with the addition of all of the

dynamic elements associated with driving. Combining the

results of both the driving simulator experiment and video-

based static evaluation led to the following conclusions:

•

Driver comprehension in the simulator experiment was

signiﬁcantly higher than the static evaluation. The results

validate two important beliefs:

– Driver decision making during a left-turn maneuver

incorporates much more than comprehension of the

PPLT display. In fact, drivers often based their deci-

sion on opposing vehicle movements rather than the

PPLT display. Drivers who do not completely com-

prehend the meaning of the PPLT display use other

available information to make their decision.

– What drivers say they will do (based on comprehen-

sion of the PPLT display alone) and what they actu-

ally do in the driving environment are not always

consistent.

•

The biggest inconsistencies occurred for displays with

the circular green permissive indication. In the simula-

tor experiment, the four scenarios with the circular green

permissive indication resulted in fail critical responses

6% of the time. By contrast, the same four scenarios in

the static evaluation resulted in fail critical responses

19% of the time. Drivers are confused by the circular

green indication and often assume it provides right-of-

way during the permissive left-turn interval. This lack

of comprehension is not directly reﬂected in left-turn

crash statistics, because drivers compensate for com-

prehension deﬁciencies by considering other informa-

tion, such as following the lead vehicle, gauging oppos-

ing traffic, and choosing acceptable gaps.

•

In the simulator experiment, the through indication had

little effect on driver comprehension, while in the static

evaluation, the circular red through indication resulted

in lower comprehension levels. Based on driver com-

ments throughout the entire experiment, drivers often

did not observe the through indication in the simulator

but noticed the through indication in the static evalua-

tion. This behavioral pattern was explainable, because,

in a pure evaluation of comprehension, drivers search

for all available information. The only information in

the static evaluation was the left-turn and through indi-

cation. Therefore, if the driver was unsure about the

meaning of the left-turn indication, he or she used the

through movement and prior experiences in order to

decide how to act. Many times, this practice still led to

an incorrect response.

•

Comparing all types of responses in both of the experi-

ments, it can be said that many drivers base their left-turn

decision on surrounding traffic, speciﬁcally the opposing

traffic, instead of the signal indication. This was shown

to result primarily from a lack of driver understanding

of the indication. Under experimental conditions, sev-

eral drivers made left-turn decisions without any con-

sideration of the PPLT display.

Implications of Driver Comprehension Study

Findings

Collectively, the study data showed a high level of com-

prehension with no variation between the different PPLT dis-

plays tested. There was no statistical difference in driver com-

prehension when the data were cross-analyzed by permissive

indication, arrangement, through indication, and location of

the display. The lack of signiﬁcant differences documented in

this study is in itself a signiﬁcant ﬁnding. Given that the con-

ﬁrmation studies found no difference in display types, there is

evidence to suggest that the PPLT indication is only one of

many elements that the driver takes into account when mak-

ing left-turn decisions. This result also explains why low level

of comprehension related to the circular green permissive

indication is not consistent with left-turn crash frequencies.

FIELD IMPLEMENTATION STUDY

To address the lack of real-world applications of the ﬂash-

ing yellow arrow PPLT display, the project panel approved

the study of the ﬂashing yellow arrow display in the ﬁeld as

part of the NCHRP 3-54 (02) amended study. The project

panel thought that this task was needed before such a display,

if deemed “best understood,” could be recommended to the

NCUTCD.

Objective

The objective of the ﬁeld implementation study was to doc-

ument the implementation of the ﬂashing yellow arrow dis-

play, associated technical and non-technical issues, and safety

and cost implications associated with implementing this dis-

play. The following sections summarize the development and

administration of the implementation study, the ﬁeld data col-

lected, and the results.

Methodology

The ﬁeld implementation study was initiated, at the request

of the project panel, to collect actual ﬁeld data on the ﬂash-

ing yellow arrow display. The implementation study collected

before and after data relevant to the implementation of a

ﬂashing yellow arrow PPLT display (e.g., conﬂict data) to

record the safety performance associated with the ﬂashing

yellow arrow display. The implementation study also collected

field operational data, such as start-up lost time and satura-

tion ﬂow rate data, to quantify the operational impacts. The

implementation study monitored ﬁeld installation by the oper-

ating agency and documented techniques as well as the issues

resolved (e.g., control logic) for successful implementation

to be achieved. In addition, agency assistance was sought to

help quantify other implementation issues, such as ﬁeld per-

sonnel reaction, as well as labor, hardware, and software costs.

The following sections discuss background information on

how the study locations were identiﬁed and what was required

of the volunteering agencies and the research team.

Implementation Plan

The research team sought the participation of volunteer

agencies on a national basis. In August 2000, the research

team issued a Request for Proposal for an Implementation

Plan for submittal to volunteer agencies, identifying the proj-

ect goals and objectives, project requirements, and responsi-

bilities of both the volunteer agencies and the research team.

A copy of the Implementation Plan is included in Working

Paper No. 8 on the accompanying CD-ROM.

The research team was very speciﬁc about the characteris-

tics of the study and control intersections, as well as the dis-

play that would be used in the implementation study.

Each volunteer agency implementing the flashing yel-

low arrow display was required to ﬁrst request (and receive)

approval for experimentation of a traffic control device

though FHWA.

Characteristics of Study Intersections

The intersections selected for evaluation currently oper-

ated PPLT signal phasing and were considered typical inter-

sections containing no unique geometric or operational fea-

tures. Speciﬁc features sought in study intersections included

the following:

•

Right angle intersections with four approaches,

•

Exclusive left-turn lane(s) on the study approach,

•

Current use of the green arrow indication for the pro-

tected left-turn movement, and the circular green indi-

cation for the permissive left-turn indication

•

Relatively ﬂat approach grades,

•

Lane widths of 12 ft,

•

No on-street parking, and

•

No other variables that directly affect the left-turn move-

ment being evaluated.

Each agency that volunteered to participate in the study was

asked to identify at least three intersections for improvement

(i.e., installation of the new ﬂashing yellow arrow PPLT sig-

nal display). An additional three intersections were identiﬁed

within each study region that did not receive any improve-

ments during the study period. These intersections served as

control sites. Therefore, at least six study sites were requested

at each study region (study regions sometimes involved mul-

tiple agencies).

Proposed Flashing Yellow Arrow Display Face

The research team, in partnership with project panel and

Technical Advisory Group members, identiﬁed several dis-

play faces that research demonstrated as having good driver

understanding. Four possible PPLT signal displays were rec-

ommended for installation of the ﬂashing yellow arrow dis-

play at locations where there was an exclusive left-turn lane,

and the left-turn display was an exclusive display (not used by

the adjacent through movements). Those alternative displays

are shown in Figure 3-18 below. As shown in Figure 3-18, the

display could be implemented in four different conﬁgura-

tions, using three or four sections and horizontal or vertical

alignments. The three-section display options involve shared

use of a bi-modal section by the green and ﬂashing yellow

arrows. The three-section display face may be desired for

clearance purposes or for ease of implementation, if an exist-

ing three-section display face is available. The signal display

face could be mounted either vertically or horizontally.

In the exclusive display application, one, and only one, of

the four arrows was illuminated at any time. The ﬂashing yel-

low arrow was illuminated during the permissive phase when

traffic could turn after yielding to opposing through traffic

and/or pedestrians. The other three arrows were used for the

normal three-color exclusive left turn display. The red arrow

indication was displayed when a left-turn movement was pro-

hibited. The green arrow indication was displayed when the

left-turn movement could be made with no conﬂicting simul-

taneous vehicle or pedestrian movement (protected operation).

The steady yellow arrow indication was illuminated for a few

seconds as a clearance indication following both the green

arrow indication and the ﬂashing yellow arrow indication.

Implementation Study Findings

Participating Agencies

Beginning in August 2000, the research team contacted

over 35 agencies from across the United States. Of those 35

agencies, 9 submitted a request to FHWA for experimenta-

tion of the ﬂashing yellow arrow display. Two of those nine

study implementation locations withdrew from participation

because of controller logic issues or implementation proce-

dures not consistent with the project objectives.

1. 2. 3.

Indicates Flashing

Figure 3-18. Exclusive ﬂashing yellow arrow display faces.

The ﬁrst agency to respond to the volunteer solicitation was

Montgomery County, Maryland. In September 2000, Mont-

gomery County implemented the ﬂashing yellow arrow dis-

play at three intersections. Maryland’s participation in the

implementation study was subsequently followed by other

agencies as summarized in Table 3-8.

