Recommended Procedures Chapter 13, "Pedestrians," of The Highway Capacity Manual

Recommended Procedures
Chapter 13, "Pedestrians,"

of the Highway Capacity Manual
Capacity Analysis of Pedestrian and Bicycle Facilities

Task Order 8: Pedestrian-Bicycle Research Program
by:
Nagui M. Rouphail, Professor, P.I.
Joseph E. Hummer, Associate Professor, co-P.I.
Joseph S. Milazzo II, Research Assistant
D. Patrick Allen, Research Assistant
North Carolina State University

Department of Civil Engineering `
Raleigh
under the direction of:

The University of North Carolina
Highway Safety Research Center
Chapel Hill
for:
United States Department of Transportation
Federal Highway Administration
Turner-Fairbank Highway Research Center
McLean, Virginia
Final Report
February 1998
Technical Report Documentation Page
1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.
FHWA-RD-98-107
4. Title and Subtitle 5. Report Date
RECOMMENDED PROCEDURES FOR CHAPTER 13, February 1998

"PEDESTRIANS," OF THE HIGHWAY CAPACITY MANUAL 6. Performing
Organization Code
7. Author(s) 8. Performing
9. Performing Organization Name and Address 10. Work Unit No.
(TRAIS)
North Carolina State University
Department of Civil Engineering 11. Contract or Grant No.
Box 7908
Raleigh, NC 27695 DTFH61-92-R-
00138
12. Sponsoring Agency Name and Address 13. Type of Report and
Period Covered
Office of Safety & Traffic Operations Research & Development
Federal Highway Administration Final Report
6300 Georgetown Pike April 1995 -
McLean, Virginia 22101-2296 February 1998
14. Sponsoring Agency
Code
15. Supplementary Notes
Contracting Officer's Technical Representative: Carol Tan Esse (HSR-20)

16. Abstract
The objective of this project is to develop revised operational analysis procedures for
transportation facilities with pedestrian and bicyclist users. This document contains both
new and revised procedures for analyzing various types of exclusive and mixed-use
pedestrian facilities. These procedures are recommended to determine the level of
service for pedestrian facilities on the basis of a summary of available U.S. and
international literature, as described in the Federal Highway Administration (FHWA)
document, "Literature Synthesis for Chapter 13, Pedestrians, of the Highway Capacity
Manual," by these same authors. These procedures are scheduled for incorporation into
a revised U.S. Highway Capacity Manual in 2000.
17. Key Words: 18. Distribution Statement
pedestrian, level of service, platoon, delay, No restrictions. This document is

capacity available to the
public through the National
Technical Information
Service, Springfield, Virginia
22161.
19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of 22. Price
Pages
Unclassified Unclassified
71
Form DOT F 1700.7 (8-72)

Reproduction of form and completed page is authorized
TABLE OF CONTENTS
Section Page
1 INTRODUCTION 1
1.1 A Note on Liability 2
1.2 Summary of Recommendations for Design and/or Analysis of Pedestrian 3
Facilities
2 SUMMARY OF PEDESTRIAN CHARACTERISTICS 4

2.1 Pedestrian Space Requirements 4
2.2 Pedestrian Walking Speeds 5
2.3 Pedestrian Start-up Times 7
2.4 Pedestrian Traffic Flow Relationships 8
3 SERVICE MEASURES OF EFFECTIVENESS 9

3.1 Uninterrupted Pedestrian Facilities 9
3.2 Interrupted Pedestrian Facilities 21
3.3 Pedestrian Networks 33
4 METHODS FOR COMPUTING MEASURES OF EFFECTIVENESS 36

4.1 Uninterrupted Facilities 36
4.2 Interrupted Pedestrian Facilities 36
4.3 Pedestrian Networks 54
5 SUMMARY OF RECOMMENDATIONS FOR DESIGN AND/OR 56

ANALYSIS OF PEDESTRIAN FACILITIES
6 APPENDIX: ADDITIONAL FORMULAS FOR COMPUTING 57
RECOMMENDED MEASURES OF EFFECTIVENESS
Interrupted Pedestrian Facilities 57
7 ACKNOWLEDGMENTS 61
8 REFERENCES 62
LIST OF FIGURES
Figure Page
FIGURE 1 A wide variety of transportation facilities must effectively serve a 2
wide variety of users .
FIGURE 2 Recommended pedestrian body ellipse for standing areas. 4
FIGURE 3 This elderly pedestrian, and others like her, may be helped by the 6
proposed revisions to crosswalk walking speeds
FIGURE 4 The proposed revisions to crosswalk walking speeds may also 6
benefit people who are not elderly, such as this pedestrian pushing a stroller
FIGURE 5 Illustration of proposed walkway Level of Service thresholds 11
FIGURE 6 Pedestrians who know each other often travel in platoons 14
FIGURE 7 Noncompliant pedestrian behavior is common at this Chicago, 22
Illinois, intersection due to low conflicting vehicle volumes
FIGURE 8 Noncompliant behavior is not limited to pedestrians at the same 22
Chicago, Illinois, intersection
FIGURE 9 Field measurements of pedestrian delay at midblock crossings in 40
Great Britain
FIGURE 10 Simulation results of pedestrian delay at fixed-time pelican 41
crossings in Great Britain
FIGURE 11 Simulation results of pedestrian delay at vehicle-actuated pelican 42
crossings in Great Britain
FIGURE 12 Simulation results of pedestrian delay at zebra crossings in Great 49
Britain
FIGURE 13 Effect of crossing width and conflicting vehicle volume on 51
pedestrian
LIST OF TABLES
Table Page
TABLE 1. Recommended pedestrian crosswalk walking speeds 5
TABLE 2. Existing HCM walkway Level of Service (LOS) criteria 9
TABLE 3. Walkway Level of Service (LOS) thresholds by space (m2/ped) and
flow rate
(ped/m/min) 10
TABLE 4. Recommended HCM walkway Level of Service (LOS) criteria 11
TABLE 5. Platoon-adjusted walkway Level of Service (LOS) thresholds 13
TABLE 6. Recommended HCM platoon-adjusted walkway Level of Service
(LOS)
criteria 13
TABLE 7. Level of Service (LOS) thresholds for platoon flow in
transportation
terminalsa 15
TABLE 8. Recommended HCM pedestrian Level of Service (LOS) criteria
for platoon flow in transportationsa 15
TABLE 9. Recommended HCM pedestrian Level of Service (LOS) criteria for 16
stairs
TABLE 10. Recommended capacity thresholds for crossflows 17
TABLE 11. Level of Service (LOS) thresholds for one-way, two-lane, mixed- 18
use paths
TABLE 12. Level of Service (LOS) thresholds for two-way, two-lane, mixed- 19
use paths
for two-lane, mixed-use paths 20
TABLE 14. Selected de facto WALK extension times 25
TABLE 15. Existing HCM signalized intersection Level of Service (LOS) 26
criteria
TABLE 16. Pedestrian and vehicle delay at midblock crossings in Great 26
Britain
TABLE 17. Selected thresholds for maximum pedestrian delay at signalized 27
intersections
for
signalized crossing delay 28
TABLE 19. Existing HCM unsignalized intersection Level of Service (LOS) 29
criteria
for unsignalized crossing delay 32
TABLE 21. Existing HCM queueing area Level of Service (LOS) criteria 33
TABLE 22. Comparison of existing HCM vehicle arterial Level of Service
(LOS)
criteria with pedestrian arterial threshold proposals by both Virkler
and North Carolina State University 34
TABLE 23. Default values of Delay Adjustment Factors (DF) for positive
pedestrian
platooning 55
1 INTRODUCTION
In the United States, the Highway Capacity Manual (HCM), published by the
Transportation Research Board (TRB), a unit of the National Research Council, provides
guidance for the analysis of transportation facilities. Chapter 13 of the 1994 (update to
the 1985) HCM discusses the operational and planning analysis of pedestrian facilities.
The HCM pedestrian chapter begins by positing some relationships between pedestrian
speed, flow, and density. It continues with analysis procedures for walkways, street
corners, and crosswalks. Although offering the traffic engineer the means to analyze the
most common pedestrian facilities, some of the procedures rely on incomplete and
outdated information. This is unfortunate, because many intersections and walkways in
downtown areas, near college campuses, by transit stops, etc., have moderate to heavy
pedestrian flows, thus warranting accurate procedures (Figure 1).
The need for new procedures stems from reasons besides outdated methods, however.
The heightened importance of "livability" in American communities presents the traffic
engineer with the challenge to fully incorporate pedestrians in transportation analysis.
The "Pedestrian Preamble" that opens the Florida Walkable Communities Guide provides
a unique perspective of the role of the pedestrian in the transportation system: "This
community, in providing for trip making, grants pedestrians and motorists of all ages and
abilities: rights, privileges, safety, mobility, and access.... Intersections should not favor
either motorist or pedestrian, but give equal service and support to both...." (Florida
DOT, 1995; emphasis added).
This report summarizes the pedestrian characteristics-related recommendations from the
companion volume, Literature Review for Chapter 13, Pedestrians, of the Highway
Capacity Manual (Rouphail et al., 1998). It also includes a comprehensive set of
recommended service measures of effectiveness, as well as methods for computing
selected service measures. Finally, this report provides a summary of recommendations,
including a listing of the affected subsections in Chapter 13 of the HCM. This summary is
provided at the end of this chapter for ease of reference, and also at the conclusion of the
report in section 5.
1.1 A Note on Liability
The current HCM provides curves for speeds greater than the maximum freeway speed
limits at the time of publishing (TRB, 1994). Consistent with the HCM's demonstrated
intent of reflecting actual conditions rather than legal thresholds, the recommendations
contained in both this Recommended Procedures for Chapter 13, "Pedestrians." and in
the companion Literature Synthesis for Chapter 13, "Pedestrians," (Rouphail et al.,
1998) are to help achieve more realistic analytical procedures for the HCM. However,
nothing in this Recommended Procedures for Chapter 13, "Pedestrians," or in the
companion Literature Synthesis for Chapter 13, "Pedestrians," or in the Highway
Capacity Manual, is to be construed as advocating the violation of traffic laws by either
pedestrians or drivers. In addition, Recommended Procedures for Chapter 13,
"Pedestrians," the companion Literature Synthesis for Chapter 13, "Pedestrians," and the
Highway Capacity Manual should not be used as a defense for the violation of traffic
laws in any of the States.
Recommended Procedures For Chapter 13
FIGURE 1
A wide variety of transportation facilities must effectively serve a wide variety of
users
1.2 Summary of Recommendations for Design and/or Analysis of Pedestrian

Facilities
RecommendationPage(s)FigureTableHCM Ch. 13 HCM variables and adjustments

Subsections affected
affected
Body ellipse for 4 2 - introductory primarily a design recommendation
standing areas narrative only
Body buffer zone 4 - - walkways, walkway LOS E/F threshold
for walking street corners, changes in Table 13-3
crosswalks
Crosswalk 5 - 1 [Ch. 9: new values replace 4.0 ft/s in
walking speeds Methodology,equation (eq.) 9-8
Input
module]
Ch. 13: new values replace 4.5 ft/s in eq.
introduction, 13-14
crosswalks
Grade and stairs 7 - - walkway speeds decrease by 0.1 m/s in eq.
walking speeds narrative, 13-14 with grades
crosswalks
Crossing speeds 7 - - N/A -- no change --
for platoons
Pedestrian start-up 7 - - N/A -- no change --
time
Capacity 8 - - walkways, walkway LOS E/F threshold
thresholds street corners, changes in Table 13-3
crosswalks
Temporal flow 8 - - N/A -- no change --
variation
LOS (Level of 11 5 4 walkways walkway LOS A/B, E/F thresholds
Service) for change, Table 13-3
walkways
LOS for 12,13 - 6 walkways new table replaces equation 13-3
walkways with
platoons
LOS for 15 - 8 walkways new table applies to terminals with
transportation (new platoon flow
terminals measure)
LOS for stairs 16 - 9 walkways new table applies only to stairs
(new
measure)
LOS for 17 - 10 walkways new table serves as secondary
crossflows (new check for walkways
measure)
LOS for mixed- 21 - 13 walkways new table applies only to mixed-use
use paths (new paths
measure)
Noncompliance 26 - - street corners, minor, major red times in equations
time adjustments crosswalks 13-6, 13-7 change;d
effective red time reduced in
computing ped delay
LOS for 29 - 18 street corners new table based on ped delay;
signalized (new space now secondary
crossingsa measure)
Swept-path 30 - - crosswalks caution to use only under
method for aggressive driver behavior
vehicle effects
LOS for 32,33 - 20 street corners new table based on ped delay
unsignalized (new
crossingsb measure)
LOS for 35,36 - 22 networks new table shows proposals for
pedestrian (new section) analysis of ped networks
networksc
Ped delay at 45 - - street corners method for computing ped delay
signalized (new
crossings measure)
Effective 48 - - crosswalks equation 13-13 corrected;
crosswalk time- calculated TSw will decrease
space
Crossing time in 49, 60 - - crosswalks new equations replace eq. 13-14
platoons with large platoons
Ped delay at 54 - - street corners method for computing ped delay
unsignalized (new
crossings measure)
a
Offers a comparison with delay-based Level of Service for drivers computed in HCM
Chapter 9, "Signalized Intersections"
b
Chapter 10, "Unsignalized Intersections"
c
Offers a comparison with Level of Service for drivers computed in HCM Chapter 11,
"Urban and Suburban Arterials"
d
Current HCM is ambiguous regarding the definition of minor and major red times (Rmi,
Rmj); therefore, the effect of the proposed noncompliance
adjustments will depend on the analyst's interpretation of the HCM
2 SUMMARY OF PEDESTRIAN CHARACTERISTICS
This section summarizes the recommendations regarding pedestrian characteristics as

reported in the companion volume, Literature Synthesis for Chapter 13, Pedestrians, of
the Highway Capacity Manual (Rouphail et al., 1998).
2.1 Pedestrian Space Requirements

Body Ellipses and Buffer Zones
Recommendation. This study recommends for design a simplified body ellipse of 50 cm x
60 cm for standing areas, with a total area of 0.3 m2, or roughly 108% of the ellipse
suggested by Fruin (1971). This shape (Figure 2) serves as an approximate metric
equivalent to Fruin's ellipse. This study also recommends a body buffer zone of 0.75 m2
for walking, near the upper end of the buffer zone range provided by Pushkarev and
Zupan (1975a) and just before "unnatural shuffling" commences.

