THESiS !l!

MIC CHIGAN STAT TEU

I!!!“ !1!2!!! !!!!!!!.!!!!!!!!!!!!!!!!!!!!!!!!!!!!

301411 2951 ’_

LIBRARY

Michigan State
Univetsl W

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

BUS PREEMPTION SIGNAL (BPS) - AN APPLICATION

OF ADVANCED PUBLIC TRANSPORTATION SYSTEM (APTS)
it

presented by
Khaled A. Al-Sahili

has been accepted towards fulfillment
of the requirements for

Ph D. degree in Waring

 

W (N fig, 44
Major professoK/

Date 7/31/95

 

MSU i: an Affirmative Action/Equal Opportunity Institution 0-12771

PLACE ll RETURN BOXto romavothbchockout’ om ywtncord.
TO AVOID FINES Mum on or baton duo duo.

DATE DUE DATE DUE DATE DUE

 

    
   

rig 0‘ '1 .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

J! I 1

MSU In An Affirmative WM Oppommuy Intuition
Mun-9.1

BUS PREEMPTION SIGNAL (BPS) - AN APPLICATION OF
ADVANCED PUBLIC TRANSPORTATION SYSTEM (APTS)

By

Khaled Ahmad Al-Sahili

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Civil and Environmental Engineering

1995

UMI Number : 9605828

 

UMI Microforn 9605828
Copyright 1995. by UMI Company. All rights reserved.

This licroforn edition is protected against unauthorized
coPying under Title 17. united States Code.

 

UMI

300 Nbrth Zeeb Road
Ann Arbor, MI 48103

ABSTRACT

BUS PREEMPTION SIGNAL (BPS) - AN APPLICATION OF
ADVANCED PUBLIC TRANSPORTATION SYSTEM (APTS)

By

Khaled A. Al-Sahili

With the emergence of Intelligent Transportation System (ITS) technologies, there has
been a renewed interest in the bus priority signal (BPS). However, at present there is no
model capable of simulating various BPS strategies and then restoring the original signal

settings after bus preemption is awarded.

The effect of providing the BPS treatment on the Washtenaw Avenue Corridor in Ann
Arbor, Michigan was studied. The NETSIM graphic animation feature was used to detect
the bus arrival, award preemption, and the signal timing plan was restored to the original
signal setting manually. The model was calibrated using field data and the sensitivity of

the model to several variables was tested.

The corridor’s signal timing was first optimized using the TRANSYT—7F model. The
green extension / red truncation with and without compensation, the skip phase with and
without compensation, and the conditional preemption plans were evaluated. It was found
that signal preemption disrupts traffic progression, and thus increases overall network

vehicle and person delay.

Results of preemption were analyzed on a cycle-by-cycle basis as well as over the entire
simulation period, and the most appropriate preemption plan for each intersection was
determined. The bus travel time and delay were reduced when the optimal BPS plan was

used.

The BPS was tested under different network traffic volumes, different main to cross street
traffic ratios for an isolated intersection, and signal preemption for carpools. It was found
that maintaining progression is most critical under heavy traffic conditions. The traffic

volume criteria that warrant signal preemption were established.

There appears to be advantages to providing carpools with preemption capability up to
between 5 and 10% of the main street traffic volume. In any corridor there is likely to be
random fluctuations in the traffic demand, and this variation may be as large as the mea-

sured effect of BPS.

I dedicate this work to my biggest supporters in this world; my family:

my parents, brother, sisters, children, and wife.

ACKNOWLEDGMENT

I would like to express my deep thanks and appreciation to Professor William C. Taylor,
my academic advisor for his continuous guidance, advice, and support since the start of

my academic program.

Also, I would like to extend my deepest thanks and appreciation for my Graduate Com-
mittee; Professor Thomas Maleck, Professor Richard Lyles, and Professor James Staple-
ton. Thanks for their support, encouragement, and advice. I would like to extend my

appreciation for Professor Francis McKelvey for his support and recommendations.

My great thanks to the Michigan Department of Transportation and the Great Lakes Cen-

ter for Truck and Transit Research, the sponsors of the project, for their financial support.