Other Agency Participation

In addition to the agencies listed in Table 3-8 that have

implemented the experimental ﬂashing yellow arrow dis-

play, three other agencies requested and received FHWA

approval to participate in the study. Two of the agencies were

ultimately unable to participate and one agency was not able

to implement within the time frame allotted to the research

project; further details are as follows:

•

The City of Kennewick, Washington, received FHWA

approval to implement and planned to do so by October

2001, but later withdrew from the study because of sig-

nal controller complications that could not be resolved

to the city traffic engineer’s satisfaction.

•

The City of Carson City, Nevada, submitted its request

for implementation to FHWA; however, the proposed

implementation included a supplemental sign explaining

the meaning of the ﬂashing yellow arrow display. After

discussions with the research team, it was decided that

FHWA would approve the installation if the supplemen-

tal sign was dropped from the implementation. As the

completion of the research project approached, Carson

City had not responded to FHWA and it remained unclear

how or if the City would proceed with implementation.

•

Snohomish County, Washington, received FHWA

approval for implementation and planed to implement at

one intersection; however, implementation was delayed

by traffic signal controller software issues that were not

resolved by the software manufacturer prior to comple-

tion of the research project.

Summary of Field Implementation Locations

Montgomery County, Maryland. Montgomery County

implemented the ﬂashing yellow arrow display in September

2000 at three intersections. A four-section vertical all-arrows

display was used. The ﬂashing yellow arrow indication was

tied to the opposing through green indication. The County

used Econolite NEMA controllers for local intersection man-

agement, and implementation required special external logic.

The County did not issue a media press release prior to imple-

mentation but did use variable message signs in advance of

the intersections for a period of 48 hr. The County has had

minimal citizen feedback during the more than 2 years of

deployment.

Tucson, Arizona. The City of Tucson, Arizona, was the

only agency to implement and then discontinue use of the

ﬂashing yellow arrow display. The City implemented ﬂash-

ing yellow arrow on May 30, 2001, at two intersections,

replacing PPLT displays with a ﬁve-section vertical display

face. The City of Tucson used the ﬂashing yellow arrow for

their right-turn overlaps for more than 10 years without any

problems. The City’s local intersection management soft-

ware was used with Econolite ASC/2 controllers. The City

used an external ﬂashing circuit tied to the ﬂashing yellow

arrow indication, which did not conﬂict with the controller’s

conﬂict monitor.

Within 1 week of implementation of the ﬂashing yellow

arrow, there was a crash at one of the study intersections. The

city manager directed the city traffic engineer to rescind par-

TABLE 3-8 Summary of implementation study sites

Agency Implementation Date

Number of

Implementation Sites

Montgomery County, Maryland September 2000

City of Tucson, Arizona May 2001

Jackson County, Oregon May 2001

Oregon Department of Transportation June 2001

City of Beaverton, Oregon April 2002

Broward County, Florida June 2002

3**

*One site in Jackson County met the NCHRP 3-54 study requirements. The County implemented non-conforming

displays at five other locations in the County with FHWA approval.

**In September 2002, Broward County received approval from FHWA to implement the flashing yellow arrow display

at two additional intersections bringing the total number of flashing yellow arrow implementation sites to five.

ticipation in the implementation study. The ﬂashing yellow

arrow was not reported to be directly linked to the cause of the

crash. Based on feedback from the operations engineer, the

intersection operations performed well during the time that

the ﬂashing yellow arrow display was in effect. The ﬂashing

yellow arrow display continues to be used for the right-turn

overlap movements, as it has for more than 10 years.

Jackson County, Oregon. Jackson County, Oregon, pur-

sued a different path with its implementation of the ﬂashing

yellow arrow display. The county traffic engineer for Jackson

County approached the research team to implement the ﬂash-

ing yellow arrow display in the county. The County had only

one intersection that could be converted from existing PPLT

to the ﬂashing yellow arrow display that met the research proj-

ect’s requirements. The County submitted the FHWA request

for experimentation and it was approved by FHWA. The sin-

gle intersection began operating with a ﬂashing yellow arrow

display using a four-section all-arrow vertical display face.

Since initial implementation in 2001, Jackson County

converted five existing exclusive (protected only) left-turn

operations to PPLT control with a ﬂashing yellow arrow dis-

play. However, in these installations, the County used a three-

section vertical display face—the center indication was used

for the yellow arrow clearance (following the circular green

indication) and the ﬂashing yellow arrow (permissive period)

indication. The County pursued this approach to eliminate

the costs of a new display face (four-section) and running

additional wire cable and to address vertical clearance issues.

The County submitted a request to FHWA for approval to

implement the three-section arrangement of the ﬂashing yellow

arrow display. The request was originally denied by FHWA

but later approved by FHWA after FHWA staff reviewed

video of the intersection showing the operation. The video

showed the adjacent through signals going to yellow at the

same time as the ﬂashing yellow arrow changed to a steady

yellow arrow in the same section. The ability of left-turn driv-

ers to see the through signals changing to yellow made it less

necessary for the change from the ﬂashing yellow arrow to

the steady yellow arrow to be positional. Accordingly,

FHWA issued a letter authorizing the County to implement

the three-section PPLT ﬂashing yellow arrow display. Fig-

ures 3-19 and 3-20 illustrate Jackson County’s retroﬁt of a

three-section display and the corresponding operation of the

signal.

The County received very positive feedback from the local

police department and citizens. The County’s local inter-

section management was a Type 170 controller with Wapiti

W4IKS ﬁrmware. Special command box logic was required

to implement the ﬂashing yellow arrow indication.

Woodburn, Oregon. The Oregon Department of Trans-

portation (ODOT) implemented the ﬂashing yellow arrow

display in June 2001 at two intersections in Woodburn, Ore-

gon. A four-section vertical all-arrows display was used. The

local intersection management was a Type 170 controller

with Wapiti W4IKS ﬁrmware. Special command box logic

was required to implement the ﬂashing yellow arrow indica-

tion. ODOT and the local City government (City of Wood-

burn) staff reported receiving minimal public feedback regard-

ing the ﬂashing yellow arrow displays.

Beaverton, Oregon. The City of Beaverton, Oregon,

implemented the ﬂashing yellow arrow display at three study

locations in April 2002 after having been granted authority

to implement the ﬂashing yellow arrow display from the City

Traffic Commission and the Mayor’s office. A four-section

all-arrows display was used. The local intersection manage-

ment was controlled by a Type 170 controller with Wapiti

W4IKS ﬁrmware. Special command box logic was required

to implement the ﬂashing yellow arrow indication. The City

was able to use the same basic logic developed by ODOT for

the Woodburn site in their implementation. At the time this

report was prepared, no problems had been experienced at

the intersections.

Broward County, Florida. Broward County was first

approached by the contractor to implement the ﬂashing yel-

low arrow display in May 2000. It was not until May 2002

that the County implemented the experimental display. This

2-year relationship between the contractor and implementing

agency is an example of the many challenges that agencies

faced in participating in this study as well as making future

changes. Implementation in Broward County was delayed by

many factors, most of which could be attributed to limited

County staff resources (time) and the desire to implement a

Red Arrow (Stop)

Bi-modal Flashing Yellow Arrow (Permissive Left-Turn)

and Clearance Steady Yellow Arrow

Green Arrow (Protected Left-Turn)

Figure 3-19. Jackson County, Oregon—3-section retroﬁt to ﬂashing

yellow arrow PPLT display.

different display from that requested by the research project

(which required an additional FHWA approval process).

Ultimately, the County implemented a ﬁve-section vertical

display, with the top two indications being a circular red indi-

cation and a circular yellow indication. The bottom three indi-

cations were all arrows with the ﬂashing yellow arrow con-

tained in the middle section as shown in Figure 3-21. It was the

County’s desire to clear the approach with all circular yellow

indications, rather than clearing the left-turn movement with a

yellow arrow indication, as the research team proposed.

In October 2002, the County received approval from FHWA

to implement the ﬂashing yellow arrow display at two addi-

tional intersections, bringing the total number of intersec-

tions with the ﬂashing yellow display to ﬁve.

Conﬂict Analysis Findings

As part of the ﬁeld implementation study effort, before and

after studies were conducted at each of the ﬂashing yellow

arrow implementation study sites. The local jurisdiction and/or

research team members videotaped 16 hr of before/after data.