FIGURE 2
Recommended pedestrian body ellipse for standing areas
2.2 Pedestrian Walking Speeds

Age
Recommendation. This study recommends a pedestrian crosswalk walking speed value of
1.2 m/sec (3.9 ft/s) for most conditions, consistent with recommendations described in
the Literature Review for Chapter 13, Pedestrians, of the Highway Capacity Manual
from several sources. In areas with large numbers of older pedestrians, this study
recommends a crosswalk walking speed value of 1.0 m/s, a nearly 30% decrease from
current HCM values.
The question may arise, "What constitutes large numbers of older pedestrians?" A
suggested answer is, "large numbers of older pedestrians exist when the elderly
proportion begins to materially affect the overall speed distribution at the facility."
Through a simple analysis of a simulated dataset, it was found that the 15th percentile
speed for the overall population will drop to 1.15 m/s (i.e., at least 0.05 m/s below the
recommended default value of 1.2 m/s) when the elderly proportion increases to about
20%. Therefore, this study recommends the use of the lower 1.0 m/s value when the
percentage of elderly using the facility in question exceeds 20%. Table 1 summarizes the
recommendations. For demographics information, consider census data at the
neighborhood level. Figures 3 and 4 illustrate some typical users who may benefit from
the proposed changes.
TABLE 1
Recommended pedestrian crosswalk walking speeds
Facility Population Above Age 65 Suggested Walking Speeda for
Time-Limited Walkwaysb
(% of all facility users) (% decrease
0-20 (m/s)(ft/s) from current
< 20 HCMc )
1.2 3.9 14%
1.0 3.3 29%
a
If necessary, adjust minimum crossing time for platoon flow
b
Crosswalks and other facilities where available user time is limited
c
Current HCM uses 1.4 m/s (4.5 ft/s) design crosswalk walking speed

FIGURE 3
This elderly pedestrian, and others like her, may be helped by the proposed
revisions to crosswalk walking speeds
FIGURE 4
The Proposed revisions to crosswalk walking speeds may also benfit people who are
not elderly, such as this pedestrian pushing a stroller
Grades and Stairs

Recommendation. This study recommends that the HCM include a policy of not
correcting for grades less than 10%. Above 10% on upgrades, this study advocates a 0.1
m/s reduction in walking speed as an approximation (roughly the amount found by the
Institute of Transportation Engineers in 1976).
Platoons
Recommendation. For most situations where platoons are prevalent, this study does not
recommend the use of walking speeds lower than 1.2 m/s (1.0 m/s for large elderly
populations).
However, in light of the research by Virkler (1997c) described in the Literature Review
for Chapter 13, Pedestrians, of the Highway Capacity Manual, this study recommends
increasing the minimum signalized intersection crossing time when typical platoons
exceed 15 people. This report details several crossing time computational methods later.
This study also cautions the analyst to consider impairments to full usage of the
crosswalk. These may include: lack of a stop bar, lack of high-visibility crosswalk
markings, crosswalks misaligned with the natural flow of the sidewalk, and corner
obstructions.
2.3 Pedestrian Start-up Times

Recommendation. For simplicity, this study recommends retaining the HCM's value of 3
s, a reasonable mid-range value between the 50th- and 85th-percentile design values (2.5
and 3.75 s, respectively) for older pedestrians suggested by Knoblauch, Pietrucha, and
Nitzburg (1996) and described in the companion Literature Review for Chapter 13,
Pedestrians, of the Highway Capacity Manual.
2.4 Pedestrian Traffic Flow Relationships

Capacity
Recommendation. Given the comfort zone requirements for Americans, it seems that
walkway capacity lies between 4,000 and 5,000 pedestrians/h/m. For simplicity, this
study recommends an assumed capacity of 75 ped/min/m (4,500 ped/h/m). This study
recommends an assumed speed at capacity of 0.75 m/s. In addition, this study
recommends the pedestrian buffer zone space of 0.75 m2/ped for a capacity threshold.
Temporal Variation
Recommendation. This study recommends the use of platoon flow Level of Service
(LOS) considerations (discussed later) in lieu of a pedestrian "peak-hour-factor" (PHF) or
similar measure. Pedestrian queueing is of comparatively short duration relative to the
vehicular queues that a PHF typically describes.
3 SERVICE MEASURES OF EFFECTIVENESS
3.1 Uninterrupted Pedestrian Facilities
Sidewalks and Walkways

The current HCM uses pedestrian space as the primary measure of effectiveness (MOE),
with mean speed and flow rates as secondary measures (Table 2; TRB, 1994). Carrying
units of area per pedestrian in the existing HCM, the measure offers a simple, intuitive
method of service evaluation. The chapter defines capacity as 6 ft2/ped (about 0.56
m2/ped).
TABLE 2
Existing HCM walkway Level of Service (LOS) criteria
Space Flow Rate Average Speed v/c ratio
LOS (m2/ped) (ft2/ped) (ped/min/m) (ped/min/ft) (m/s) (ft/min) -

A >12 >130 <7 <2 <1.32 >260 0.08
B 3.7-12 40-130 23-Jul 7-Feb 1.27-1.32 250-260 0.08-0.28
C 2.2-3.7 24-40 23-33 10-Jul 1.22-1.27 240-250 0.28-0.4
D 1.4-2.2 15-24 33-49 15-Oct 1.14-1.22 225-240 0.4-0.6
E 0.6-1.4 15-Jun 49-82 15-25 0.76-1.14 150-225 0.6-1.0
F <0.6 <6 var. var. <0.76 <150 var.
SOURCE: TRB, 1994.
Alex Sorton of Northwestern University suggests that the current LOS A space
requirement is excessive, and should be reduced from 130 ft2/ped (12 m2/ped) to 60
ft2/ped (5.6 m2/ped). Indeed, the Interim Materials on Highway Capacity (TRB, 1980)
recommended an even lower space threshold (3.7 m2/ped or 40 ft2/ped) than Sorton's
recommendation. This report has stated earlier that capacity probably occurs around 75
peds/min/m, somewhat lower than HCM values.
As a point of comparison, Table 3 compares LOS values in the HCM with those reported
from other researchers. Tanaboriboon and Guyano (1989) developed LOS standards for
Bangkok, Thailand. Although probably not useful for most areas of the United States,
their data in the table highlight the importance of cultural values and physical
characteristics on LOS breakpoints. The authors note that one result of the difference
between Thai and American LOS standards is that pedestrian facilities in Thailand can
accommodate higher flows at a given LOS. Stating that capacity limitations do not
normally dominate pedestrian facility concerns, Brilon stated that Germany's revised
pedestrian LOS standards will have breakpoints based on density (1994). The boundaries
for Polus et al.'s work correspond to the three regimes of pedestrian flow reported by
those researchers.
TABLE 3
Walkway Level of Service (LOS) thresholds by space (m2/ped) and flow rate
(ped/m/min)
United States of America Germany Israel Thailand

HCM Fruin Pushkarev- Brilon Polus et al. Tanaboriboon-Guyano
Zupan
LOS (m2/ped ) (m2/ped) (m2/ped) (m2/ped) (m2/ped) (m2/ped)
49a
A 12 3.2 Dec-49 10 2.38
B 3.7-12 2.3-3.2 12-Apr 3.3-10 1.60-2.38
b
C 2.2-3.7 1.4-2.3 4-Feb 2-3.3 1.67 0.98-1.60
D 1.4-2.2 0.9-1.4 1.5-2 1.4-2 1.33-1.66 0.65-0.98
0.8-1.33
E 0.6-1.4 0.5-0.9 1-1.5 0.6-1.4 0.5-0.8 0.37-0.65
F 0.6 0.5 0.2-1 0.6 unknown 0.37
LOS (ped/min/m) (ped/min/m) (ped/min/m) (ped/min/m) (ped/min/m)

1.6a
A 6.6 23 1.6-7.0 28
B 6.6-23 23-33 20-Jul 28-40
b
C 23-33 33-49 20-33 40 40-61
D 33-49 49-66 33-46 40-50 61-81
50-75
E 49-82 66-82 46-59 75-95 81-101
F var. var. 0-82 unknown 101 or var.
a
Instead of HCM LOS designations "A"-"B"-"C"-"D"-"E"-"F", Pushkarev and Zupan
use"Open"-
"Unimpeded"- "Impeded"-"Constrained"- "Crowded"- "Congested"- "Jammed"
b
Instead of HCM LOS designations "A"-"B"-"C"-"D"-"E"-"F", Polus et al. use A-B-C1-
C2-D
SOURCES: TRB, 1994; Fruin, 1971; Pushkarev and Zupan, 1975b; Brilon, 1994; Polus
et al., 1983; Tanaboriboon and Guyano, 1989.
Recommendation. This study recommends keeping the current HCM walkway LOS B, C,
and D thresholds. This study also recommends changing the capacity thresholds to the
values mentioned earlier. Table 4 summarizes the recommendations. Figure 5
approximates the revised service levels.
TABLE 4
Recommended HCM walkway Level of Service (LOS) criteria
Space Flow Rate Average Speed v/c ratio
2 2
LOS (m /ped) (ft /ped) (ped/min/m) (ped/min/ft) (m/s) (ft/min)
A >5.6 >60 <16 <5 >1.3 >255 0.21
B 3.7-5.6 40-60 16-23 7-May 1.27-1.30 250-255 0.21-0.31
C 2.2-3.7 24-40 23-33 10-Jul 1.22-1.27 240-250 0.31-0.44
D 1.4-2.2 15-24 33-49 15-Oct 1.14-1.22 225-240 0.44-0.65
E 0.75-1.4 15-Aug 49-75 15-23 0.75-1.14 150-225 0.65-1.0
F <0.75 <8 var. var. <0.75 <150 var.
Platoons. The companion volume, Literature Review for Chapter 13, Pedestrians, of the
Highway Capacity Manual noted the effect of platoons on walkway flow. Table 5
summarizes the initial research on platoons. Pushkarev and Zupan (1975b) note that
earlier research found the ability to pass slow-moving pedestrians to be relatively
unrestricted at space modules above 3.3 m2/ped, difficult between 1.7 and 3.3 m2/ped,
and essentially impossible below 1.7 m2/ped. Pushkarev and Zupan also compared
average flow rates with possible flow in platoons. They found no difference between the
flow conditions at any service level, except at that point in "Impeded" flow
(approximately LOS B) when platoons begin (1975b).
The Interim Materials on Highway Capacity (TRB, 1980) contained platoon flow criteria.
This work, relying on the "rule of thumb" mentioned earlier, simply rewrote the
recommended walkway values up one level for platoons. The current HCM, which does
not contain a platoon flow service level table, uses different walkway values for average
flow rate and space at most service levels than those in the Interim Materials. Therefore,
one cannot simply apply the values listed in the Interim Materials to the current HCM.
One can develop platoon flow LOS criteria based on a synthesis of the relationship
between average and platoon flow described in the companion Literature Review, the
existing HCM walkway standards for midrange LOS values (TRB, 1994), and the earlier
work of Pushkarev andZupan (1975b) for extreme values. For LOS A, this report uses
Pushkarev and Zupan's relationship between average and platoon flow (Figure 20 in the
Literature Review) and defines this breakpoint to be just before the discontinuity, at 1.6
ped/min/m (0.5 ped/min/ft), identical to the "Open" flow of Pushkarev and Zupan. For
LOS B through D, this study applies metricized "rule of thumb" to 1994 HCM walkway
values, by subtracting 13 ped/min/m from walkway flow rates. For LOS E, and thus LOS
F, this report uses the highest platoon flow rate found by Pushkarev and Zupan, 59
peds/min/m. The resulting values, shown in Table 6, provide a sound basis for
determining the level of service experienced by people who travel in platoons, such as the
pedestrians shown in Figure 6.
Recommendation. This study recommends incorporating the walkway platoon criteria in
Table 6 into the HCM.
FIGURE 5
Illustration of proposed walkway Level of Service thresholds
SOURCE: TRB, 1994; adapted from FRUIN, 1971.
TABLE 5
Platoon-adjusted walkway Level of Service (LOS) thresholds
Space Flow Rate
Pushkarev-Zupan Interim Materials Pushkarev-Zupan Interim Materials
2 2 2 2
LOS (m /ped) (ft /ped) (m /ped) (ft /ped) (ped/min/m) (ped/min/ft) (ped/min/m) (ped/min/ft)
Aa >49b >530 12c 130 >1.6b >0.5 6c 2
B 6-Apr 40-60 12-Apr 40-130 15-20 4.5-6 20-Jun 6-Feb
C 4-Feb 24-40 4-Feb 24-40 20-33 10-Jun 20-33 10-Jun
D 1.5-2 16-24 1.5-2 16-24 33-46 14-Oct 33-46 14-Oct
E 1-1.5 16-Nov 1-1.5 16-Nov 46-59 14-18 46-59 14-18
F 1 11 0.6-1 11-Jun 59 18 59-82 18-25
a
Instead of HCM LOS designations "A"-"B"-"C"-"D"-"E"-"F", Pushkarev and Zupan use
"Open"- "Impeded"-"Constrained"- "Crowded"- "Congested"- "Jammed"
b
Values given by Pushkarev and Zupan for flow rates and space are within platoons
c
Values given in the Interim Materials for flow rates and space are under average flow
conditions
The LOS shown at each flow rate or pedestrian space level represents the walkway LOS
(based on Interim Materials service levels) under these average flow rates when platoons
arise
SOURCE: Pushkarev and Zupan, 1975b; TRB, 1980.
TABLE 6
Recommended HCM platoon-adjusted walkway Level of Service (LOS) criteria
Space Flow Ratea