My deepest love and thanks for my parents, who brought me up in a best manner and set
the best example for me for life, for their unconditional support from the day I was born,

and for their prayers that do not stop.

I can not express my heartfelt thanks to my wife and two boys. They supported me and

provided me with comfort, joy and happiness. They are my inspiration for life.

Above all, my greatest thanks for Allah (God) as He deserves. I certainly believe that any

success I have accomplished in my life is from Him and by His guidance.

TABLE OF CONTENTS

Chapter Page
LIST OF TABLES .................................................................................................. vii
LIST OF FIGURES ................................................................................................ xiii
1. INTRODUCTION ............................................................................................... l
1.1 Research Objectives ............................................................................. 3
1.2 Bus Priority (Preemption) .................................................................... 4
1.2.1 Passive Priority ............................................................................ 5
1.2.2 Active Priority ............................................................................... 6
2. LITERATURE REVIEW ................................................................................... 9
2.1 Isolated Intersections ........................................................................... l()
2.2 Network and Arterials .......................................................................... 17
2.3 Signal Technology ................................................................................. 23
2.4 Summary ................................................................................................ 26
3. RESEARCH APPROACH AND DATA COLLECTION ................................... 36
3.1 Alternative Plans ................................................................................... 36
3.1.1. Automatic Signal /Eagle Signal Software-Hardware Interface. 37
3.1.2 THOREAU Model ......................................................................... 39
3.1.3. TRAF-NETSIM Graphics and Simulation ................................. 39
3.2 Data Collection ...................................................................................... 42
3.2.1 Background ................................................................................... 42
3.2.2 Network Selection ........................................................................ 44
3.2.3 Data Collected .............................................................................. 44
3.3 Model Calibration .................................................................................. 47
3.4 Sensitivity Analysis .............................................................................. 50
4. BPS SCHEMES AND ALGORITHM ............................................................... 55
4.1 Bus Detectors ........................................................................................ 55
4.2 BPS Schemes ......................................................................................... 57
4.3 BPS Logic ............................................................................................... 58
5. BPS SIMULATION RESULTS AND ANALYSIS ........................................... 60
5.1 Study Cases ........................................................................................... 60
5.2 Analysis of BPS Time Period Specific Statistics ................................. 63
5.2.1. Case 1 Preemption ...................................................................... 64
5.2.2. Case 2 Preemption ...................................................................... 69

vi

5.2.3. Case 3 Preemption ...................................................................... 71

5.2.4. Case 4 Preemption ...................................................................... 72
5.2.5. Case 5 Preemption ...................................................................... 73
5.2.6. Case 6 Preemption ...................................................................... 74
5.3 Intersection and Link Overall Statistics .............................................. 74
5.4 Cumulative Network Measures of Effectiveness (MOEs) ................ 85
6. THE DYNAMICS OF BPS ................................................................................ 91
6.1 BPS Sensitivity to Volume .................................................................... 91
6.2 BPS Sensitivity to Volume Ratio .......................................................... 97
6.2.1. BPS Overall Statistics ................................................................ 99
6.2.2. Before and After Analysis .......................................................... 105
6.3 BPS and Carpools ................................................................................. 110
6.4 Test of Random Vehicle Generation .................................................... 114
7. CONCLUSIONS ................................................................................................. 119
8. APPENDICES:
Appendix A: Network Peak Hourly Traffic Volume .................................. 125
Appendix B: BPS Algorithm (Flow-Charts) ............................................ 127
Appendix C: Case 1 Preemption Results .................................................. 137
Appendix D: Case 3 Preemption Results .................................................. 145
Appendix E: Case 4 Preemption Results .................................................. 155
Appendix F: Volume Sensitivity Results .................................................. 165
Appendix G: Results of Carpool ................................................................ 173
9. LIST OF REFERENCES ................................................................................. 175

vii

Table 1

Table 2

Table 3

Table 4

(

Table 5A

Table SB

Table 5C

Table 6

Table 7

Table 8

Table 9

LIST OF TABLES

Summary of Previous BPS Simulation Studies at Isolated
Intersections ...................................................................................... 28

Summary of Previous BPS Field Studies at Isolated Intersections 30

Summary of Previous BPS Simulation Studies for Networks and
Arterials ........................................................................................... 31