The research team reviewed the videotapes to conduct a con-

ﬂict rate analysis.

Opposing

Through Signal

All Red

Protected Left-Turn

Change Interval

Permissive Phase

(Flashing Yellow Arrow)

Change Interval

(Steady Yellow Arrow)

Indicates Flashing

Figure 3-20. Operation of Jackson County, Oregon’s 3-section ﬂashing yellow arrow

PPLT display.

The before and after conﬂict analysis focused on conﬂicts

and events speciﬁcally related to the left-turn signal display.

This typically included driver hesitation on a permissive indi-

cation (circular green indication or ﬂashing yellow arrow) or

driver hesitation on a protected indication (green arrow).

Driver hesitation in either instance provided an indication

about drivers’ comprehension of the presented indication.

Left-turn conﬂicts, such as a driver failing to yield the right

of way to opposing traffic on a permissive indication, were

also observed because such conﬂicts provided the research

team with additional information about drivers’ comprehen-

sion of the particular left-turn indication. Although many con-

flicts were observed, this analysis focused on only those

related to the left-turn signal display. The results of the con-

ﬂict analysis are shown in Tables 3-9 and 3-10 for the before

and after analysis periods, respectively.

As shown by the two tables, there was little notable dif-

ference in the before and after conﬂict rate and no difference

that could be attributed to the change in PPLT display.

Follow-up headway information and corresponding ﬂow

rate were also evaluated for the permissive left-turn move-

ments. The headways between queued permissive left-turn

vehicles selecting the same gap in opposing traffic were

determined from the ﬁeld videos. Tables 3-9 and 3-10 also

show the follow-up headway ﬁndings. At many intersec-

tions, few instances with two permissive left-turns being

made in the same gap were observed, and the corresponding

headway and ﬂow rate information is based on only a small

sample size. Based on the data available, the change in PPLT

display to the ﬂashing yellow arrow had a negligible impact

on follow-up headway. A site-by-site summary of the analy-

sis ﬁndings is documented in Working Paper 8.

Field Observations During Flashing Yellow

Arrow Activation

In addition to the before and after data collection and analy-

sis, members of the research team were present as the ﬂash-

ing yellow arrow permissive indication was activated at six

of the site locations. The premise for these visits was to

observe whether there was any evidence of initial driver

confusion that might not be apparent from observation of the

after videos, particularly if drivers had learned the meaning

of the experimental display before the video data were col-

lected. Researchers present at the three Beaverton, Oregon,

ﬂashing yellow arrow installation sites observed no unusual

or signiﬁcant ﬁndings at the time the display was activated.

Similar results were observed at the three Broward County,

Florida, experimental sites as the ﬂashing yellow arrow indi-

cation was activated. The observations made at the time the

ﬂashing yellow arrow display was switched on are consistent

with both the before and after video data at these intersections,

indicating no signiﬁcant difference when changing from a cir-

cular green to a ﬂashing yellow arrow permissive indication.

The research team conducted on-site observations of ﬂashing

yellow arrow operation used in Montgomery County, Mary-

land, and Woodburn, Oregon. Conditions similar to those in

Beaverton, Oregon, and Broward County, Florida, were

observed.

Post-Implementation Survey of Volunteer

Agencies

A post-implementation survey was administered to the

agencies that participated in the implementation study. The

survey sought to identify issues that had to be dealt with to

implement the ﬂashing yellow arrow display, the cost to the

agency to implement the ﬂashing yellow arrow display, and

whether there was support within the agency and outside the

agency for the ﬂashing yellow arrow display.

Overall, each of the participating volunteer agencies expe-

rienced favorable results with the ﬂashing yellow arrow dis-

play implementation. The most commonly reported problem

was overcoming the current design of controllers and conﬂict

monitors. In all cases, the participating agencies had to use

either internal logic (e.g., command box in the Wapiti ﬁrm-

ware for the Type 170 controller) or some type of external

logic or relay device to implement the ﬂashing yellow arrow

display. These changes were necessary because the permis-

sive ﬂashing yellow arrow and circular green through move-

ment indications could not illuminate simultaneously (prior

to conversion, the circular green indication was used to com-

municate both the through and permissive movements). It is

assumed that new controller software and any signiﬁcant

upgrade of existing controller software will include this func-

tionality so that, over time, external logic will no longer be

needed. The special logic described above can be imple-

mented using a “logic box” external to the signal controller,

or with software enhancements in the signal controller.

The cost to implement the ﬂashing yellow arrow display

was relatively low (approximately $750 for new signal heads

and about 200 staff-hours total). All agencies received sig-

12345

Clear with adjacent

through lanes

Indicates Flashing

Figure 3-21. Broward County ﬂashing yellow arrow

display arrangement.

TABLE 3-9 Study intersection conﬂict data before ﬂashing yellow arrow implementation—Page 1 of 2

GB=Circular Green Indication

GA=Green Arrow Indication

Observational Safety Analysis (Display

Based)

Event

State

Observation

Period

Location Intersection

Site

Type

PPLT

Display

Data

Source

Hours of

Video

Reviewed

Conflicts

Hesitation on

Follow-Up

Headway

(veh/hr) /

Sample Size

Woodburn

Hwy 99 @

Hardcastle St.

Study

5-Section

Cluster

ODOT 16 0 0 0 N

Woodburn

Hwy 99 @ Lincoln

St.

Study

5-Section

Cluster

ODOT 17 0 1 0

1,321 (12)

Woodburn

Hwy 99 @ Young

St.

Control

5-Section

Cluster

ODOT 26 0 3 1

1,440 (10)

OR May 2001

McMinnville

Hwy 99 @ Baker

Creek Rd.

Control

5-Section

Cluster

ODOT 22 0 2 0

1,460 (9)

Beaverton

SW Allen Blvd. @

Wilson Ave.

Study

5-Section

Cluster

UMass 23 0 2 3

1,545 (40)

Beaverton

SW Allen Blvd. @

Menlo Dr.

Study

5-Section

Cluster

UMass 23 0 0 4

1,410 (32)

Beaverton

SW 125

St. @

Longhorn Ln.

Study

5-Section

Cluster

UMass 21 0 0 3

1,470 (18)

Beaverton

SW 72

St. @

Bonita St.

Control

5-Section

Cluster

UMass 24 1 10 12

1,790 (35)

Beaverton

SW Oleson St. @

Vermont St.

Control

5-Section

Cluster

UMass 20 0 1 0 1,750 (25)

OR April 2002

Beaverton

NW Murray Rd. @

Science Park Rd.

Control

5-Section

Cluster

UMass 16 1 1 0

1,661 (25)

AZ May 2001 Tucson

Ajo Way @ Park

Ave.

Study

5-Section

Vertical

Tucson 24 0 0 0

1,408 (15)

TABLE 3-9 Study intersection conﬂict data before ﬂashing yellow arrow implementation—Page 2 of 2

Observational Safety Analysis (Display

Based)

Event

State

Observation

Period

Location Intersection Site Type

PPLT

Display

Data

Source

Hours of

Video

Reviewed

Conflicts

Hesitation on

Follow-Up

Headway

(veh/hr) /

Sample Size

Mont. Cty

Cecil St. @ E.

Gude Dr.

Study

5-Section

Cluster

County 16 0 1 0

1,925 (64)

Mont. Cty

E. Randolph Rd @

Fairland Rd.

Study

5-Section

Cluster

County 16 0 0 0

1,809 (83)

MD August 2000

Mont. Cty

Montrose Rd. @

Tower Oaks Blvd

Study

5-Section

Cluster

County 16 1 1 0

1,800 (28)

Broward Cty

Broward Blvd. @

SW 69

Ave.

Study

5-Section

Cluster

UMass 8 1 3 2

1,470 (77)

Broward Cty

Coral Springs Dr.

@ Wiles Rd.

Study

5-Section

Cluster

UMass 8 1 0 13

1,598 (13)

Broward Cty

Sample Rd. @

Riverside Dr.

Study

5-Section

Cluster

UMass 8 1 6 4

1,933 (17)

Broward Cty

Broward Blvd. @

SW 70

Ave.

Control

5-Section

Cluster

UMass 8 3 0 9

1,735 (37)

Broward Cty

University Dr. @

Wiles Rd.