LOS (m2/ped) (ft2/ped) (ped/min/m) (ped/min/ft)
A 49 530 1.6 0.5
B Aug-49 90-530 1.6-10 0.5-3
C 8-Apr 40-90 20-Oct 6-Mar
D 4-Feb 23-40 20-36 11-Jun
E 2-Jan 23-Nov 36-59 18-Nov
F 1 11 59 18
*Flow rate in the table represent average flow rates over a 5-6 min period. The LOS
shown is the
walkway LOS under these average flow rates when platoons arise
Transportation terminals provide a special case of platoon flow. Davis and Braaksma
(1987) analyzed the pedestrian flow within an airport corridor by a "floating pedestrian"
method, in which the surveyor measures traffic parameters from within the pedestrian
stream. Table 7 shows the LOS standards developed by the authors for platoon flow in
transportation terminals. By implication, the use of the term "transportation terminal"
refers to both an airport and to those other locations with tendencies for the platooning
behavior common in airport walkways. Note that, although maximum speed and space
occur at the highest LOS (A+ in the table), the maximum flow occurs at the boundary
between LOS D and E. Also of note, the extremely high flows in these facilities warrant
much less restrictive service criteria. To facilitate incorporation into the HCM, one can
eliminate or consolidate one of their seven service levels (Table 8). This report
consolidates Davis and Braaksma's LOS A and B into LOS B and redesignates LOS A+
as LOS A. In effect, this expands the transportation terminal LOS B to a range roughly
coincident with platoon-adjusted walkway criteria LOS E. In addition, LOS E reflects the
capacity thresholds suggested earlier.

FIGURE 6
Pedestrians who know each other travel in platoons
Recommendation. This study recommends incorporating the transportation terminal
criteria adapted from Davis and Braaksma in Table 8 into the HCM.
TABLE 7
Level of Service (LOS) thresholds for platoon flow in transportation terminalsa
Space Flow Rate Speed
LOS (m2/ped) (ped/min/m) (m/s)
A+ >2.3 <37 >1.4
A 1.7-2.3 37-46 1.3-1.4
B 1.3-1.7 46-57 1.2-1.3
C 1.0-1.3 57-68 1.1-1.2
D 0.8-1.0 68-75 1.0-1.1
E 0.7-0.8 57-75 0.7-1.0
F <0.7 <57 <0.7
a
Airports or other facilities where platoon flow is prevalent along pedestrian walkways
SOURCE: Davis and Braaksma, 1987.
TABLE 8
Recommended HCM pedestrian Level of Service (LOS) criteria
for platoon flow in transportationsa
Space Flow Rate Speed

LOS (m2/ped) (ped/min/m) (m/s)
A+ >2.3 <37 >1.4
A 1.7-2.3 37-46 1.3-1.4
B 1.3-1.7 46-57 1.2-1.3
C 1.0-1.3 57-68 1.1-1.2
D 0.8-1.0 68-75 1.0-1.1
E 0.7-0.8 57-75 0.7-1.0
F <0.7 <57 <0.7
aAirports or other facilities where platoon flow is prevalent along pedestrian walkways
Stairs. To allow for the determination of pedestrian arterial ("network" in this review)
LOS, Virkler utilized a 20-year-old proposed ITE stairways standard (ITE, 1976), which
provided space and flow values at various stairway LOS. Virkler states that he modified
this standard somewhat "to ensure that the basic equation of traffic flow is satisfied,"
although this review of his research could discern no difference between his values and
the space and flow values given in the ITE stairways standard.
Table 9 shows the recommended HCM pedestrian LOS criteria. The values reflect ITE's
flow values, Fruin's (1971) original breakpoints for stairway level of service, and
Virkler's values for speed and volume-capacity ratio. Note that the LOS E values of 49
and 56 ped/min/m for Virkler and Fruin, respectively, are noticeably less than the 62-73
ped/min/m capacity ranges found in the Hong Kong and London transit systems listed
earlier by Lam et al. (1995).
Recommendation. In so far as Virkler's paper corrects earlier research by ensuring
congruence with pedestrian traffic flow theory, his work remains the best available for
American users. This study recommends this material (Table 9) for the HCM, pending
further research on capacity limits.
TABLE 9
Recommended HCM pedestrian Level of Service (LOS) criteria for stairs
Space Flow Rate Avg. Horiz. Speed v/c ratio

LOS (m2/ped) (ped/min/m) (m/min) (m/s)
A 1.9 16 32 0.53 0.33
B 1.6-1.9 16-20 32 0.53 0.33-0.41
C 1.1-1.6 20-26 29-32 0.48 0.41-0.53
D 0.7-1.1 26-36 25-29 0.42 0.53-0.73
E 0.5-0.7 36-49 24-25 0.4 0.73-1.00
F < 0.5 var. < 24 < 0.40 var.
Crossflows. A crossflow is a pedestrian flow that is roughly perpendicular to and crosses

another pedestrian stream. In general, one refers to the smaller of the two flows as the
crossflow. Khisty (1982) notes that pedestrian crossflows occur in hallways and corridors
and are "ubiquitous." Table 10 notes his suggestions for acceptable criteria regarding
corridor crossflows. These values correspond roughly with the bottom half of HCM
walkway LOS E; by terming them minimums and maximums, he implies that his values
establish LOS boundaries for crossflows.
Recommendation. This study recommends the incorporation into the HCM of Khisty's
crossflow standards listed in Table 10 below as an interim measure pending further
research.
TABLE 10
Recommended capacity thresholds for crossflows
Speed Flowb Density Space

LOS (m/s) (ped/min/m) (ped/m2) (m2/ped)
Ea 1 75 0.8 1.25
a
Khisty terms these threshold values "minimums" and "maximums"; by implication, this
is
LOS E.
b
Total of the major and minor flow
SOURCE: Khristy, 1982.

Off-street paths
Exclusive pedestrian trails. Virkler and Balasubramanian (1997), in their discussion of
flow characteristics on shared user trails, imply that the current HCM's LOS walkway
guidelines apply for exclusive pedestrian trails.
Shared pedestrian-bicycle paths. Virkler and Balasubramanian (1997) describe a 1995
study by Hein Botma of shared pedestrian-bicycle facilities in The Netherlands. This
study develops LOS guidelines for both pedestrians and bicyclists on the basis of the
frequency of passing (same direction) and meeting (opposite direction) other users on the
trail. Botma characterizes these two occurrences as "events," with an overtaking equal to
one event and a meeting equivalent to one-half of an event. Under this framework, LOS F
refers to "very bad quality of traffic operation," not congestion (Botma, 1995). More
specifically, it refers to a situation where an average user experiences "hindrance" more
than 1.0 times in a 1-km trail segment. Virkler and Balasubramanian note that, for one-
way paths, pedestrians seldom overtake other pedestrians, and thus the LOS afforded a
pedestrian on a shared path depends on the frequency with which an average pedestrian
would be overtaken by bicyclists. In Botma's discussion of his own work (1995), he
poses the question of whether it is justified to neglect hindrance due to pedestrian
interactions. As the authors of this report have observed moderate pedestrian-pedestrian
hindrances on various mixed-use trails, it is likely that Botma's assumption of negligible
pedestrian interactions is not entirely correct.
Botma's expression describing the total number of overtakings of pedestrians by
bicyclists, Nf/s, is:
Nf/s = X T Qf Qs (1/Us - 1/Uf)
where:
X = length of site, m;
T = time period considered, s;
Qf = flow of faster group in subject direction, bicyclists/s;
Qs = flow of slower group in subject direction, pedestrians/s;
Uf = mean speed of faster group, m/s (for bicyclists); and
Us = mean speed of slower group, m/s (for pedestrians).
Using an average pedestrian speed of 1.25 m/s and an average bicyclist speed of 5 m/s,
Botma developed a LOS table for pedestrians on one-way, two-lane shared-use paths.
Table 11, which converts "frequency" of events into period between events to eliminate
fractions, provides Botma's LOS thresholds.
TABLE 11
Level of Service (LOS) thresholds for one-way, two-lane, mixed-use paths
LOS Period Between Events Service Volume
(s/event) (bicycles/h)
A > 150 < 33
B 75-150 33 - 64
C 35 - 75 65 - 136
D 20 - 35 137 - 240
E 15 - 20 240 - 320
F < 15 > 320
SOURCE: Adapted from Botma, 1995.
For two-way trails, Botma states that pedestrians still seldom overtake other pedestrians,
and thus the LOS afforded a pedestrian on a shared path depends on the frequency with
which an average pedestrian experiences meetings of and overtakings by bicyclists.
Using the speed assumptions listed above for one-way paths, Botma established a table
for pedestrians traveling on two-lane, two-way, shared-use paths. Table 12 , again
substituting period for frequency, shows Botma's service levels.
As an aside, if one applied either of the tables to an exclusive pedestrian trail, one would
always have a service level of A, regardless of pedestrian volume, since the tables depend
entirely on bicycle volume. Therefore, Virkler and Balasubramanian's implication to use
existing walkway standards certainly seems more reasonable than the use of Botma's
method for an exclusive pedestrian facility.
TABLE 12
Level of Service (LOS) thresholds for two-way, two-lane, mixed-use paths
LOS Period Between Events Service Volume
(s/event) (bicycles/h)
A > 95 < 29
B 60 - 95 29 - 44
C 35 - 60 45 - 75
D 25 - 35 76 - 105
E 20 - 25 106 - 131
F < 20 > 131
SOURCE: Adapted from Botma, 1995.
Virkler and Balasubramanian (1997) studied flow characteristics on two-way, shared-use

trails in both Columbia, Missouri, and Brisbane, Australia. They found bicycling speeds
of 5.95 m/s and 5.76 m/s in Missouri and Australia, respectively, both of which were
somewhat higher than the 5 m/s speed used by Botma. However, Botma uses 5 m/s for
simplicity; field studies of trails in The Netherlands show slightly higher average speeds
of 5.28 m/s (Botma, 1995). Also, they found that the standard deviations of bicycling
speeds, 2.1 m/s for Missouri and 1.33 m/s for Australia, were much higher than the 0.83
m/s average speed reported by Botma. Finally, they observed average "hiking"
(presumably walking) speeds of 1.59 and 1.56 m/s in Missouri and Australia,
respectively. Upon comparison between predicted (by Botma's tables) and observed
values, Virkler and Balasubramanian found that their results generally supported the
framework espoused by Botma for bicyclists overtaking pedestrians.
Although not mentioned by Virkler and Balasubramanian (1997), if one rounds the
Missouri average speed measurements to the nearest 0.5 m/s (i.e., rounding bicycling
speeds from 5.95 m/s to 6 m/s and walking speeds from 1.59 m/s to 1.5 m/s), then the
resulting table of values for both one- and two-way trails will be identical to that by
Botma. Therefore, in so far as Botma's assumptions are correct, one can directly apply
Botma's pedestrian LOS tables listed above to at least one American mixed-use trail.
Recommendation. In light of the validation of Botma's method on an American mixed-
use path by Virkler and Balasubramanian (1997), this study recommends the
incorporation of the Botma mixed-use path criteria in the HCM. Table 13 summarizes the
recommended LOS thresholds for these paths, identical to Botma's values.
TABLE 13
for two-lane, mixed-use paths
One-Way Paths Two-Way Paths
Perioda Service Volume Perioda Service Volume