Summary of Previous BPS Field Studies for Networks and
Arterials .......................................................................................... 33

West Bound MOEs as a Result of lO-Seconds Preemption at
Intersection #8 (IO-Seconds Green Extension) .......................... 52

East Bound MOEs as a Result of lO-Seconds Preemption at
Intersection #8 (IO-Seconds Green Extension) .......................... 53

North & South Bound MOEs as a Result of lO-Seconds
Preemption at Intersection #8 (IO-Seconds Green Cut) ............. 54

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 10. Main Street (E-W) Green is Extended for
3-Seconds, Cross Street (N-S) is Cut by 3-Seconds. Third Time

Period is Preempted. Bus Direction is East Bound ....................... 65

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 6. lO-Seconds of Advance Green For Main

Street EB, lO-Seconds Cut From Cross Street NB and Main

Street WB left. Third Time Period is Preempted. Bus Direction

is East Bound .................................................................................. 66

Time Period Movement Specific Statistics as a Result of Pre-

emption For Intersection Number 11. Main Street (E-W) R & T

are Advanced for lO—Seconds, Cross Street (S-N) Left Turns are

Cut by 10-Seconds.Third Time Period is Preempted. Bus

Direction is East Bound .................................................................. 67

Time Period Movement Specific Statistics as a Result of Preemption
and Compensation For Intersection Number 11. Main Street (E-W)

R & T are Advanced for lO-Seconds and Then Cut in The 3rd and 4th
Periods, Respectively / Main Street (E-W) Left Turns are Cut by

viii

Table 10

Table 11

Table 12

Table 13

Table 14

Table 15

Table 16

Table 17

Table 18

Table 19

Table 20

Table 21

Table 22

Table 23

Table 24

Table 25

Table 26

lO-Seconds and Compensated in the 3rd and 4th period,

Respectively ....................................................................................

Case 1 Total Link Statistics mm and Without Preemption ........
Case 2 Total Link Statistics With and Without Preemption ........
Case 3 Total Link Statistics With and \Vrthout Preemption

Case 4 Total Link Statistics Mth and VVrthout Preemption ........

Average Delay Over The 45-Minute Simulation Period ...............

Cumulative Network Statistics; With and Without Preemption ..

Cumulative NETSIM Person Measures of Effectiveness; Before

and After Preemption ......................................................................

Cumulative Network-Wide Bus Statistics; With and Without

Preemption ......................................................................................

Total NETSIM Bus LINK Statistics; With and Without

Preemption ......................................................................................

Cumulative Network Statistics; Mth and Without Preemption

Cumulative NETSIM Person Measures of Effectiveness; Before

and After Preemption ......................................................................

Cumulative Network-Wide Bus Statistics; With and Without

Preemption .......................................................................................

Cumulative NETSIM Bus Statistics; With and \Vrthout

Preemption .......................................................................................

Cumulative Network Statistics; With and Without Preemption

for a Different Random Seed Numbers ..........................................

Cumulative NETSIM Person Measures of Effectiveness; Before

and After Preemption for a Different Random Seed Numbers ......

Cumulative Network-Wide Bus Statistics; With and Without

Preemption for a Different Random Seed Numbers ......................

Cumulative NETSIM Bus Statistics; With and “Without

Preemption for a Different Random Seed Numbers ......................

ix

7()

75

77

79

81

84

86

87

88

90

92

93

95

96

115

115

116

116

Table 27

Table 28

Table 29

Table 30

Table A1

Table C]

Table C.2

Table C.3

Table C.4

Table C.5

Table C.6

Summary of Statistics For Several Preemption Plans ................... 121

Summary of The Overall Statistics For The Volume Sensitivity

Test .................................................................................................... 122
Summary of The Impact of Preemption on Carpools ..................... 123
Comparison Between The Results of a Different Random Seed

Number and Case 5 Preemption ..................................................... 124
RM. Peak Hourly Volume Along Washtenaw Avenue ................ 125

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 6. lO-Seconds of Advance Green For Main

Street EB/ lO-Seconds Cut From Cross Street NB and Main

Street WB left. Period Extended is The Third Time Period. Bus
Direction is East Bound .................................................................. 137