Control

5-Section

Cluster

UMass 8 0 7 1

1,711 (20)

FL January 2002

Broward Cty

Sample Rd. @

Rock Island Rd.

Control

5-Section

Cluster

UMass 8 0 2 1

1,931 (21)

GB=Circular Green Indication

GA=Green Arrow Indication

TABLE 3-10 Study intersection conﬂict data after ﬂashing yellow arrow implementation—Page 1 of 3

Observational Safety Analysis (Display

Based)

Events

State

Observation

Period Location Intersection

Site

Type

PPLT

Display

Data

Source

Hours of

Video

Reviewed Conflicts

Hesitation on

FYA

Hesitation on

Follow-Up

Headway

(veh/hr) /

Sample Size

Woodburn

Hwy 99 @

Hardcastle St.

Study

4-Section

Vertical

ODOT 12 0 3 0

1,375 (15)

Woodburn

Hwy 99 @

Lincoln St.

Study

4-Section

Vertical

ODOT 12 0 1 0

1,465 (19)

Woodburn

Hwy 99 @ Young

St.

Control

5-Section

Cluster

ODOT 12 0 1 (GB) 0

1,535 (45)

December

2001

McMinnville

Hwy 99 @ Baker

Creek Rd.

Control

5-Section

Cluster

ODOT 12 0 1 (GB) 0

1,962 (10)

Beaverton

SW Allen Blvd. @

Wilson Ave.

Study

4-Section

Vertical

UMass 16 0 4 0

1,650 (47)

Beaverton

SW Allen Blvd. @

Menlo Dr.

Study

4-Section

Vertical

UMass 16 0 1 0

1,415 (9)

Beaverton

SW 125

St. @

Longhorn Ln.

Study

4-Section

Vertical

UMass 15 0 1 0 N

Beaverton

SW 72

St. @

Bonita St.

Control

5-Section

Cluster

UMass 20 0 6 (GB) 9

1,728 (35)

Beaverton

SW Oleson St. @

Vermont St.

Control

5-Section

Cluster

UMass 18 0 0 0

1,760 (24)

OR June 2002

Beaverton

SW Murray Rd.

@ Science Park

Rd.

Control

5-Section

Cluster

UMass 18 1 0 1

1,603 (22)

FYA=Flashing Yellow Arrow Indication

GA=Green Arrow Indication

TABLE 3-10 Study intersection conﬂict data after ﬂashing yellow arrow implementation—Page 2 of 3

Observational Safety Analysis (Display

Based)

Events

State

Observation

Period Location Intersection

Site

Type

PPLT

Display

Data

Source

Hours of

Video

Reviewed Conflicts

Hesitation on

FYA

Hesitation on

Follow-Up

Headway

(veh/hr) /

Sample Size

Jackson Cty

Hamrick Rd. @ E.

Pine St.

Study

4-Section

Vertical

UMass 20 0 6 3

1,550 (61)

Jackson Cty

E. Pine St. @

Peninger St.

Study

3-Section

Vertical

UMass 17 1 0 8

1,440 (35)

Jackson Cty

Table Rock Rd.

@ Vilas Rd.

Study

3-Section

Vertical

UMass 19 0 1 1

1,584 (15)

OR June 2002

Jackson Cty

Stewart St. @

Columbus St.

Control

5-Section

Cluster

UMass 16 0 0 6 N

AZ June 2001 Tucson Ajo Way @ Park

Ave.

Study

5-Section

Vertical

Tucson 12 0 0 0

1,735 (14)

Mont. Cty

Cecil St. @ E.

Gude Dr.

Study

4-Section

Vertical

County 16 0 0 0

1,818 (38)

Mont. Cty

E. Randolph Rd

@ Fairland Rd.

Study

4-Section

Vertical

County 16 0 0 0

1,827 (70)

MD October 2000

Mont. Cty

Montrose Rd. @

Tower Oaks Blvd

Study

4-Section

Vertical

County 16 1 1 0

1,800 (42)

FYA=Flashing Yellow Arrow Indication

GA= Green Arrow Indication

TABLE 3-10 Study intersection conﬂict data after ﬂashing yellow arrow implementation—Page 3 of 3

Observational Safety Analysis (Display

Based)

Events

State

Observation

Period Location Intersection

Site

Type

PPLT

Display

Data

Source

Hours of

Video

Reviewed Conflicts

Hesitation on

FYA

Hesitation on

Follow-Up

Headway

(veh/hr) /

Sample Size

Broward Cty

Broward Blvd. @

SW 69th Ave.

Study

4-Section

Vertical

UMass 17

12 5

1,644 (31)

Broward Cty

Coral Springs Dr.

@ Wiles Rd.

Study

4-Section

Vertical

UMass 23

02 4

1,492 (12)

Broward Cty

Sample Rd. @

Riverside Dr.

Study

4-Section

Vertical

UMass 19

05 1

1,933 (35)

Broward Cty

Broward Blvd. @

SW 70th Ave.

Control

5-Section

Cluster

UMass 16

0 2 (GB) 2

1,593 (33)

Broward Cty

University Dr. @

Wiles Rd.

Control

5-Section

Cluster

UMass 20

0 2 (GB) 10

1,803 (24)

FL July 2002

Broward Cty

Sample Rd. @

Rock Island Rd.

Control

5-Section

Cluster

UMass 18

10 1

2,016 (20)

GA=Green Arrow Indication

FYA=Flashing Yellow Arrow Indication

niﬁcantly more positive than negative comment from the pub-

lic and from their own staffs.

Agency Feedback

In addition to the formal written survey response com-

ments, there was considerable anecdotal evidence that pro-

vided preliminary insight into the agencies’ perspectives on

use of the ﬂashing yellow arrow display. The overall response

to the ﬂashing yellow arrow display from traffic engineers

around the county was positive. In general, traffic engineers

expressed their approval of the ﬂashing yellow arrow display

because of the following:

•

The conﬁguration provided an exclusive signal display

for the left-turn control.

•

The indication was ﬂashing, which attracted more atten-

tion.

•

The traffic engineers had more operational control.

In general, local law enforcement agencies were support-

ive of the ﬂashing yellow arrow display; however, there was

some hesitation from city councils and county commission-

ers because of concern about trying something new that was

not formally part of the MUTCD.

Public Reaction

Public comments from citizens who experienced the ﬂash-

ing yellow arrow display in the ﬁeld were generally positive.

Several volunteer agencies reported receiving e-mails or writ-

ten letters from the motoring public with most, if not all, in

support of the ﬂashing yellow arrow display.

As part of the ongoing study activities, the research team

observed driver reaction to the ﬂashing yellow arrow display

immediately upon implementation in the ﬁeld. The drivers’

responses to the new displays suggest there was very little con-

fusion, with most drivers driving through the intersection as if

nothing were changed. Interestingly, at least one agency that

has implemented the ﬂashing yellow arrow display reported

that drivers waiting to make a permissive left turn now stop

behind the stop bar and wait for a gap in opposing traffic

(rather than entering the intersection before stopping).

Agencies Declining Participation

A significant challenge encountered by the research team

was the recruitment of volunteer agencies to participate in

the field implementation study. The research team solicited

involvement from 35 individual operating agencies, of which

26 declined for various reasons. Through the course of ver-

bal discussions with agency representatives and responses to

a follow-up survey that the research team conducted, the pri-

mary reason cited by agencies for not participating was their

lack of resources. Other explanations for not participating in

the implementation study included controller capabilities,

changes in management staff, and/or lack of technical and/or

managerial staff.

ENGINEERING ASSESSMENT

In the initial stages of the research, the research team devel-

oped an engineering assessment intended to explore the many

subjective elements affecting the use of traffic signal dis-

plays that were not measured through scientiﬁc experiment.

To perform the Engineering Assessment, the research team

considered practical issues related to how an agency would

actually implement a particular signal display or indication.

The discussion below presents the ﬁnal engineering assess-

ment that was developed after completion of the Driver Con-

ﬁrmation and Field Implementation studies.

Objective

The Engineering Assessment considered scientiﬁc and non-

scientiﬁc implementation issues in the following areas: safety,

operations, implementability, human factors, and versatility.

The updated engineering assessment identiﬁed objective and

subjective information needed in order to evaluate the signal

displays/indications. The assessment provided a thorough eval-

uation based upon sound engineering practice and the ﬁndings

of the various elements composing the research project.