LOS (s/event) (bicycles/h) (s/event) (bicycles/h)
A > 150 < 33 > 95 < 29
B 75-150 33 - 64 60 - 95 29 - 44
C 35 - 75 65 - 136 35 - 60 45 - 75
D 20 - 35 137 - 240 25 - 35 76 - 105
E 15 - 20 240 - 320 20 - 25 106 - 131
F < 15 > 320 < 20 > 131
a
Period between events; where an event is either a bicycle meeting or passing a
pedestrian.
3.2 Interrupted Pedestrian Facilities

Signalized Crossings
Overview of Noncompliance. The pedestrian literature contains several articles dealing
with pedestrian disobedience of traffic signals. In addition, anecdotal evidence suggests
that assuming legal behavior will not sufficiently resemble reality for analysis purposes.
Therefore, before considering a delay-based service measure of effectiveness in detail,
one should examine the effects of pedestrian noncompliance. Figures 7 and 8 are
illustrative of the problem.
Middleton (1981), bemoaning the levels of pedestrian accidents in Australia and the
United States, notes the presence of what he terms an "over-supply of pedestrian facilities
at signalized intersections." He notes that safety-motivated pedestrian control signals at
signalized intersections may actually reduce safety by encouraging noncompliance to
avoid the "largely unnecessary delay imposed" on pedestrians. Indeed, the author
observed disobedience rates as high as 70 percent in Queensland, Australia. Stating that
the "very existence of this widespread lawbreaking in the community should be sufficient
evidence that the system needs attention," he reiterates an earlier suggestion by F.R.
Fulsher to change the legal meaning of the DON'T WALK signal to "Yield to Vehicles."
In so far as the resulting change in pedestrian signals from regulation to guidance may
discourage avoidance of pedestrian signals, he hypothesizes that safety improvements
may result.
The Literature Review for Chapter 13, Pedestrians, of the Highway Capacity Manual
included a study by Rouphail (1984) noting the preference of pedestrians for midblock
crossings. However, when a pedestrian crossing is displaced from the intersection, the
increase in travel path for many users walking along the cross-street may tend to breed
signal noncompliance in some situations, as people tend to use the intersection
crosswalks regardless of signal indication (Pretty et al., 1994).
Hunt and Griffiths (1991) note that pedestrians experience very little delay at zebra
crossings, since they always have the right-of-way. However, they note that pedestrians
who are unable or unwilling to accept gaps in traffic during the DON'T WALK period at
the signalized pelican (pedestrian light controlled) midblock crossings in Britain incur
substantial delay. They suggest that pedestrians crossing illegally at signalized
intersections could be less safe than those crossing at random points along a roadway
since drivers approaching a green signal will not expect to have to yield to a pedestrian.
Griffiths et al. (1984a) observed during their field studies that significant numbers of
pedestrians are prepared to begin crossing during either flashing or steady DON'T
WALK pedestrian indications. They noted that noncompliant behavior occurred almost
exclusively when two-way conflicting vehicle flows were below 1,500/h.
Gordon and Robertson (1988) noted that driver noncompliance with traffic signals is a
serious problem as well, particularly at low-volume intersections. They recommend a
combination of higher enforcement levels, stiffer violation penalties, education of the
public, and the removal of unnecessary informational or regulatory control devices
adjacent to intersection approaches.
Knoblauch, Pietrucha, and Nitzburg (1996) noted that, of the pedestrians they observed
during their field study of intersections in eastern cities, those who crossed against the
signal (i.e., noncompliant pedestrians) tended to walk faster than those who crossed
legally.
A study of Hong Kong pedestrians noted that pedestrians walk faster during the red phase
at signals, confirming the ubiquity of noncompliant pedestrians (Lam et al., 1995). The
authors report an average noncompliant pedestrian crosswalk speed of 1.5 m/s in Hong
Kong crosswalks, much higher than the 1.27 m/s level observed at those facilities during
the WALK indication.
Virkler (1997a) noted that, based on his observations of intersections in Brisbane,
Australia, pedestrians typically treat about 69 percent of the flashing DON'T WALK
signal as an effective WALK. He discerns two groups of noncompliant pedestrians:
"jumpers," who start crossing before the WALK indication begins, and "runners," who
begin crossing after the flashing DON'T WALK signal commences. Between the two
groups, he observed that the runners saved over 7 times as much delay per person as
jumpers, so he focused on the behavior of the former group.
North Carolina State University (NCSU) also noticed similar noncompliant behavior at
several sites during its field study of American intersections. NCSU calculated that
pedestrians typically treated the first 5 s of flashing DON'T WALK as a de facto WALK
signal indication. Indeed, the NCSU data-collection team observed some crossings during
both flashing DON'T WALK (which typically coincides with the latter part of the
vehicular green) and the vehicular clearance interval. Milazzo II (1996) adjusted his
volume-occupancy data collection framework to allow for pedestrian occupancy of the
crosswalk at any time during the pedestrian clearance interval.

FIGURE 7
Noncompliant pedestrian behavior is common at this Chicago, Illinois intersection
due to low conflicting vehicle volumes
Viney and Pretty (1982) examined pedestrian and vehicle interactions at Brisbane,
Australia, intersections. They observed an average WALK "extension time" (i.e., de facto
WALK) of 1.95 s with a standard deviation of 2.7 s. They used 2 s as an allowance for
disobedient pedestrians.
It is important to note that changes in signal timing can affect noncompliance. For
example, the slight increase in green time and cycle length that may occur under the
assumption of reduced walking speeds will increase pedestrian delay and probably
increase pedestrian noncompliance. Of course, the presence of excessive cycle lengths
and/or unnecessary phases also causes pedestrian delay and noncompliance. Some
jurisdictions use "early release" signal timing, where pedestrians receive the WALK
before the concurrent vehicles receive the green, in an effort to reduce pedestrian delay.
Regardless of the phasing scheme chosen, most facility users are local pedestrians who
will learn the signal timing and try to reduce their own delay.

FIGURE 8
Noncompliant behavior is not limited to pedestrians at the same Chicago, Illinois
intersection
In summary, any delay measure to pedestrians should include some mechanism for
considering noncompliance. Table 14 summarizes the findings of the last three research
groups mentioned; coincidentally, two of the three groups examined downtown Brisbane,
Australia. All of this empirical research seems to indicate that pedestrians, recognizing
the margin of safety built into the pedestrian clearance interval, treat the initial part as an
effective walk time.
TABLE 14
Selected de facto WALK extension times
Location De Facto WALK Interval,a s

Viney and Pretty Brisbane, Australia WALK + 2
NC State United Statesb WALK + 5
Virkler Brisbane, Australia WALK + 8c or
WALK + 69% of flashing DON'T WALK
a
Observed or calculated total effective WALK interval as used by pedestrians
b
Washington, D.C.; Portland, Oregon; Atlanta; and Chicago
c
Virkler only reports the percentage of flashing DON'T WALK (clearance) time;
8 s is approximately 69% of the 11.1-s mean clearance time for Virkler's study
SOURCES: Viney and Pretty, 1982; Virkler, 1997a.
Recommendation. Based on the middle range of values from the above research on
noncompliance, this study suggests the following adjustments to pedestrian signalized
crossing timing for simplicity:
WALKe = WALK + 5
Flashing DON'T WALKe = Flashing DON'T WALK - 5
where:
WALK = nominal WALK time, s;
WALKe = effective WALK time, s;
Flashing DON'T WALK = nominal flashing DON'T WALK time, s; and
Flashing DON'T WALKe = effective flashing DON'T WALK time, s.
The analyst should be aware, however, that intersections with high conflicting traffic
and/or large street widths have excellent compliance, primarily because pedestrians have
no choice but to wait.
Delay. Currently, no LOS standard based on pedestrian delay at signalized intersections

exists in the HCM. However, the HCM does incorporate vehicular delay at these facilities
into its LOS criteria for vehicles at signalized intersections (Table 15). The Australian
signalized intersection software package SIDRA considers pedestrian delay (Akçelik,
1989) as a performance measure.
The following paragraphs, which give the results of several delay studies, provide a
useful frame of reference for establishing a suitable pedestrian delay criteria at signalized
crossings.
TABLE 15
Existing HCM signalized intersection Level of Service (LOS) criteria
Stopped Delay per Vehicle

LOS (s)
A <5
B 15-May
C 15-25
D 25-40
E 40-60
F 60
SOURCE: TRB, 1994.

Noland (1996) states that any street crossing, regardless of the control device used, will
result in some delay to pedestrians due to caution before entering the crosswalk. He also
argues that, since average delay to pedestrians is frequently ignored at signals, total "costs
to society" may rise due to unfavorable timing patterns. He notes that, if pedestrian green
phases remain constant while cycle lengths increase, average delay to pedestrians can
increase quite rapidly.
Griffiths et al. (1984a) examined pedestrian delay at both signalized and unsignalized
crossings in Great Britain. Table 16 shows the results.
TABLE 16
Pedestrian and vehicle delay at midblock crossings in Great Britain
Unsignalized Signalized
Zebra Fixed-time Pelican Vehicle-actuated
Pelican
(s) (s) (s)
Pedestrian Delay 1.4 10.1 9.8
Vehicle Delay 5.2 4.2 3.9
SOURCE: Griffiths et al., 1984a.
The authors note that Great Britain began installing unsignalized pedestrian crossings in
1935, with signalized pelican installations commencing in 1969. The latter device was
introduced to provide the "flexibility of a Zebra" with the "positive command to drivers
to stop." Along these lines, Dunn and Pretty (1984) state that, provided that pelicans are a
legal device in the jurisdiction, one should always install a pelican crossing over a
standard pedestrian signalized crossing, because they provide reduced vehicular delay
with no detriment to pedestrian delay.
At a field survey of fixed-time signalized crossings in Great Britain, Griffiths et al.
(1984a) found significant increases in pedestrian delay for increases in vehicular delay
over a wide (400-1,400 veh/h) range. They did not observe any additional effect on
pedestrian delay at signalized crossings with vehicle actuation over these volume levels.
MacLean and Howie (1980) examined the performance of pedestrian crossings in
Victoria, Australia. They found that mean pedestrian delay was 17 s at signalized
crossings.
Table 17, based on anecdotal evidence and empirical observation, provides some
maximum delay thresholds recommended by various researchers for signalized
intersections. Dixon (1996) terms the choice of 40 s for Gainesville, Florida, a
compromise value. Kaiser (1994) notes the increase in pedestrian impatience and risk-
taking behavior beyond 30 s of delay; Dunn and Pretty (1984) also mention that
pedestrians become increasingly impatient when delayed beyond 30 s. Hunt and Griffiths
(1991), noting that risk-taking behavior increases with pedestrian delays of 30 s or more,
state that the vehicle precedence time should not exceed 40 s at a pelican crossing in
Great Britain. Of course, with sufficiently high conflicting vehicle volume, pedestrians
can face delays above 60 s (Dunn and Pretty, 1984). Under these conditions, pedestrian
compliance increases, because sufficient gaps do not exist in the vehicle stream for
pedestrians to utilize.
TABLE 17
Selected thresholds for maximum pedestrian delay at signalized intersections
Location Maximum Recommended Pedestrian Delaya
(s)
Brilon Germany 60
Dixon Gainesville, 40
Florida
Dunn and Australia 30
Pretty
Hunt and Great 30
Griffiths Britain
Kaiser United 30
States
a
Values typically based on observation of pedestrian impatience and noncompliance
SOURCES: Brilon, 1994; Dixon 1996; Dunn and Pretty; 1984; Hunt and Griffiths, 1991;
Kaiser 1994.
Recommendation. This study recommends the incorporation of pedestrian delay as a

measure of effectiveness for signalized intersections. This study recommends the
establishment of the delay thresholds shown in Table 18, based on both the anecdotal
evidence of pedestrian tolerance of delay tabulated above and congruence with similar
values for vehicles in Chapter 9, Signalized Intersections, of the current HCM. As
mentioned above, the current HCM contains no procedures for midblock crossings. The
signalized type resembles an intersection crossing in that the signal incorporates a time
element with a limited, predictable duration for pedestrians to legally complete their
crossing. Therefore, this study recommends the above intersection crossing criteria for
signalized midblock crossings.
TABLE 18
Recommended HCM pedestrian Level of Service (LOS) criteria for signalized
crossing delay
Average Delay Per Likelihood of Pedestrian Noncompliance
Pedestrian
LOS (s)
A < 10 Low
B 20-Oct
C 20-30 Moderate
D 30-40
E 40-60 High
F 60 Very High
Space. As was the case with walkways, the current HCM uses pedestrian walkway space
criteria as the primary MOE for street corners. However, in this case, the methodology
centers on a validated "time-space" framework developed by Fruin and Benz (1984). It
provides average space values of 5 ft2/ped in a queue and average time values of 3 to 5 s
for moving through the corner.
The existing HCM also offers a crude method of describing the effect of turning vehicles
on pedestrians at intersections, by assuming a swept-path for a vehicle and decrementing
the crosswalk time-space available to pedestrians. Indeed, despite the legal precedence of
pedestrians over vehicles in the crosswalk, Virkler (1997c) found that vehicles
occasionally occupy a portion of the crosswalk during the pedestrian phase.
Unsignalized Crossings.
Delay. The current HCM does not have a method for analyzing unsignalized crossing
facilities. However, the HCM unsignalized intersection chapter does provide a
mechanism for computing vehicular delay at these locations. Table 19 provides delay
thresholds for vehicles at (two- or all-way) stop-controlled intersections, the most
common unsignalized intersection types in the United States.
TABLE 19
Existing HCM unsignalized intersection Level of Service (LOS) criteria
Average Total Delay