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 10. Main Street (E-W) Green is Extended

for 3-Seconds / Cross Street (N-S) is cut by 3-Seconds. Third

Time Period is Preempted. Bus Direction is East Bound ............. 138

Time Period Movement Specific Statistics as a Result of Pre-

emption For Intersection Number 11. Main Street (E-W) R & T

are Advanced for lO-Seconds / Main Street (S-N) Left Turns are

Cut by lO-Seconds. Third Time Period is Preempted. Bus

Direction is East Bound .................................................................. 139

Time Period Specific Statistics as a Result of Preemption For

Intersection Number 6. lO-Seconds of Advance Green For Main

Street EB/ lO-Seconds Cut From Cross Street NB and Main

Street WB Left Turn. Period Extended is The Third Time Period.

Bus Direction is East Bound ................................................................. 140

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 3. Main Street (E-W) Green is Advanced for
3-Seconds / Cross Street (N-S) is Cut by 3-Seconds. Third Time

Period is Preempted. Bus Direction is East Bound ...................... 141

Time Period Specific Statistics as a Result of Preemption For

Intersection Number 6. lO-Seconds of Green Extension For Main

Street EB / lO-Seconds Cut From Cross Street NB and Main

Street WB Left Turn. Period Extended is The Third Time Period.

Bus Direction is East Bound .................................................................. 142

Table C.7

Table C.8

Table D.1

Table D2

Table D.3

Table D4

Table D5

Table D6

Table D7

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 10. Main Street (E-W) Green is Extended for
3-Seconds / Cross Street (N-S) is Cut by 3-Seconds. Third Time

Period is Preempted. Bus Direction is East Bound ...................... 143

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 9. Main Street (E-W) Green is Extended for
5-Seconds / Cross Street (N-S) is Cut by 5-Seconds. Third Time

Period is Preempted. Bus Direction is East Bound ...................... 144

Time Period Movement Specific Statistics as a Result of Preemption

For Intersection Number 11. Cross Street (N-S) T & R Phase is

Cut By ll-Seconds, Main Street (E-W) Left Phase is Skipped,

And Main Street (E-W) R & T Phase is Started 35-Seconds

Earlier. 3rd Period is Preempted. Bus Direction is WB ............... 145

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 6. lO-Seconds of Early Start For Main Street

EB/ lO-Seconds Cut From Cross Street NB and Main Street WB

Left Turn. Third Time Period is Preempted. Bus Direction is East

Bound ............................................................................................... 146

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 3. Main Street (E-W) Green Has Started

3-Seconds Earlier] Cross Street (N-S) is Cut by 3-Seconds. Third

Time Period is Preempted. Bus Direction is East Bound ............. 147

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 10. Main Street (E-W) Green is Extended for
3-Seconds / Cross Street (N-S) is Cut by 3-Seconds. Third Time

Period is Preempted. Bus Direction is East Bound ...................... 148

Time Period Movement Specific Statistics as a Result of Pre-

emption For Intersection Number 11. Main Street (E-W) R & T

Phase rs Extended For 25-Seconds, While Cross Street (N-S) Left

Phase is Skipped. Part of the 2nd and Part of the 3rd Time Periods

are Preempted. Bus Direction is WB ............................................. 149

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 6. lO-Seconds of advanced Green For Main

Street EB / lO—Seconds Cut From Cross Street NB and Main

Street WB left. Period Extended is The Third Time Period. Bus
Direction is East Bound .................................................................. 150

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 2. Main Street (E—W) Right & Thru Phase is

xi

Table D8

Table D9

Table D.lO

Table B]

Table E.2

Table E.3

Table E.4

Table E.5

Extended For 16 Seconds / Cross Street (N-S) is Skipped. Third
Time Period is Preempted. Bus Direction is East Bound ............. 151

Time Period Movement Specific Statistics as a Result of Pre-

emption For Intersection Number 11. Main Street (E-W) R & T

Phase Started Earlier, While Cross Street (N -S) T & R and Main

Street Left Phases Were Skipped. 3rd Time Periods is Preempted.