Methodology

The research team identiﬁed and answered questions related

to safety, operations, implementability, human factors, and

versatility. Experiences of practicing traffic engineers and

basic engineering judgment were applied when appropriate.

Two signiﬁcant tables were developed in the engineering

assessment. The first table, Table 3-11, lists each of the

major categories (e.g., safety and operations) and each ques-

tion within each category. Each answer to each question was

rated from highest to lowest. By design, quantitative scores

were not associated with any display and/or indication (indi-

vidual questions were not weighted equally). Therefore, judg-

ment was used to assess which displays and/or indications

performed better than others. Many of the questions in Table

3-11 require further explanation or clariﬁcation of intent

and/or meaning. Each question in Table 3-11 is explored in

more detail in the full technical report documented in Work-

ing Paper No.1, Engineering Assessment, included in the

appendixes on the accompanying CD-ROM.

The second table, Table 3-12, identiﬁed allowable combi-

nations of placement, display faces, and left-turn phasing. This

TABLE 3-11 Engineering assessment evaluation matrix—Page 1 of 4

Solid Circular Green Indication

# Questions to be answered

Traditional

Five-

Section

Circular

Green +

Flashing

Arrow

Dallas

Display

Flashing

Yellow

Arrow

Comments

SAFETY

S-1

Does it fail safe? Is a misunderstanding of the indication likely

to result in a safe action?

The driver

simulation/confirmation study

has shown safer operation for

the flashing yellow arrow

S-2

Can the indication eliminate the yellow trap under all

operational and field condition

S-3 Can a red clearance be displayed after leading left

S-4 Can the start of permissive indication be delayed?

S-5

Does it avoid dilution of the safety or meaning of other

indication

S-6 Are traffic violations minimized?

S-7 Are accident costs reduced

S-8 Are conflicts reduced?

OPERATIONS

O-1

Does the indication increase total delay to the driver due to

indecision, increased start-up lost times, reduced travel

speeds, and/or lower saturation flow

rates?

O-2 Does the indication impact pedestrian movements?

All indications mean yield to

left-turn driver

O-3 Can the indication be used with lead/lag operation?

O-4 Does the indication impact opposing left-turning traffi

O-5

Does the indication allow the skipping of all side-street

phas

es?

O-6 Is the indication consistent with flashing indication

O-7

Does operating the intersection in flashing mode provide

negative consequences?

O-8 Does the indication lead to false starts or related driver errors?

Ranking scale: = highest/best; = lowest/worst

TABLE 3-11 Engineering assessment evaluation matrix—Page 2 of 4

Solid Circular Green

# Questions to be answered

Traditional

Five-

Section

Circular

Green +

Flashing

Arrow

Dallas

Display

Flashing

Yellow

Arrow

Comments

IMPLEMENTABILITY

I-1

Are there significant issues with installation? Can the

indication be placed to meet with the current MUTCD

requirements?

Circular Green + Flashing Arrow

and flashing yellow arrow will

require amendment of MUTCD

I-2

Are there issues with conversion of existing indications?

-Convert a signal currently using traditional five-section

indication?

-Convert a signal currently using permissive-only?

-Convert a signal currently using protected-only?

I-3

Are there legal issues to consider including the Uniform

Vehicle Code and state and local laws?

I-4

Does the signal indication permit maximum number of

signal phasing strategies?

HUMAN FACTORS

H-1

Is the indication universally understood? Does the

indication meet both a priori and ad hoc driver

expectancies?

H-2 Do drivers respond correctly to the information presented?

H-3

Do drivers accept the indication? Does the indication

increase driver workload, reduce conspicuity, or increase

driver error?

H-4 Are supplemental signs required for understanding?

H-5

Do drivers exposed to the "new" indication easily learn the

meaning?

H-6

Does the signal indication fail safe? What are the

consequences of a driver misinterpreting the signal

indication message?

Ranking scale: = highest/best; = lowest/worst

TABLE 3-11 Engineering assessment evaluation matrix—Page 3 of 4

Solid Circular Green

# Questions to be answered

Traditional

Five-

Section

Circular

Green

+Flashing

Arrow

Dallas

Display

Flashing

Yellow

Arrow

Comments

VERSATILITY

V-1 Does it allow permissive-only operation?

V-2 Does it allow protected-only operation?

V-3

Does it allow change between mode of operation by

time of day?

V-4 Can it be used on curved approaches?

V-5

Does it allow two far-side LT heads in customary

locations?

V-6 Does it allow use of any phase sequ

ence?

V-7 Is it applicable to right turns as well as left

V-8 Can it be used with span wire-mounted sign

als?

V-9

Can heads be in same location as permanent

protected-only heads for easy conversio

V-10

Can heads be in same location as permanent

permissive-only heads for easy conv

ersion?

V-11

Does it allow use of all of the opposing through green

time for permissive turns?

V-12

Can it be used when the left-turn lane is shared with

through traffic?

V-13

Can permissive, turning traffic proceed legally without

stopping?

V-14

Could it replace all current standard and non-standard

PPLT indicatio

ns?

V-15

Can it be used where there is no adjacent through

movem

ent?

V-16

Can it be used where the adjacent through movement

is unsignalize

Ranking scale: = highest/best; = lowest/worst

TABLE 3-11 Engineering assessment evaluation matrix—Page 4 of 4

Solid Circular Green

# Questions to be answered

Traditional

Five-

Section

Circular

Green +

Flashing

Arrow

Dallas

Display

Flashing

Yellow

Arrow

Comments

VERSATILITY (continued)

V-17

Can it be used when the left-turn slot is physically

separated or on different alignment than through lane

(wide median, etc.)?

V-18

Can the signal indication be placed horizontally or

vertically in the same arrangement?

V-19 Does it work under all preemption scenarios?

V-20

Does it avoid the yellow trap situation under all

circumstances?

V-21

Can the permissive indication be easily applied to other

than PPLT situations?

V-22

Will practitioners likely use the indication if made the

standard, or allowed alternat

Ranking scale: = highest/best; = lowest/worst

table identiﬁed the practical considerations regarding display

or indication, based on (1) whether the display/indication

could be used in a shared display face or an exclusive display

and (2) whether the display/indication could be used in a

lead-lead, lag-lag, or lead-lag left-turn signal operation.

Findings

The Engineering Assessment material presented in Table

3-11 focused on key issues surrounding safety, operations,

implementability, human factors, and versatility. The ﬁnd-

ings of this assessment are highlighted below:

•

Within the category of safety, the assessment ﬁndings

suggest that the ﬂashing yellow arrow display offered the

highest level of safety, followed by the circular green

indication using the Dallas Display.

•

Within the category of operations, the circular green

indication display using the Dallas Display performed

similarly to the ﬂashing indications, with the remaining

circular green indication displays performing at lesser

levels.

•

Within the category of implementability, the circular

green indication was identiﬁed as being easiest to imple-

ment. This ﬁnding reﬂects the status of the circular green

indication as being the current standard for most agen-

cies. By comparison, the ﬂashing yellow arrow display

was found to be easier to implement than a standard cir-

cular green indication in an exclusive left-turn display

and, overall, nearly equal to the implementability of a cir-

cular green indication display using the Dallas Display.

•

Within the category of human factors, the ﬂashing yel-

low arrow display was found to rank the best.

•

Within the category of versatility, the ﬂashing yellow

arrow display was clearly shown to offer the most ver-

satility while the standard circular green display shared

with a through lane offered the least.

Table 3-12 identiﬁed allowable combinations of place-

ment, display face, and left-turn phasing and highlights the

fact that only the circular green indication and the ﬂashing

yellow arrow can be used in both the shared display and in

the exclusive left-turn display. The circular green indication

has some limitations in the shared display placement (e.g., it

cannot be used for lead-lag phasing, must serve both lagging

lefts at the same time, and has yellow trap potential). Conse-

quently, the comparison provided in Table 3-12 points out

that the ﬂashing yellow arrow appears to be the most widely

applicable option.