LOS (s/vehicle)
A <5
B 10-May
C 20-Oct
D 20-30
E 30-45
F 45
SOURCE: TRB, 1994.
As in the signalized intersection case, it is useful to examine existing research on

pedestrian delay at unsignalized crossings to gain a feel for actual delay levels at these
facilities. Dunn and Pretty (1984) examined pedestrian and vehicle delay at Australian
and New Zealand midblock crossings. They neglected pedestrian delay at unsignalized
(zebra) crossings, however, effectively terming it negligible. They therefore focused
solely on vehicle delay for the unsignalized case.
MacLean and Howie (1980) examined the performance of pedestrian crossings in the
Australian state of Victoria. They found that mean pedestrian delay was 1.7 s at
unsignalized midblock crossings in Victoria, dramatically (and somewhat surprisingly)
less than the 17-s mean delay at signalized midblock crossings. Mean pedestrian delay at
zebra crossings was 2.3 s in metro Melbourne but negligible in rural areas.
At low to moderate vehicle volumes, Griffiths et al. (1984a) found little pedestrian mean
delay at unsignalized crossings. They also noted that average pedestrian delay decreases
as pedestrian flow increases because more pedestrians can take advantage of "an
established pedestrian precedence."
Song, Dunn, and Black (1993) examined the interaction of pedestrians and vehicles for
pedestrians crossing at least 10 m away from a designated crossing. The authors collected
pedestrian gap acceptance characteristics at several streets in Sydney, Australia. They
divide pedestrian crossing tactics into four categories: "double-gap," "risk-taking," "two-
stage," and "walk'n-look," each of which serves to minimize crossing time while still
providing a degree of safety. Their approach assumes that each crossing tactic, rather
than each person, has critical gaps for the near lane, far lane, and combined traffic
streams associated with it. A corollary is that different demographic groups will typically
use a particular crossing tactic; for example, disabled and elderly pedestrians, and
mothers with children, will often use the cautious "double-gap" tactic.
The "double-gap" tactic involves identifying a gap of size a in the near stream and 2a in
the far stream, in order to ensure successful crossing of the entire street in one continuous
motion. The "risk-taking" tactic involves selecting individual gaps of a in each of the
lane-by-lane traffic streams. A "two-stage" crossing involves the use of a median as a
refuge. The "walk'n-look" tactic involves walking parallel to the street in the direction of
desired travel until a suitable gap arrives, then crossing using one of the previous three
tactics. Users of this tactic can essentially eliminate crossing delay under low to moderate
conflicting vehicle volumes; in addition, the authors note that by minimizing interaction
with vehicles, little accident risk exists for users of this tactic (Song et al., 1993).
Palamarthy et al. (1994) describe available crossing tactics to pedestrians at signalized
intersections analogous to those described by Song et al. (1993), except that a lane-by-
lane crossing substitutes for the "walk'n-look" at these locations. Palamarthy et al. found
that pedestrians are more likely to look for an overall gap rather than separate gaps in
individual traffic streams. The authors found mean critical gaps of 3.33 s for the near
traffic stream under all crossing tactics, 7.14 s for the far stream under a double-gap
crossing, 3.58 s for the far stream under a risk-taking crossing, and 3.81 s for the far
stream under a two-stage crossing.
TABLE 20
for unsignalized crossing delay
Average Delay Per Pedestriana Likelihood of Risk-Taking Behavior

LOS (s) by Pedestriansb
Average Delay Per Pedestriana Likelihood of Risk-Taking Behavior

LOS (s) by Pedestriansb
A <5 Low
B 10-May
C 20-Oct Moderate
D 20-30
E 30-45 High
F >45 Very High
a
Delay includes waiting on one side to begin crossing and/or waiting in the median to
complete
the crossing
b
Likelihood of acceptance of short gaps
Finally, the HCM contains no provision for a space-based measure of effectiveness for
unsignalized crossings. In this case, the periodic element found at a signalized
intersection is not as pronounced, and the delay to pedestrians predominates.
Other Waiting Areas
Space. The current HCM uses pedestrian space as the primary MOE. Based on average
pedestrian space, personal comfort, and degree of internal mobility, capacity here is 2
ft2/pedestrian (0.19 m2/ ped). The values of space for queueing or waiting areas at each
level of service shown in Table 21 vary from 10 to 50 percent of the space required for
circulation on walkways.
TABLE 21
Existing HCM queueing area Level of Service (LOS) criteria
Space Interperson Spacing

LOS (m2/ped) (ft2/ped) (m) (ft)
A >1.21 >13 1.2 >4
B 0.93-1.21 13-Oct 0.9-1.2 3.5-4
C 0.65-0.93 10-Jul 0.7-0.9 3-3.5
D 0.27-0.65 7-Mar 0.3-0.7 3-Feb
E 0.19-0.27 3-Feb < 0.3 <2
F < 0.19 <2 negligible negligible
SOURCE: TRB, 1994; from Fruin, 1971.
3.3 Pedestrian Networks

The German Highway Capacity Manual recommends a maximum pedestrian delay of 90
s total for a series of signals (Brilon, 1994).
Virkler (1996) notes that the HCM's arterial analysis chapter (11) uses overall average
travel speed as the measure of effectiveness in determining LOS. He recommends use of
average travel speed for pedestrian arterials (routes) as well. Table 22 compares his
recommended travel speed values with appropriate speed values from the HCM's
vehicular arterial analysis chapter. Virkler's values represent an adaptation of pedestrian
walkway and signalized intersection vehicular delay LOS standards. For a given LOS, the
pedestrian arterial values represent an average travel speed, assuming the average
walkway speed at that LOS with a delay over a 100-m length equal to that signalized
intersection LOS. Examination of the above criteria reveals that one would have to
maintain normal walking speeds throughout the entire arterial (i.e., essentially no
stopping at signals or other nodes) in order to achieve the upper levels of service.
TABLE 22
Comparison of existing HCM vehicle arterial Level of Service (LOS) criteria with
pedestrian arterial threshold proposals by both Virkler and North Carolina State
University
Pedestrian arterial
threshold proposals
Cl. I Vehicle
Cl. III Vehicle Arterials Virkler NCSU
Arterials
LOS (m/s) (mi/h) (%FFSa) (m/s) (mi/h) (%FFSb) (m/min) (m/s) (mi/h) (%FFSc) (ratiod)
A 16 35 88 11 25 93 80 1.33 3 95 90
B 13 28 70 8.5 19 70 70 1.17 2.6 84 70-90
C 9.8 22 55 5.8 13 48 60 1 2.2 71 50-70
D 7.6 17 42 4 9 33 50 0.83 1.9 59 40-50
E 5.8 13 33 3.1 7 26 35 0.58 1.3 41 30-40
F < 5.8 < 13 < 33 < 3.1 < 7 < 26 < 35 < < 1.3 < 41 30
0.58
a
Percent of 18 m/s (40 mi/h) free-flow speed for Class I vehicle arterials that favor
mobility
b
Percent of 12 m/s (27 mi/h) free-flow speed for Class III vehicle arterials that favor
access
c
Percent of 1.4 m/s (3.1 mi/h or 84 m/min) free-flow speed for pedestrian walkways
d
Ratio of calculated minimum travel time to actual travel time, multiplied by 100 for
easier
comparison with "percent of free-flow speed" used for vehicle arterial thresholds
SOURCES: TRB, 1994; Virkler, 1996.
One alternative method could be to determine the minimum travel time at a given LOS,
compare this with the actual travel time (i.e., incorporating any delay), and define a
pedestrian network LOS based on this ratio. The following equation shows this:
Minimum Travel Time . Calculated Minimum Travel Time
Time Ratio = =
Delay + Minimum Travel Time Actual Observed Travel Time
where:
0 < Time Ratio 1
With this expression, a trip with no delay will have a time ratio of 1.0, while a trip taking
four times as long as the minimum will have a time ratio of 0.25. In addition, this
formulation allows for calibration of minimum times at a local site. However, Virkler's
proposal offers the convenience of a fixed-speed criterion for each service level.
4 METHODS FOR COMPUTING MEASURES OF EFFECTIVENESS
Note: All values for walking speeds are those used by the original researcher, rather than
those recommended in this report, unless otherwise noted.
4.1 Uninterrupted Facilities
Sidewalks and Walkways
The existing HCM contains detailed analysis procedures for these facilities (TRB, 1994).
Although this report recommends new LOS thresholds, the basic procedures for the
facilities will not change.
Off-street paths
Exclusive pedestrian trails. As stated earlier, the existing HCM procedures for walkway
analysis apply here. Although this study recommends new service level thresholds, the
basic procedure for these facilities will not change.
Shared pedestrian-bicycle paths. For these facilities, Botma's procedure, described
earlier, is the only viable alternative in the literature. The procedure consists of measuring
bicycle volume and then assigning a pedestrian LOS based on this volume.
4.2 Interrupted Pedestrian Facilities
The existing HCM contains detailed analysis procedures for these facilities (TRB, 1994).
Chapter 13 notes that one can analyze a crosswalk as a time-space zone, similar to a
street corner. According to the HCM, the demand for space equals the product of
pedestrian crossing flow and average crossing time. The chapter notes that a surge
condition exists when the two opposing platoons meet. One determines the primary
measure of effectiveness, space per pedestrian, using this time-space methodology. No
delay measures exist, as stated earlier.
Delay: Pretty's Method. Pretty (1979) analyzed the delays to pedestrians at signalized
intersections using relatively simple models. For pedestrians crossing one street at an
intersection, he developed the following formula for pedestrian delay, based on uniform
arrival rates and equal pedestrian phases:
where:
d1 = total delay to pedestrians crossing one street, ped-h/h;
P = pedestrian volume crossing one street, peds/h;
C = cycle length, s; and
w = WALK time, s.
For pedestrians crossing two streets at an intersection, he offers the following formula,
which assumes that one-half the cycle length separates the two WALK periods:
d2 = Pd (0.75C - w)2
where:
d2 = total delay to pedestrians crossing two streets successively, ped-h/h;
Pd = pedestrian volume crossing two streets, ped/h.
For an all-pedestrian phase, sometimes referred to as a "barn dance" or "Barnes dance,"
the total pedestrian delay is of the same form as that for a single crossing:
where:
d1&2 = total delay to pedestrians crossing two streets diagonally, ped-h/h.
Delay: Dunn and Pretty's Method. Dunn and Pretty (1984) determined the following
formulas for pedestrian delay at signalized pedestrian (Pelican) crossings:
for a narrow roadway (about 7.5 m or two lanes)
for a wider roadway (about 15 m or four lanes)
where:
d = average delay per pedestrians, s; and
g = vehicular green signal.
The parenthetical expressions in the denominator represent the cycle length for the above
expressions, which assume pedestrian signal compliance.
Delay: Griffiths et al.'s Method. Griffiths et al. (1984a) conducted field surveys of
delay at midblock pedestrian crossings in Great Britain. Figure 9 shows the results of the
authors' field study. The top graph represents zebra crossings. The middle graph
represents fixed-time pelican crossings. The lower graph represents vehicle-actuated
pelican crossings.
As mentioned earlier, Griffiths et al. (1984c) performed extensive simulation analyses on
a 10-m-wide pelican crossing. The authors found an increase in pedestrian delay with
increases in vehicle flow, because the former group enjoys reduced opportunities to cross
in gaps in traffic under higher vehicle flow conditions. The authors found a moderate
increase in pedestrian delay with increases in pedestrian flow. Under vehicular actuation,
the authors found that an increase in vehicular green from 20 to 40 s (in response to
higher vehicle flows) resulted in rapid increases in pedestrian delay above two-way flows
of 1,000 veh/h but no change in pedestrian delay below these levels.
Figures 10 and 11 graphically depict the results of their simulation analyses on pelican
crossings. Figure 10 refers to a fixed-time pelican crossing with a vehicle precedence
period (fmax) of 20 s. Figure 11 shows a vehicle-actuated pelican crossing, with the solid
lines representing fmax = 20 s and the dashed line representing fmax = 40 s. The latter
figure shows the effect of increasing pedestrian flow on pedestrian delay at higher vehicle
flows.
This report has also mentioned that Griffiths et al. (1985) developed mathematical
expressions for delay at fixed-time signal crossings. The appendix describes these
formulas.
Delay: Roddin's Method. Roddin (1981) offers the following equation for average delay
(D) to pedestrians at signalized intersections:
where:
F = fraction of pedestrians who wait when they arrive at a red, amber, or flashing
DON'T WALK signal; i.e., compliant pedestrians;
R = duration of red or DON'T WALK signal;
A = duration of amber or flashing DON'T WALK signal; and
G = duration of green or WALK signal.
Roddin assumes random pedestrian crossings during WALK and random vehicle arrivals
throughout the cycles.