TABLE 3-12 Allowable combinations of placement, display face, and phasing for potential display type

D I S P L A Y T Y P E

Placement

Indication

Arrangement Phasing

Traditional Five-

Section Circular

Green

Solid Circular

Green

– Dallas Display

Flashing

Yellow

Arrow

Lead-Lead Lefts

Y N Y

Lag-Lag Lefts

Five-section

Cluster

Lead-Lag Lefts

N N Y

Lead-Lead Lefts

Y N Y

Lag-Lag Lefts

Five-section

Vertical

Lead-Lag Lefts

N N Y

Lead-Lead Lefts

Y N Y

Lag-Lag Lefts

Shared

Indication with

Through

Five-section

Horizontal

Lead-Lag Lefts

N N Y

Lead-Lead Lefts

Y Y Y

Lag-Lag Lefts

Five-section

Cluster

Lead-Lag Lefts

N Y Y

Lead-Lead Lefts

Y Y Y

Lag-Lag Lefts

Five-section

Vertical

Lead-Lag Lefts

N Y Y

Lead-Lead Lefts

Y Y Y

Lag-Lag Lefts

Exclusive

Indication

Five-section

Horizontal

Lead-Lag Lefts

N Y Y

Footnotes: 1. Assumes that the yellow arrow indication serves to both clear the green arrow indication and flash for the permissive interval. Use the bi-modal in

the bottom and use the yellow for the clearance.

2. Works only if serve both lagging lefts at the same time, otherwise a yellow trap may be created.

CHAPTER 4

DISCUSSION

The NCHRP 3-54 (02) research project looked at all pos-

sible PPLT displays and shortened the list of PPLT displays

to a select few. The research team collected sufficient data to

narrow the list of potential PPLT displays after extensive dis-

cussions with the project panel members.

The ﬁnal study tasks of the project were focused on an eval-

uation of the ﬂashing yellow arrow display in comparison with

the circular green indication, the MUTCD standard. The appli-

cation of a driver simulator provided an in-depth analysis of

driver understanding of the displays studied. When compared

with PPLT displays with a circular green permissive indica-

tion, the ﬂashing yellow arrow indication was shown to exhibit

superior qualities in some analyses and equal qualities in oth-

ers. In only a few cases did the ﬂashing yellow arrow indica-

tion not yield a higher level of understanding as compared with

the circular green indication, such as when the display was

“ﬁrst observed.” The laboratory-based research techniques did

not universally demonstrate that the ﬂashing yellow arrow

indication represents a far superior indication to the circular

green indication. The ﬁndings of the laboratory-based research

support the position that the ﬂashing yellow arrow indication

represents the best alternative to the circular green indication.

Complementing the laboratory-based research techniques

was the in-the-ﬁeld-based implementation study in which the

ﬂashing yellow arrow indication was deployed in the real

world and thoroughly studied. The Implementation Study

results demonstrated that the ﬂashing yellow arrow indication

was well understood in almost all deployment cases. Analy-

sis of the conﬂict data suggests that the ﬂashing yellow arrow

indication is at least as safe as the current MUTCD standard

(circular green indication). Where deployed, the ﬂashing yel-

low arrow indication was preferred by almost all of the traffic

engineers, ﬁeld technicians, and citizens when compared with

the circular green PPLT indication. The Implementation Study

identiﬁed hardware and software issues related to implement-

ing the ﬂashing yellow arrow display that would need to be

addressed before any widespread deployment could be con-

ducted. Through the Engineering Assessment, the ﬂashing yel-

low arrow display was identiﬁed in three of the ﬁve categories

as “best” and near equal to the circular green indication in the

remaining two categories.

All PPLT displays and/or indications have been researched

thoroughly. The ﬂashing yellow arrow indication/display was

found to result in a high level of understanding and a lower

fail critical rate as compared with the circular green indica-

tion. The ﬂashing yellow arrow display offers more versatile

ﬁeld application features (e.g., the display can be (1) operated

in various operational modes by time of day and (2) imple-

mented on any signal mount and intersection conﬁguration)

as compared with the circular green indication. To that end,

the recommendations presented in Chapter 5 focus on the

flashing yellow arrow display as preferrable to the circular

green indication.

The various research activities completed throughout the

course of the project consistently supported the use of the ﬂash-

ing yellow arrow indication as an alternative to the MUTCD

circular green permissive indication. The Engineering Assess-

ment summarized many of the aspects by which the ﬂashing

yellow arrow display was found to be equal to or superior to

existing PPLT displays. Similarly, the Conﬁrmation Study

and the Field Implementation Plan demonstrated the viability

of the flashing yellow arrow display, as well as its perfor-

mance and acceptance upon introduction to the motoring pub-

lic. The remainder of this chapter outlines the research team’s

basis for the recommendations documented in Chapter 5.

FINDINGS OF THE CONFIRMATION STUDY

The results of the Conﬁrmation Study showed the following:

•

The ﬂashing yellow arrow display was as well under-

stood (measured in terms of correct responses to ques-

tions presented) as the circular green indication.

•

There was no signiﬁcant difference in the motoring pub-

lic’s correctly interpreting the meaning of the ﬂashing yel-

low arrow indication as compared with the circular green

indication. The data demonstrated that drivers’ under-

standing of the ﬂashing yellow arrow display increased

with exposure.

•

The flashing yellow arrow display showed a higher

fail-safe response as compared with the circular green

indication.

•

The conﬂict studies demonstrated that drivers interpret

the ﬂashing yellow arrow display correctly.

FINDINGS OF THE FIELD IMPLEMENTATION

STUDY

The results of the Field Implementation Study demon-

strated the following:

•

The flashing yellow arrow display was successfully

implemented in the field with relatively little or no

technical or political issues. Post-implementation pub-

lic testimony almost unanimously supported use of the

experimental display. For example, the Jackson County,

Oregon, traffic engineer received the following voice

message referring to the new ﬂashing yellow arrow PPLT

display: “I’m [name removed for privacy] calling to com-

ment about your ﬂashing yellow arrow on a turn lane. I

think it’s the best thing since sliced bread. It is easy to

understand, makes traffic flow much more smoothly,

makes a lot better sense, you don’t sit there waiting to

make a left turn when nobody’s coming. The one in

downtown Medford confused me, the green light and

the green arrow [referring to the standard MUTCD ﬁve-

section cluster PPLT display]. I’m confused with that;

but with the ﬂashing yellow light—I think it’s wonder-

ful. Don’t change a thing.” Letters and additional testi-

mony are presented later in this chapter.

•

Most practicing traffic engineers contacted during the

study endorse the ﬂashing yellow arrow display.

•

The ﬁeld data support high understanding of the ﬂash-

ing yellow arrow display.

In addition to the many technical results of the research

elements supporting the flashing yellow arrow display, it is

also appropriate to reflect on the practical elements of why

the flashing yellow arrow display provides equal or supe-

rior performance as compared with the circular green PPLT

indication.

OPERATIONAL ADVANTAGES

There are many reasons some modes of left-turn operation

(e.g., permissive only, protected only, protected permissive,

lead/lag, and so forth) are chosen or the mode is changed dur-

ing the day. In most cases, however, it is for operational effi-

ciency, such as to increase the left-turn capacity, improve

traffic progression through coordinated signals, or reduce the

duration required for the protected phase, including full sup-

pression of the protected phase. Each of the various control

modes of left-turn operation has certain advantages and dis-

advantages. It has also been demonstrated that the current

MUTCD standard circular green indication display presents

safety problems.

As a result, traffic engineers have had to balance efficiency

and safety when designing and operating intersections. Inno-

vative traffic engineers have searched for ways to maintain

intersection safety and improve intersection operations, as

demonstrated by the multiple non-standard PPLT displays

currently in use within the United States. Given this discus-

sion as a preamble, the ﬂashing yellow arrow display, in an

all-arrow display face, supports the following modes of left-

turn operations.

Protected-Only Operation

The ﬂashing yellow arrow display can be operated in a

protected-only mode. In the protected-only mode, the ﬂash-

ing yellow arrow permissive indication is not used.

Protected-Permissive Operation

The flashing yellow arrow display is “logically tied” to

the green output of the opposing through movement to avoid

the yellow left-turn trap. Logically tying a phase means that

the traffic signal control software outputs the ﬂashing yel-

low arrow indication (permissive turn) only during the green

interval of the opposing through phase. This ensures that when

the permissive display terminates, the opposing through phase

will terminate simultaneously.

The original research problem statement addressed the

yellow trap issue as a major concern and it was, in fact, fun-

damental to initiation of this research project. As shown in

Figure 4-1, logically tying the ﬂashing yellow arrow indica-

tion to the opposing through green indication completely elim-

inates the yellow trap, even when lead/lag phasing is used or

when side street phases are skipped and the leading left-turn

phase now follows the through phase.