FIGURE 9
Field Measurements of pedestrian delay at midblock crossings in Great Britain

FIGURE 10
Simulation results of pedestrian dealy at fixed-time Pelican crossings in Great
Britain
Delay: Virkler's Method. Virkler clearly states that "pedestrians can save significant
amounts of delay by using more than just the WALK interval to enter the intersection."
He develops a new model of pedestrian delay that reflects the benefits of noncompliance
on pedestrian delay:
where:
D = average delay per pedestrian, s;
C = cycle length, s;
G = duration of WALK signal; and
A = duration of flashing DON'T WALK signal.
Virkler applied this equation to actual measured delay at 18 Brisbane, Australia,
crosswalks and found that the equation predicted delay about 1 percent higher than
observed values
FIGURE 11
Simulation results of pedestrian delay at vehicle-actuated Pelican crossings in Great
Britain
Delay: NCSU's Method. Gerlough and Huber (1975) discuss several intersection delay
and queueing models for vehicles. They derive a fluid (or continuum) delay model for a
pretimed signal, and then note that this formulation is identical to the first term of the
famous Webster analytical model for computing delay. One can write this portion of the
model as:
where:
d = delay, s;
C = cycle length, s;
g = effective green time, s;

x = q C / (g s);
q = arrival rate on approach, ped/h; and
s = saturation flow rate on approach per/hour green.
The NCSU research team observed flow rates up to 5,000 ped/h ped-green at some
locations. Inserting this value for the maximum pedestrian saturation flow rate, one has:
However, the NCSU team observed no capacity constraints, even with pedestrian flow
rates of 5,000 ped/h. Therefore, rather than substitute this value for a maximum saturation
flow rate, one could alternatively assume that the maximum saturation flow rate(s)
approaches infinity for pedestrians. In this case, the term in brackets {1 - q/s} will tend to
unity as s approaches infinity, and the following simple formula remains:
d = r2 / 2C
This last expression is identical to that found in the Australian Road Research Board
Report 123 for pedestrian delay (Akçelik, 1989).
Space.The HCM contains a detailed analysis procedure for determining the space
measure of effectiveness for pedestrians for signalized crossings. Although the general
time-space framework appears sound, several researchers have noted problems associated
with particular aspects of the procedure. These areas include street corner waiting areas,
corner circulation times, start-up times, and minimum crossing times.
Space: Fruin, Ketcham, and Hecht's Method. Fruin, Ketcham, and Hecht (1988)
recommend several changes to the HCM method based on time-lapse photographic
observations of Manhattan Borough, New York City, street corners and crosswalks. First,
they advocate the use of 7 ft2/person (about 0.65 m2/person) for standing area on a street
corner, rather than the HCM value of 5 ft2/person (0.46 m2/person). They also
recommend a change in corner circulation time from a constant of 4 s to the following
formula based on corner dimensions:
To = 0.12 (Wa + Wb) + 1.4
where:
To = circulation time (s); and
Wa, Wb = intersecting sidewalk widths (ft).
Metricized, the equation becomes:
To = 0.37 (Wa + Wb) + 1.4
where:
Wa, Wb = intersecting sidewalk widths (m)
Space: Virkler's et al. Methods. Virkler, Elayadath, and Geethakrishnan (1995) note
that the signalized intersection chapter of the HCM, among other references, contains a
basic crossing time (T) equation of the following general form:
T=D+L/u
where:
D = initial startup delay, s
L = walking distance, m or ft; and
u = walking speed, m/s or ft/s.
Virkler and Guell (1984) provide a method for determining intersection crossing time (T)
that incorporates platoon size:
T = t + (L/V) + H (N/W)
where:
T = crossing time;
t = pedestrian startup time = 3 s (from TRB, 1980);
L = length of crosswalk, ft or m;
V = walking speed = 4.5 ft/s or 1.5 m/s;
H = time headway between persons = 6.7 s / (ped/ft) or 2 .0 s / (ped/m). A later article
(Virkler et al., 1995) uses a higher value for H of 2.61 s/(ped/m);
N = number of pedestrians crossing during an interval; and
W = crosswalk width, ft or m.
Virkler, Elayadath, and Geethakrishnan (1995) note that the Virkler and Guell equation
does not address the problem of opposing platoons meeting in a crosswalk. In addition,
these authors state that the current HCM time-space methodology suffers from two flaws
dealing with the available time-space and walking time. Concerning the former, Virkler
et al. (1996) believe that the HCM methodology overestimates the available time-space
by about 20 percent, because legally crossing pedestrians cannot reach the space in the
center of the crosswalk at the beginning of the phase and must have cleared this space by
the end of the phase. Regarding the latter, Virkler et al. note that the time-space product
ignores the fact that pedestrians must have sufficient time to physically traverse the entire
length of the crosswalk. They imply that one
should subtract the quotient of the crosswalk length and twice the assumed walking speed
from the crosswalk time-space product for accuracy.
Those authors advocate the use of an approach based on shockwave theory when crossing
pedestrian platoons are large, perhaps seven pedestrians per platoon (very roughly, 300
per hour). The shockwave assumptions include: opposing platoons occupy the full
walkway width until they meet and one-half of the walkway width upon meeting,
pedestrian speeds fall upon meeting and remain low after platoon separation, and
pedestrian density increases at platoon meeting. The Appendix contains the expression
for required effective green time from Virkler et al. This model implies that, for a 60-s
cycle length and a 30-s effective green time, 1,000 ped/h in the major direction would
result in inadequate crossing time for crosswalk lengths greater than about 16 m despite
an average LOS of B and a surge LOS of C. In addition, 250 ped/h in the major direction
would have inadequate time for crosswalk lengths above 27 m, even with a surge LOS of
A.
More recently, Virkler conducted a study in the Brisbane, Australia, area to determine
appropriate crossing time parameters for two-way and scramble (all-pedestrian phase)
crosswalk flow (Virkler, 1997c). Virkler notes that, for both types of facilities, the width
of crosswalk actually used by pedestrians increases with increasing crosswalk volume. As
mentioned earlier, Virkler also states that vehicles often use a portion of the crosswalk
during the pedestrian phase. He therefore cautions that engineers should treat measured
crosswalk widths merely as the width "intended for pedestrian use" rather than as an
exact measurement of the width pedestrians will actually use. As an aside, Pretty argued
against the use of exclusive pedestrian phases because of the considerable increase in
pedestrian delay (Pretty et al., 1994).
As mentioned in the Literature Review for Chapter 13, Pedestrians, of the Highway
Capacity Manual, Virkler (1997c) found that speeds at the rear of the platoons are not
independent of concurrent or opposing platoon sizes. He found that, with large platoon
sizes, the typical 7-s WALK interval is insufficient to allow all pedestrians to enter the
crosswalk. For platoons of about 15 people or more, he states that the engineers should
extend the minimum crossing time on a typical 3-m-wide crosswalk by 0.27 s/ped
headway plus 1.71 s. The Appendix contains these calculations.
Unsignalized Crossings
The existing HCM contains no procedures for analyzing these facilities.
Delay: Roddin's Method. Roddin (1981) describes a method by another researcher for
calculating moderate (less than 18 s) mean pedestrian delay (D) at unsignalized
intersections:
D = 6.7 x 10-4 (Q - 0.3)
where:
Q = total hourly vehicle flow, both directions, if less than 1,600/h.
Delay: Virkler's Method. Virkler (1996) describes a similar equation for calculating
delay from other research, based on queueing theory. Assuming random vehicle arrivals
and normal crossing speeds, the expression is:
D = 6.7 x 10-6 Q2 + 0.3
where:
D = delay, s; and
Q = total hourly vehicle flow, both directions.
Delay: Griffiths' et al. Method. As described earlier, Griffiths et al. (1984b) performed
extensive simulation analysis on zebra crossings. They found that pedestrian delay
depends heavily on both pedestrian and vehicle flows; however, they noted that the effect
of increasing vehicle flow occurs primarily at low pedestrian volumes. In fact, as vehicle
volumes continue to increase, pedestrian delay actually decreases, because most vehicles
begin from a stopped (queued) position and pedestrians can establish precedence easier.
Figure 12 depicts the authors' field results.

FIGURE 12
Simulation results of pedestrian delay at Zebra crossings in Great Britain
This report has also mentioned that these same authors developed mathematical delay
models. The Appendix contains the expression developed by Griffiths et al. for pedestrian
delay at a zebra crossing.
Smith's et al. Method. Smith et al. (1987) refer to an earlier study that demonstrated the
effect of crossing width and conflicting vehicle volume on pedestrian delay (Figure 13).
Palamarthy's et al. Method. Palamarthy et al. (1994) present the following model for
mean pedestrian delay for all pedestrians employing one of the crossing tactics mentioned
earlier in the discussion of unsignalized service measures of effectiveness:
where:
di = mean delay to pedestrians using tactic i, s per pedestrian;
q = vehicular flow in one direction, vehicles/s;
"n = critical gap in near-lane traffic stream, s; and
"f = critical gap in far-lane traffic stream, s.
It follows that the mean delay across all tactics is:
where:
d = mean delay to all crossing pedestrians, s/pedestrian; and
.

FIGURE 13
Effect of crossing width and conflicting vehicle volume on pedestrian delay
NCSU's Method. The NCSU research team has developed a formulation for computing
pedestrian delay at unsignalized intersections based on gap acceptance by platoons. Since
"delays are relatively insensitive to the form of the distribution of the arriving traffic"
(Gerlough and Huber, 1975), the research team assumed random arrivals for both
pedestrians and vehicles. In addition, the procedure described in the following paragraphs
assumes that start-up times, headways, walking speeds, and minimum pedestrian body
ellipses retain constant values.
The ITE Manual of Traffic Engineering Studies (Robertson et al., 1994) contains a
general equation describing the minimum safe gap (G) in traffic:
G = (W/S) + (N-1) H + R
where:
W = crossing distance or width of roadway, ft or m;
S = walking speed, ft/s or m/s;
N = predominant number of rows (group size);
H = time headway between rows, s; and
R = pedestrian start-up time, s.
Gerlough and Huber (1975) note that, for a group of pedestrians, the pedestrian and
vehicle volume together determine the size of the platoon:
where:
E(nc) = size of typical pedestrian crossing platoon, ped;
p = pedestrian flow rate, ped/h;
q = vehicular flow rate, veh/h;
One can make an estimate of a critical gap, , for a single pedestrian by substituting N = 1
into the ITE equation above and simplifying:
Then, one substitutes this value for critical gap, ,into the Gerlough and Huber expression
above to determine the number of pedestrians in a typical crossing platoon. To determine
the spatial distribution of pedestrians, the research team developed a simple geometric
expression incorporating the crosswalk width and the pedestrian body buffer zone:
crosswalk width
As stated earlier, the research team recommends a value of 0.75 m2 for a design body
buffer zone.
Given the critical gap for a single pedestrian computed previously, the ITE equation
simplifies to:
G = + (N-1) H
The ITE Manual suggests 2 s as a typical value of headway, H. To avoid confusion, this
report will refer to the pedestrian group critical gap (G in the previous equation) as G .
The final issue concerns the average delay to all pedestrians, whether waiting or not.
Again, Gerlough and Huber (1975) provide guidance:
where:
E(t) = average delay to all pedestrians, s;
T = 1/q = mean vehicle headway, 1/s; and
Other Waiting Areas

Space. The existing HCM does not contain detailed analysis procedures for waiting areas,
because the methodology is extremely simple. One simply computes the available
waiting area and determines the actual or expected number of pedestrians during the
critical time period, and then determine the LOS from the average space per pedestrian.
Fortunately, queueing areas sufficiently resemble street corners such that one can apply
those procedures if needed.
4.3 Pedestrian Networks
Travel Time: Roddin's Method. Roddin (1981) mentions one quantitative factor, travel
time, in the evaluation of pedestrian transportation. His narrative implies that the
following equation applies to pedestrian networks:
Total travel time = Number of ped x (Route length / Walking speed + Signal Delay)
where route length is:
estimated from plans,
generally < 3000 ft (915 m),
generally < 1.4 x straight-line distance, ideally < 1.2 x straight-line distance,
weighted by proportion of ped using alternate routes if available
and:
signal delay is as computed by the method presented earlier.
Travel Time: Virkler's Method. Virkler (1997b) provides an extensive method of
calculating travel time along a pedestrian network. Incorporating both link and note
components, his methodology determines the total walking plus queueing time along the
extended pedestrian facility. For congruence with vehicle arterial measures of
effectiveness, the method determines the average travel speed along the route as a final
step.
Virkler notes that platooning due to an upstream signal can either increase or decrease
pedestrian delay at a downstream signal, depending on the offset and the green time at the
upstream signal (1997d). He argues that one can use field measurements of arrival
patterns at signals to modify random arrival-based delay results. Table 24 shows his
recommended default delay adjustment factors (DFs) to achieve positive pedestrian
platooning.
Examination of the table demonstrates that DF between 0.45 and 0.64 lie within the likely
range at all listed green time/cycle length ratios. In addition, the table demonstrates that
one will achieve better (lower) delay adjustment factors at higher green ratios (g/C). He
notes that the best offsets for pedestrian progression do not necessarily occur when one
achieves the highest arrival rate during the green; rather, one must consider the green
time itself. Virkler found that, as green times increase, the best offsets are shorter, in
order to maximize the benefits of pedestrian platooning.
TABLE 23
Default values of Delay Adjustment Factors (DF) for positive pedestrian platooning
g/C Default DF Likely Range of DF Values