Permissive-Protected Operation

Historically, most traffic engineers with lagging protected

operations serve both left-turn phases for the same duration

so as to not create the yellow trap. Because the ﬂashing yel-

low arrow display is tied to the opposing through movement,

thereby eliminating the yellow trap, this operation becomes

much more efficient. This operation is especially useful for

skipping the protected-only sequence when there is inade-

quate vehicular demand.

Permissive-Only Operation

The permissive left-turn operation allows the drivers to exe-

cute left turns when gaps occur within the opposing through

movement. The ﬂashing yellow arrow display allows the sig-

nal operation to change modes during the day in any way, or

for longer periods of time, if desired. A key advantage to this

ﬂexibility is that the signal operation could, for example,

change from an eight-phase protected-only operation during

peak hours, to an eight-phase protected/permissive operation

during the day, to a simple two-phase permissive-only signal

operation during low volume conditions. It is also feasible

for one direction to operate protected-permissive while the

opposite approach uses a different mode. All combinations

are feasible and can be selected to optimize the operational

efficiency as conditions change.

OTHER CONSIDERATIONS

Supplemental Display Arrangements

The ﬂashing yellow arrow display supports all of the modes

of left-turn operation as identiﬁed above and can be imple-

mented in states that require supplemental signal faces. States

such as California use supplemental signal faces that are nor-

mally located on the far left side of the intersection. This type

of operation is not possible with the Dallas display because

of its need for optical shielding.

Right-Turn Overlap Display

The ﬂashing yellow arrow display would solve the prob-

lem of having to prohibit the conﬂicting U-turn when oper-

ating a right turn overlap during a side street left-turn phase.

Similarly, the ﬂashing yellow arrow display could be used

for drivers who might have to yield to pedestrians while turn-

ing right from approaches with no through movement and in

Figure 4-1. Flashing yellow arrow logical link.

Opposing

Through Signal

All Red

Protected Left-Turn

Change Interval

(End Protected Left-Turn)

Permissive Phase

Through Traffic

Change Interval

Opposing Through

Phase Still Running

Indicates Flashing

cases where a circular green indication cannot be displayed

(e.g., because of a one-way opposite approach).

UNIVERSAL APPLICATION

The flashing yellow arrow is the only display that pro-

vides a universal solution. It can be used at every intersec-

tion, no matter how unusual. Other PPLT displays cannot be

used in some situations, such as when the signal arrange-

ments are mounted on span wires, when dual far-side signal

arrangements are needed (e.g., in California), when there is

no adjacent through movement, when the approach is

curved, or when the turn must be held (not even permissive)

during railroad preempt although both through movements

can proceed.

Figure 4-2a. Letter in support.

A NEED FOR CHANGE

There was much discussion within traffic engineering

forums that left-turning drivers need to know that there is a

change in right-of-way permissions. It has been proposed

that a new display was needed that would change in color,

mode, position, and shape. That way, when in operation,

drivers would be alerted to new information on the signal

face so they could align their movements with the correct

right-of-way permissions.

The ﬂashing yellow arrow display, in a four-section dis-

play face, provides a change in color (green to yellow), a

change in mode (steady to ﬂashing), a change in position (a

lateral change in display placement and a vertical change in

the display arrangement), and a change in shape (circular

green to a yellow arrow).

Figure 4-2b. Letter in support.

PUBLIC SUPPORT FOR THE DISPLAY

The reaction from the public toward the ﬂashing yellow

arrow has been positive. The Field Implementation Study

identiﬁed all seven implementing agencies as receiving

mostly positive feedback from local citizens. For example, as

found in Jackson County, Oregon, most of those in favor of

the ﬂashing yellow arrow display wanted the displays retro-

ﬁtted to other intersections where protected-only left turn

phasing was in operation. Figures 4-2a and 4-2b contain let-

ters submitted by local citizens supporting use of the ﬂashing

yellow arrow displays.

Even outside the auspices of the NCHRP 3-54 project,

public reaction has been overwhelmingly favorable. For

example: the City of Reno, Nevada, received tremendous

public response immediately upon converting several inter-

sections to using the flashing yellow arrow display. A total

of 66% of the callers favored the new display, 27% were

neutral (asked for information), and 7% were negative (they

did not see a reason for the change). Although there were

some negative feelings about changing to the flashing yel-

low arrow display, such a response can be expected anytime

there is change in operation.

IMPLICATIONS FOR NONCONFORMING

PPLT DISPLAYS

This research project was initiated at the request of the

Signals Technical Committee of the National Committee on

Uniform Traffic Control Devices. A goal of the study was to

identify the “best” PPLT display that would lead to the

determination of a national standard, whether that display was

the circular green indication or some other indication. This

study has shown that the ﬂashing yellow arrow indication is

better understood than the circular green indication when used

for PPLT operation. The recommendations identiﬁed in Chap-

ter 5 identify the ﬂashing yellow arrow display as the preferred

alternative to the circular green indication. The future status of

other displays/indications currently in use around the United

States that do not conform to either the circular green indi-

cation or the new proposed alternative indication will need to

be determined by FHWA.

CHAPTER 5

RECOMMENDATIONS

Through the course of this study, the individual study tasks

such as the Agency Evaluation, Photographic Driver Study,

Traffic Operations Study, Traffic Conﬂict Study, Crash Analy-

sis, Driver Simulation Study, Field Implementation Study, and

Engineering Assessment have provided the research team with

a tremendous amount of data. Chapters 2, 3, and 4 of this

report have presented these data and captured key ﬁndings

derived from the study task activities. This chapter provides

the research team’s recommendations based on the collective

project ﬁndings.

Based on the ﬁndings of this project, the research team

makes the three recommendations identiﬁed below.

RECOMMENDATION #1: INCORPORATE

FLASHING YELLOW ARROW DISPLAY

INTO THE MUTCD

It is recommended that the ﬂashing yellow arrow display

be included in the MUTCD as an allowable alternative display

to the circular green indication when used in PPLT control/

operation.

RECOMMENDATION #1A: DISPLAYS

The four-section, all-arrow display face should be the only

display allowed. The only display that justiﬁes an exception

to this recommendation is the three-section display face with

bi-modal lens. The three-section display face with bi-modal

lens should also be allowed because it operates in the same

way as does the four-section display face. Only one indica-

tion shall be illuminated at any time. Allowed variations could

be as follows:

•

The agency may use a circular red indication in lieu of

the red arrow.

•

A three-section all-arrow display using the center sec-

tion as a change interval and the permissive interval is

allowed, provided that all yellow change intervals for

the approach are initiated simultaneously.

•

A ﬁve-section display may be used.

Figure 5-1 illustrates two potential options for retroﬁtting

a ﬁve-section display to incorporate the ﬂashing yellow arrow.

The five-section heads shown in Figure 5-1 would serve

only as the left-turn signal and would have to be shielded

from the adjacent through movement lanes. The adjacent

through movement indication would have to be provided on

an exclusive display. Furthermore, only one indication in the

ﬁve-section display would be illuminated at any time.

As shown in Figure 5-1, the five-section display could

incorporate an all-arrow arrangement or a yellow circular

indication and a red circular indication. If the red circular indi-

cation is used, a yellow circular indication or yellow arrow

indication could be used for the permissive phase change inter-

val. However, if a red arrow indication is used, the yellow

arrow indication must be provided for the change interval. In

either case, the protected left-turn elements of the display

would remain in their current location and the ﬂashing yellow

arrow would replace the current green circular permissive

indication in the bottom right corner of the display.

RECOMMENDATION #1B: LOCATION

The ﬂashing yellow arrow operation shall only be used in

an exclusive signal arrangement. The left-turn signal face

should be placed over the left turn lane.

RECOMMENDATION #1C:

SUPPLEMENTAL SIGNS

Supplemental signing is not warranted with a ﬂashing yel-

low arrow display. Use of supplemental signing is optional. If

a supplemental sign is deemed necessary, the following sign

should be used: YIELD ON FLASHING YELLOW ARROW

or YIELD ON FLASHING YELLOW [Symbolic Arrow].

RECOMMENDATION #1D: PHASING

When used for left-turn treatments, the flashing yellow

arrow shall be tied to the opposing through-green indication/

display. There may be a delay in the illumination of the ﬂash-

ing yellow arrow display.