0.1 0.65 0.45-0.80
0.2 0.57 0.38-0.77
0.3 0.5 0.30-0.74
0.4 0.42 0.23-0.72
0.5 0.35 0.16-0.68
0.6 0.27 0.12-0.64
SOURCE: Virkler, 1997d.
5 SUMMARY OF RECOMMENDATIONS FOR DESIGN AND/OR ANALYSIS

OF PEDESTRIAN FACILITIES
Recommendation Page(s) Figure Table HCM Ch. 13 HCM variables and
Subsections adjustments affected
affected
Body ellipse for 4 2 - introductory primarily a design
standing areas narrative only recommendation
Body buffer zone 4 - - walkways, street walkway LOS E/F
for walking corners, threshold changes in
crosswalks Table 13-3
Crosswalk walking 5 - 1 [Ch. 9: new values replace
speeds Methodology, 4.0 ft/s in equation
Input module] (eq.) 9-8
Ch. 13: new values replace
introduction, 4.5 ft/s in eq. 13-14
crosswalks
Grade and stairs 7 - - walkway speeds decrease by
walking speeds narrative, 0.1 m/s in eq. 13-14
crosswalks with grades
Crossing speeds for 7 - - N/A 1 no change 1
platoons
Pedestrian start-up 7 - - N/A 1 no change 1
time
Capacity thresholds 8 - - walkways, street walkway LOS E/F
corners, threshold changes in
crosswalks Table 13-3
Temporal flow 8 - - N/A 1 no change 1
variation
LOS (Level of 11 5 4 walkways walkway LOS A/B,
Service) for E/F thresholds
walkways change, Table 13-3
LOS for walkways 12,13 - 6 walkways new table replaces
with platoons equation 13-3
LOS for 15 - 8 walkways (new new table applies to
transportation measure) terminals with
terminals platoon flow
LOS for stairs 16 - 9 walkways (new new table applies
measure) only to stairs
LOS for crossflows 17 - 10 walkways (new new table serves as
measure) secondary check for
walkways
LOS for mixed-use 21 - 13 walkways (new new table applies
paths measure) only to mixed-use
paths
Noncompliance 26 - - street corners, minor, major red
time adjustments crosswalks times in equations
13-6, 13-7 change;d
effective red time
reduced in computing
ped delay
LOS for signalized 29 - 18 street corners (new new table based on
crossingsa measure) ped delay; space now
secondary
Swept-path method - - crosswalks caution to use only
for vehicle effects under aggressive
30 driver behavior
LOS for 32,33 - 20 street corners (new new table based on
unsignalized measure) ped delay
crossingsb
LOS for pedestrian 35,36 - 22 networks (new new table shows
networksc section) proposals for analysis
of ped networks
Ped delay at 45 - - street corners (new method for
signalized crossings measure) computing ped delay
Effective crosswalk 48 - - crosswalks equation 13-13
time-space corrected; calculated
TSw will decrease
Crossing time in 49, 60 - - crosswalks new equations
platoons replace eq. 13-14
with large platoons
Ped delay at 54 - - street corners (new method for
unsignalized measure) computing ped delay
crossings
a
Chapter 9, "Signalized Intersections"
b
Chapter 10, "Unsignalized Intersections"
c
Offers a comparison with Level of Service for drivers computed in HCM Chapter 11,
"Urban and Suburban Arterials"
d
Current HCM is ambiguous regarding the definition of minor and major red times (Rmi,
Rmj); therefore, the effect of the proposed noncompliance
adjustments will depend on the analyst's interpretation of the HCM
6 APPENDIX: ADDITIONAL FORMULAS FOR COMPUTING

RECOMMENDED MEASURES OF EFFECTIVENESS
Interrupted Pedestrian Facilities

Delay: Griffiths et al.'s Method. Griffiths et al. (1985) derived the following expression
for pedestrian delays with two-way vehicle volumes below 1,500 vehicles/h:
For vehicle volumes at or above 1,500/h (with pedestrian noncompliance less likely at
these high vehicle volumes), they found that the following formula best fit their
simulation results:
where:_
dp = mean overall delay to pedestrians, s;
V = vehicle mean two-way arrival rate, veh/h;
µ = pedestrian mean arrival rate, ped/s;
dT = µFDWd{a + b + e + f + d/2} + µR(e + f) {a + b + (e + f) /2} + µR(a + b)2/2
+ (a + b)e-{µFDWd + µR(e+f)}
_
y2 = a + b+ c+ k+ (1/_µR)e-{µFDWd + µR(e+f)}
µFDW = pedestrian flow rate during flashing DON'T WALK, ped/s;
µR = pedestrian flow rate during steady DON'T WALK, ped/s;
a = vehicular yellow time, s;
b = all red period, s;
c = WALK time, s;
d = flashing DON'T WALK time during vehicle red indication, s;
e = flashing DON'T WALK time during vehicle "yield to peds" indication, s;
f = vehicle green time, s;
k = pedestrian effective red time = d + e + f
Under vehicle actuation, they found the following best matched simulation results:
where all variables and parameters are as before, except:_
k = d + e + fmin ; and
fmin = minimum vehicular green, s.
Space: Virkler et al.'s Method. Virkler, Elayadath, and Geethakrishnan (1995) offer the
following expression for required effective green time (Greq):
Greq = t0 + t1 + t2
where:
and where:
R = effective red time, s;
qA = flow rate of peds approaching the queue, ped/min;

qC = flow rate of peds leaving the queue, ped/min;
kB = density of arriving pedestrians, ped/m2;
kA = density of platooned pedestrians, ped/m2;
Lq = maximum depth of the standing queue (waiting to cross), m;
L = walking distance, m;
uC = pedestrian walking speed before meeting opposing platoon, m/s; and
uD = pedestrian walking speed after meeting opposing platoon, m/s.
Regarding space-based methods at signalized intersections, Virkler assumes 1.2 m/s
platoon flow speeds in the following calculation, which reflects this:
WALK interval = 3.2 s + (0.19 s/ped) * N1
where:
N1 = number of people in the primary movement who arrive before the WALK
indication and exit the curb during the WALK indication.
For larger effective crosswalk widths, he offers the following modification:
WALK interval = 3.2 s + (0.57 s/ped/m) * (platoon size/W)
where:
W = effective crosswalk width, m.
Virkler then offers the following method of determining sufficient total crossing time
(WALK plus flashing DON'T WALK), which accounts for the effects of dispersion of
platoons larger than 15 persons, for crosswalks with effective widths up to about 3 m:
Recommendation.
This report recommends the above methods of determining sufficient total crossing
time(WALK plus flashing DON'T WALK) proposed by Virkler into the HCM.
Unsignalized Crossings
Delay: Griffiths et al.'s Method. Griffiths et al. (1985) established the following
expression for pedestrian delay at a zebra crossing:
where:
G = mean pedestrian group size, ped.

The exponential portion of the expression reflects the authors' observation that pedestrian
groups experience no delay when their arrival at the curbside occurs before a preceding
pedestrian group has reached about halfway across the road. 7
7 ACKNOWLEDGMENTS
The authors would like to thank several graduate students at North Carolina State
University who were instrumental to the successful completion of this report. Conredge
Lewis helped with field observations in Chicago and Washington. Aemal Khattak offered
input into the development of recommended walking speed values for elderly pedestrians.
Craig Sheffler gave additional information regarding shared-use paths. Tahsina Ahmed
assisted with the development of pedestrian delay models. Brian Eads provided
invaluable administrative support. Michael Surasky and Steven Click offered assistance
regarding the appearance of the report.
We would also like to thank members of the sponsoring agencies for this report. David
Harkey and Charles Zegeer of the University of North Carolina Highway Safety Research
Center and Carol Tan Esse of the Federal Highway Administration provided the
necessary support and guidance for the project. We would also like to thank our
reviewers who gave excellent feedback on a wide variety of issues.
The authors would like to thank Dr. Alex Sorton of Northwestern University and Dr.
Mark Virkler of the University of Missouri at Columbia for providing timely assistance
and information throughout this project. Alex Sorton and Dan Burden also provided the
research team with excellent photographs for inclusion in the document.
8 REFERENCES
Akçelik, Rahmi. "ARR 180: Calibrating SIDRA," 2nd edition, 1st reprint. Australian
Road Research Board, 1993.
Akçelik, Rahmi. "ARR 123: Traffic Signals: Capacity and Timing Analysis," 4th reprint.
+Australian Road Research Board, 1989.
Allen, D. Patrick. "The Effect of Bicycles on the Capacity of Signalized Intersections."
Master's Thesis, North Carolina State University, Raleigh, 1996.
Arasan, V. Thamizh and K. Jagadeesh. "Effect of Heterogeneity of Traffic on Delay at

Signalized Intersections." Journal of Transportation Engineering, September-October
1995.
Botma, Hein. "Method to Determine Levels of Service for Bicycle Paths and Pedestrian-
Bicycle Paths." In Transportation Research Record 1502, Transportation Research
Board, 1995.
Bowman, Brian and Robert Vecellio. "Pedestrian Walking Speeds and Conflicts at Urban
Median Locations." In Transportation Research Record 1438, Transportation Research
Board, 1994.
Brilon, Werner. "A New German Capacity Manual." The Second International
Symposium on Highway Capacity, 1994 Proceedings.
Coffin, Ann and John Morall. "Walking Speeds of Elderly Pedestrians at Crosswalks." In
Transportation Research Record 1487, Transportation Research Board, 1995.
Davis, Dennis and John Braaksma. "Level-of-Service Standards for Platooning

Pedestrians in Transportation Terminals." In ITE Journal, April 1987.
Davis, Scott, L. Ellis King, and H. Douglas Robertson. "Predicting Pedestrian Crosswalk
Volumes." In Transportation Research Record 1168, Transportation Research Board,
1988.
Dixon, Linda B. "Bicycle and Pedestrian Level of Service Performance Measures and
Standards for Congestion Management Systems." In Transportation Research Record
1538, Transportation Research Board, 1996.
Dunn, Roger and Robert Pretty. "Mid-Block Pedestrian Crossings - An Examination of

Delay." In the 12th Annual Australian Road Research Board Conference (at Hobart,
Tasmania, Australia) Proceedings, August 1984.
Dunne, M.C. and D.J. Buckley. "Delays and Capacities at Unsignalized Intersections." In
the 6th Annual Australian Road Research Board Conference (at Canberra, Australia)
Proceedings, 1972.
Florida Department of Transportation (DOT). "Twelve Steps Toward Walkable

Communities." April 1995.
Franck, Lukas. "Intersection Design, Signalling, and Control Issues for Pedestrians Who
Are Blind." Presentation at the ITE International Conference on Transportation and
Sustainable Communities, Tampa, Florida, March 1997.
Fruin, John J. "Pedestrian Planning and Design." Metropolitan Association of Urban
Designers and Environmental Planners, New York, N.Y., 1971.
Fruin, John J. and Gregory Benz. "Pedestrian Time-Space Concept for Analyzing
Corners and Crosswalks." Transportation Research Record 959, Transportation Research
Board, 1984.
Fruin, John J., Brian Ketcham, and Peter Hecht. "Validation of the Time-Space Corner
and Crosswalk Analysis Method." In Transportation Research Record 1168,
Transportation Research Board, 1988.
Gerlough, Daniel and Matthew Huber. "Transportation Research Board Special Report
165: Traffic Flow Theory: A Monograph." Transportation Research Board, 1975.
Gordon, Stewart and H. Douglas Robertson. "A Study of Driver Noncompliance with
Traffic Signals." In Transportation Research Record 1168, Transportation Research
Board, 1988.
Griffiths, J.D., J.G. Hunt, and M. Marlow. "Delays at Pedestrian Crossings: 1. Site
Observations and the Interpretation of Data." In Traffic Engineering and Control,
July/August 1984a.
Griffiths, J.D., J.G. Hunt, and M. Marlow. "Delays at Pedestrian Crossings: 2. The
Development and Validation of a Simulation Model of a Zebra Crossing." In Traffic
Engineering and Control, October 1984b.
Griffiths, J.D., J.G. Hunt, and M. Marlow. "Delays at Pedestrian Crossings: 3. The
Development and Validation of a Simulation Model of a Pelican Crossing." In Traffic
Engineering and Control, December 1984c.
Griffiths, J.D., J.G. Hunt, and M. Marlow. "Delays at Pedestrian Crossings: 4.

Mathematical Models." In Traffic Engineering and Control, May 1985.
Hunt, John C. and J.D. Griffiths. "Vehicle Capacity at Pedestrian Crossings in the United
Kingdom," Highway Capacity and Level of Service, 1991.
Institute of Transportation Engineers (ITE). "Characteristics and Service Requirements of

Pedestrians and Pedestrian Facilities." ITE Technical Council Committee 5-R
Informational Report, Edmund F. Cantilli, chairman. Traffic Engineering, May 1976.
Kaiser, Steven H. "Urban Intersections That Work for Pedestrians: A New Definition for
Level of Service." Informal paper presented at 1994 Transportation Research Board
Annual Meeting, Session 15.
Khisty, C. Jotin. "Evaluation of Pedestrian Facilities: Beyond the Level-of-Service

Concept." In Transportation Research Record 1438, Transportation Research Board,
1994.
Khisty, C. Jotin. "Pedestrian Cross Flows in Corridors." In Transportation Research

Record 847, Transportation Research Board, 1982.
Knoblauch, R., M. Nitzburg, R. Dewar, J. Templer, and M. Pietrucha. "Older Pedestrian

Characteristics for Use in Highway Design." Office of Safety and Traffic Operations
Research and Development, Federal Highway Administration, February 1995.
Knoblauch, Richard, Martin Pietrucha, and Marsha Nitzburg. "Field Studies of Pedestrian
Walking Speed and Start-up Time." In Transportation Research Record 1538,
Transportation Research Board, 1996.
Kyte, Michael, Zongzhong Tian, Zakir Mir, Zia Hameedmansoor, Wayne Kittelson,
Mark Vandehey, Bruce Robinson, Werner Brilon, Lothar Bondzio, and Rod Troutbeck.
"Capacity and Level of Service at Unsignalized Intersections." National Cooperative
Highway Research Program Project 3-46, Volume Two, Final Report, December 1995.
Kyte, Michael, Zongzhong Tian, Zakir Mir, Zia Hameedmansoor, Wayne Kittelson,
Mark Vandehey, Bruce Robinson, Werner Brilon, Lothar Bondzio, Ning Wu, and Rod
Troutbeck. "Highway Capacity Manual: Chapter 10 - Unsignalized Intersections, 1997
Update, Draft 5", December 1996. (Based on National Cooperative Highway Research
Program Project 3-46, Volume One, Final Report, April 1996.)
Lam, William H.K., John F. Morrall, and Herbert Ho. "Pedestrian Flow Characteristics in
Hong Kong." Transportation Research Record 1487, Transportation Research Board,
1995.
Lautso, Kari and Pentti Murole. "A Study of Pedestrian Traffic in Helsinki: Methods and
Results." In Traffic Engineering and Control, January 1974.
MacLean, A. Scott and Don Howie. "Survey of the Performance of Pedestrian Crossing
Facilities." In Australian Road Research, Vol. 10, No. 3, September 1980.
Middleton, G. "Pedestrian Crossings at Signalised Intersections in Australia: The Need

for Change." In Australian Road Research, Vol. 11, No. 3, September 1981.
Milazzo II, Joseph S. "The Effect of Pedestrians on the Capacity of Signalized

Intersections." Master's Thesis, North Carolina State University, Raleigh, 1996.
Navin, Francis P.D. and R. J. Wheeler. "Pedestrian Flow Characteristics." In Traffic

Engineering, June 1969.
Noland, Robert. "Pedestrian Travel Times and Motor Vehicle Traffic Signals." In
Palamarthy, Srinivas, Hani Mahmassani, and Randy Machemehl. "Models of Pedestrian
Crossing Behavior at Signalized Intersections." Research Report 1296-1. Center for
Transportation Research, Austin, Texas, January 1994.
Polus, Abishai, Joseph Schofer, and Ariela Ushpiz. "Pedestrian Flow and Level of
Service." Journal of Transportation Engineering, American Society of Civil Engineers,
January 1983.
Pretty, Robert. "The Delay to Pedestrians and Vehicles at Signalized Intersections." In

ITE Journal, May 1979.
Pretty, Robert, Brigid Broekstra, and Shane Healey. "Pedestrians at Signalized

Intersections." Proceedings of the Second International Symposium on Highway
Capacity, 1994.
Pushkarev, Boris, with Jeffrey Zupan. "Urban Space for Pedestrians." Cambridge,
Massachusetts: The MIT Press, 1975a.
Pushkarev, Boris and Jeffrey Zupan. "Capacity of Walkways." In Transportation

Research Record 538, Transportation Research Board, 1975b.
Robertson, H. Douglas, Joseph Hummer, and Donna Nelson. "Manual of Transportation

Engineering Studies." Institute of Transportation Engineers, 1994.
Roddin, Marc. "NCHRP Report 240: A Manual to Determine Benefits of Separating

Pedestrians and Vehicles." Transportation Research Board, 1981.
Rouphail, Nagui. "Midblock Crosswalks: A User Compliance and Preference Study." In

Rouphail, Nagui, Joseph Hummer, Joseph Milazzo II, and D. Patrick Allen. "Literature
Review for Chapter 13, Pedestrians, of the Highway Capacity Manual." Federal Highway
Administration Report, February 1998.
Sarkar, Sheila. "Determination of Service Levels for Pedestrians, With European

Examples." In Transportation Research Record 1405, Transportation Research Board,
1993.
Sarkar, Sheila. "Measuring Comfort Levels for Pedestrians in Urban Sidewalks."

Unpublished work obtained from the author, 1996.
Smith, S.A., K.S. Opeila, L.L. Impett, M.T. Pietrucha, R. Knoblauch, and C. Kubat.
"NCHRP Report 294A: Planning and Implementing Pedestrian Facilities in Suburban and
Developing Rural Areas - Research Report." Transportation Research Board, 1987.
Song, Lida, M.C. Dunne, and J.A. Black. "Models of Delay and Accident Risk to
Pedestrians." In Transportation and Traffic Flow Theory, 1993.
Tan Esse, Carol and Charles Zegeer. "European Practices and Innovations for Pedestrian
Crossings." ITE Journal, November 1995.
Tanaboriboon, Yordphol and Jocelyn A. Guyano. "Level of Service Standards for

Pedestrian Facilities in Bangkok: A Case Study." ITE Journal, November 1989.
Tanaboriboon, Yordphol, Sim Siang Hwa, and Chin Hoong Chor. "Pedestrian
Characteristics Study in Singapore." In Journal of Transportation Engineering, May
1986.
Transportation Research Board (TRB). Interim Materials on Highway Capacity.

Transportation Research Circular 212, 1980.
Transportation Research Board (TRB). "Highway Capacity Manual." Special Report 209,
1994 update to 1985 edition.
Viney, N.D. and Robert Pretty. "Saturation Flow of a Movement Subject to a Pedestrian
Stream at Traffic Signals." In the 11th Annual Australian Road Research Board
Conference (at Melbourne, Victoria, Australia) Proceedings, August 1982.
Virkler, Mark. "Pedestrian Compliance Effects on Signal Delay." Unpublished work

obtained from the author, 1997a. (Submitted for possible publication at the Annual
Meeting of the Transportation Research Board at Washington, D.C., in January 1998.)
Virkler, Mark. "Prediction and Measurement of Travel Time Along Pedestrian Routes."
Unpublished work obtained from the author, 1997b. (Submitted for possible publication
at the Annual Meeting of the Transportation Research Board at Washington, D.C., in
January 1998.)
Virkler, Mark. "Quality of Flow Along Pedestrian Arterials." In the Combined 18th
Annual Australian Road Research Board Transport Research Conference / Transit New
Zealand Land Transport Symposium (at Christchurch, New Zealand), September 1996.
Virkler, Mark "Scramble and Crosswalk Signal Timing." Unpublished work obtained
from the author, 1997c. (Submitted for possible publication at the Annual Meeting of the
Transportation Research Board at Washington, D.C., in January 1998.)
Virkler, Mark. "Signal Coordination Benefits for Pedestrians." Unpublished work

obtained from the author, 1997d. (Submitted for possible publication at the Annual
Meeting of the Transportation Research Board at Washington, D.C., in January 1998.)
Virkler, Mark. and Rajesh Balasubramanian. "Flow Characteristics on Shared Hiking /

Biking / Jogging Trails." Unpublished work obtained from the senior author, 1997.
(Submitted for possible publication at the Annual Meeting of the Transportation Research
Board at Washington, D.C., in January 1998.)
Virkler, Mark and Sathish Elayadath. "Pedestrian Speed-Flow-Density Relationships."

Transportation Research Record 1438, Transportation Research Board, 1994a.
Virkler, Mark and Sathish Elayadath. "Pedestrian Density Characteristics and

Shockwaves." Proceedings of the Second International Symposium on Highway
Capacity, 1994b, Volume 2.
Virkler, Mark, Sathish Elayadath, and Geethakrishnan. "High Volume Pedestrian

Crosswalk Time Requirements." Transportation Research Record 1495, Transportation
Research Board, 1995.
Virkler, Mark and David Guell. "Pedestrian Crossing Time Requirements at

Intersections." In Transportation Research Record 959, Transportation Research Board,
1984.
Wegmann, F.J., K.W. Heathington, D.P. Middendorf, M.W. Redford, A. Chatterjee, and
T.J. Bell. "NCHRP Report 262: Planning Transportation Services for Handicapped
Persons - User's Guide." Transportation Research Board, 1983.
Wigan, Marcus. "Treatment of Walking as a Mode of Transportation." In Transportation

Research Record 1487, Transportation Research Board, 1995.

Recommended Procedures Chapter 13, "Pedestrians," of The Highway Capacity Manual

Uploaded by

Copyright:

Available Formats

Recommended Procedures Chapter 13, "Pedestrians," of The Highway Capacity Manual

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Recommended Procedures Chapter 13, "Pedestrians," of The Highway Capacity Manual

Uploaded by

Copyright:

Available Formats

Recommended Procedures

Chapter 13, "Pedestrians,"

Capacity Analysis of Pedestrian and Bicycle Facilities

D. Patrick Allen, Research Assistant

North Carolina State University

under the direction of:

1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.

RECOMMENDED PROCEDURES FOR CHAPTER 13, February 1998

15. Supplementary Notes

Contracting Officer's Technical Representative: Carol Tan Esse (HSR-20)

pedestrian, level of service, platoon, delay, No restrictions. This document is

Form DOT F 1700.7 (8-72)

2 SUMMARY OF PEDESTRIAN CHARACTERISTICS 4

3 SERVICE MEASURES OF EFFECTIVENESS 9

4 METHODS FOR COMPUTING MEASURES OF EFFECTIVENESS 36

5 SUMMARY OF RECOMMENDATIONS FOR DESIGN AND/OR 56

1.2 Summary of Recommendations for Design and/or Analysis of Pedestrian

RecommendationPage(s)FigureTableHCM Ch. 13 HCM variables and adjustments

This section summarizes the recommendations regarding pedestrian characteristics as

2.1 Pedestrian Space Requirements

Recommended Procedures For Chapter 13

2.2 Pedestrian Walking Speeds

Recommended Procedures For Chapter 13

Grades and Stairs

2.3 Pedestrian Start-up Times

2.4 Pedestrian Traffic Flow Relationships

3.1 Uninterrupted Pedestrian Facilities

Sidewalks and Walkways

LOS (m2/ped) (ft2/ped) (ped/min/m) (ped/min/ft) (m/s) (ft/min) -

United States of America Germany Israel Thailand

LOS (ped/min/m) (ped/min/m) (ped/min/m) (ped/min/m) (ped/min/m)

Space Flow Ratea

Recommended Procedures For Chapter 13

SOURCE: Davis and Braaksma, 1987.

Space Flow Rate Speed

Space Flow Rate Avg. Horiz. Speed v/c ratio

Crossflows. A crossflow is a pedestrian flow that is roughly perpendicular to and crosses

Speed Flowb Density Space

SOURCE: Khristy, 1982.

SOURCE: Adapted from Botma, 1995.

SOURCE: Adapted from Botma, 1995.

Virkler and Balasubramanian (1997) studied flow characteristics on two-way, shared-use

One-Way Paths Two-Way Paths

Perioda Service Volume Perioda Service Volume

3.2 Interrupted Pedestrian Facilities

Recommended Procedures For Chapter 13

Recommended Procedures For Chapter 13

Location De Facto WALK Interval,a s

SOURCES: Viney and Pretty, 1982; Virkler, 1997a.

Delay. Currently, no LOS standard based on pedestrian delay at signalized intersections

Stopped Delay per Vehicle

SOURCE: TRB, 1994.

Recommendation. This study recommends the incorporation of pedestrian delay as a

Average Total Delay

SOURCE: TRB, 1994.

As in the signalized intersection case, it is useful to examine existing research on

Average Delay Per Pedestriana Likelihood of Risk-Taking Behavior

Average Delay Per Pedestriana Likelihood of Risk-Taking Behavior

Space Interperson Spacing