RECOMMENDATION 2: CONDUCT FOLLOW-UP

STUDY

It is recommended that a follow-up study be conducted for

this project. The follow-up study should be conducted after

there has been sufficient time for an implementation trial

period for agencies currently participating in the ﬁeld imple-

mentation as well as for any additional agencies that may

choose to implement based on the ﬁndings of this research

effort. Sufficient time should also be allowed to acquire before

and after crash data at the study intersections and correspond-

ing control sites. Given the difficulties encountered in obtain-

ing useful crash data for this research project, it is strongly

encouraged that any participating agencies carefully compile

and maintain crash data records for the study intersections.

Once initiated, the follow-up study should

•

Analyze all available crash data for the experimental

ﬂashing yellow arrow displays implemented as part of

this study;

•

Identify whether the ﬂashing yellow arrow display should

become the only display allowed in the MUTCD for a

protected-permissive left-turn operation;

•

Identify whether, if the ﬂashing yellow arrow is selected to

become the only display allowed for PPLT, the MUTCD

should also be changed to add the following prohibition:

“For Permissive Only Mode operation, a signal face dis-

playing a circular green indication shall not be located

directly over or in line with a left-turn lane”; and

•

Identify an implementation plan.

RECOMMENDATION 3: RESTRICT USE

OF FLASHING RED INDICATIONS

The use of the ﬂashing red indication should only be

implemented where an engineering study has identiﬁed that

all drivers must come to a complete stop before proceeding

on the permissive interval.

Flashing

Figure 5-1. Potential ﬂashing yellow arrow

retroﬁt to ﬁve-section cluster display (exclusive

left-turn display).

REFERENCES

1. American Traffic Safety Services Association/Institute of Trans-

portation Engineers/American Association of State Highway and

Transportation Officials/Federal Highway Administration, “Uni-

form Traffic Control Devices,” Washington, DC (2001).

2. Fambro, D.B., Messer, C.J., and Woods, D.L., “Guidelines for

Signalized Left-Turn Treatments,” FHWA Report IP-81-4,

FHWA, U.S. Department of Transportation, Washington, DC

(1981).

3. Fambro, D.B, Gaston, G.D., and Hoff, C.M., “Comparison of

Two Protected-Permitted Lead-Lag Left-Turn Phasing Arrange-

ments,” Research Report 989-1F, Texas Transportation Institute

(1991).

4. King, G.F., NCHRP Research Results Digest 97: Guidelines for

Uniformity in Traffic Signal Design Conﬁgurations. Trans-

portation Research Board, National Research Council, Wash-

ington, DC (1977).

5. Institute of Transportation Engineers, Washington Section of

District 6, “Flashing Yellow Protected-Permissive Signal Eval-

uation,” Final Report (1985).

6. Kua, C., “Left-Turn Arrow: Flashing versus Non-Flashing,”

Written correspondence to Brian Latte, Project 158, Uniform

Traffic Signals Displays, Chairman (1991).

7. Mintz, S., The Canadian Association of Optometrists. Written

correspondence to Alfred Guibert on color blindness and traf-

ﬁc signal display (1992).

8. Holowachuk, L. M., Director of Highway Safety, “Impact

Statement of New Signal Standards,” Prepared for the National

Council on Uniform Traffic Control Devices, Canada (1993).

9. Pitzinger, P., “Guidelines for Signal Light Installations,”

Research Society for Highways and Traffic Workshop Team

for Traffic Strategy and Traffic Security: Maastricher Straße,

Germany (1981).

10. Pitzinger, P. and Hasler, H., “The Flashing Yellow Arrow,”

Confederate Department of Traffic and Energy Economics,

Federal Office for Road Construction, Germany (April 1998).

11. Pitzinger, P., “Three Years of Experience with Traffic Acci-

dents with the Flashing Yellow Arrow,” Zurich City Police,

Zurich, Germany (1990).

12. Brebner, J., Ward, L., and Taplin, J., “Study of Comprehen-

sion of Turn Right With Care (TRWC) Message at Intersec-

tions and Related Crash Patterns,” University of Adelaide,

Australia (2002).

13. Brookes, J., Collins, R., and Haenel, H., “Study of MUTCD

and City of Dallas Permitted/Protected Lead/Lag Left-Turn

Phasing,” Internal Paper Austin, Texas, Texas Department of

Transportation (1990).

14. Collins, R., “ A comparative analysis of left-turn delay associ-

ated with two difference lead lag phasing arrangements,” Mas-

ters Thesis, The University of Texas at Austin (1988).

15. Albert Grover & Associates, “Minimize Delay, Maximize Pro-

gression with Protected Permissive Lead/Lag Phasing,” Tech-

nical Workshop, Inland Empire Section of ITE (1995).

16. Highway Capacity Manual 2000, Transportation Research

Board, National Research Council, Washington, DC (2000).

17. Robertson, H.D., Hummer, J.E., and Nelson, D.C., Manual of

Traffic Engineering Studies, ITE, Prentice Hall, Englewood

Cliffs, NJ (1994).

18. Parker, M.R. Jr., and Zegeer, C.V., “Traffic Conﬂict Tech-

niques for Safety and Operations: Engineers Guide,” FHWA

Report IP-88-026, FHWA, U.S. Department of Transportation,

Washington DC (1988).

19. Parker, M.R. Jr., and Zegeer, C.V., “Traffic Conﬂict Tech-

niques for Safety and Operations: Observers Manual,” FHWA

Report IP-88-027, FHWA, U.S. Department of Transportation,

Washington DC (1988).

20. “Recommended Warrants for the Use of Protected/Permissive

Left-Turn Phasing,” Technical Committee Project 4A-30, Insti-

tute of Transportation Engineers, Washington DC, (1994). Man-

ual on Uniform Traffic Control Devices, Federal Highway

Administration, Washington, DC (1988).

21. Hummer, J. E., Montgomery, R. E., and Sinha, K. C., “Guide-

lines for Use of Leading and Lagging Left-Turn Signal Phas-

ing,” Transportation Research Record 1324 (1991).

22. Upchurch, J., “Comparison of Left-Turn Crash Rates for Dif-

ferent Types of Left-Turn Phasing.” Transportation Research

Record 1324 (1991).

23. JHK & Associates, “Signal Displays for Left Turn Control.”

Task A Report (Unpublished), prepared for Federal Highway

Administration, Contract # DTFH61-85-C-00164 (August

1988).

24. Hauer, E, J.C.N. Ng, and J. Lovell, “Estimation of Safety at

Signalized Intersections, Transportation Research Record

1185 (1988).

25. Bonneson, J.A. and McCoy, P.T., “Evaluation of Protected/

Permitted Left-Turn Traffic Signal Displays,” Nebraska Depart-

ment of Roads, Research Report No. TRP-02-27-92 (1993).

26. Institute of Transportation Engineers, Washington Section of

District 6, “Flashing Yellow Protected-Permissive Signal Eval-

uation,” Final Report (1985).

27. Agent, K.R, “Development of Warrants for Left-Turn Phas-

ing,” Research Report 456, Kentucky Department of Trans-

portation (1976).

28. Agent, K.R., “An Evaluation of Permissive Left-Turn Phas-

ing,” Research Report 519, Kentucky Department of Trans-

portation (1979).

29. Agent, K.R., “Guidelines for the Use of Protected/Permitted

Left-Turn Phasing,” Research Report UKTRP-85-19, Kentucky

Transportation Research Program (1985).

Abbreviations used without deﬁnitions in TRB publications:

AASHO American Association of State Highway Officials

AASHTO American Association of State Highway and Transportation Officials

APTA American Public Transportation Association

ASCE American Society of Civil Engineers

ASME American Society of Mechanical Engineers

ASTM American Society for Testing and Materials

ATA American Trucking Associations

CTAA Community Transportation Association of America

CTBSSP Commercial Truck and Bus Safety Synthesis Program

FAA Federal Aviation Administration

FHWA Federal Highway Administration

FMCSA Federal Motor Carrier Safety Administration

FRA Federal Railroad Administration

FTA Federal Transit Administration

IEEE Institute of Electrical and Electronics Engineers

ITE Institute of Transportation Engineers

NCHRP National Cooperative Highway Research Program

NCTRP National Cooperative Transit Research and Development Program

NHTSA National Highway Traffic Safety Administration

NTSB National Transportation Safety Board

SAE Society of Automotive Engineers

TCRP Transit Cooperative Research Program

TRB Transportation Research Board

U.S.DOT United States Department of Transportation