Bus Direction is WB ....................................................................... 152

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 9. Main Street (E-W) Green Has Started

S-Seconds Earlier/ Cross Street (N-S) is Cut by 5-Seconds.

Third Time Period is Preempted. Bus Direction is East Bound 153

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 9. Main Street (E-W) Green is Extended for

19 Seconds / Cross Street (N-S) Phase is Skipped. Third Time

Period is Preempted. Bus Direction is West Bound ..................... 154

Time Period Movement Specific Statistics as a Result of Pre-

emption For Intersection Number 11. Cross Street (N-S) T & R

Phase is Cut By ll-Seconds, Main Street (E-W) Left Phase is

Skipped, And Main Street (E-W) R & T Phase Started 35—Seconds
Earlier. Preemption and Compensation are in the 3rd and 4th Periods,
respectively. Bus Direction is West Bound ...................................... 155

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 6. lO—Seconds of Early Start For Main Street

EB/ lO—Seconds Cut From Cross Street NB and Main Street WB

Left Turn. Third Time Period is Preempted. Bus Direction is East

Bound ............................................................................................... 156

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 3. Main Street (E-W) Green Has Started

3-Seconds Earlier/ Cross Street (N-S) is Cut by 3-Seconds. Third

Time Period is Preempted. Bus Direction is East Bound ............. 157

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 5. Main Street (E-W) Green is Extended /

Cross Street (N-S) is Skipped. Third Time Period is Preempted.

No Compensation. Bus Direction is West Bound ........................ 158

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 10. Main Street (E-W) Green is Extended for
3-Seconds / Cross Street (N-S) is Cut by 3-Seconds. Third Time

Period is Preempted. Bus Direction is East Bound ...................... 159

xii

Table E.6

Table E.7

Table E.8

Table E.9

Table E.10

Table F.1

Table F.2

Table F.3

Table F.4

Table F.5

Table F.6

Time Period Movement Specific Statistics as a Result of Pre-

emption For Intersection Number 11. Cross Street (N-S) T & R &

Main Street (E-W) Left Phases are Skipped, And Main Street

(E-W) R & T Phase Started Earlier. Preemption and Compensation

are in the 3rd and 4th Periods, respectively. Bus Direction is

West Bound ..................................................................................... 160

Time Period Specific Statistics as a Result of Preemption For

Intersection Number 6. lO-Seconds of Advance Green For Main

Street EB / lO-Seconds Cut From Cross Street NB and Main

Street WB Left Turn. Period Extended is The Third Time Period.

Bus Direction is East Bound ................................................................. 161

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 3. Main Street (E-W) Green Has Started

3-Seconds Earlier/ Cross Street (N-S) is Cut by 3—Seconds.

Third Time Period is Preempted. Bus Direction is East Bound 162

Time Period Specific Statistics as a Result of Preemption For
Intersection Number 9. Main Street (E-W) Green Has Started

5-Seconds Earlier/ Cross Street (N-S) is Cut by 5-Seconds. Third

Time Period is Preempted. Bus Direction is East Bound ............. 163

Time Period Movement Specific Statistics as a Result of Pre-

emption For Intersection Number 11. Cross Street (N-S) T & R

Phase is Cut By ll-Seconds, Main Street (E-W) Left Phase is

Skipped, And Main Street (E-W) R & T Phase Started 35-Seconds
Earlier. Preemption and Compensation are in the 3rd and 4th Periods,
respectively. Bus Direction is West Bound ....................................... 164

Cumulative Network Statistics; With and Without Preemption 165

Cumulative Network-Wide Bus Statistics; With and Without
Preemption. Route 1 ........................................................................ 166

Cumulative Network-Wide Bus Statistics. With and Without
Preemption. Route 2 ........................................................................ 167

Cumulative NETSIM Person Measures of Effectiveness For;
Before and After Preemption .......................................................... 168

Cumulative NETSIM Bus Statistics; With and Without Preemption. 169

Cumulative Network Statistics; Four Periods Before and Four

Periods After For Each of the Four Selected Preemptions.
(No Compensation) ........................................................................ 170

xiii

Table 0.1

Table G.2

Table G.3

Table 6.4

Cumulative Network Statistics; With and Without Preemption 171
Cumulative NETSIM Person MOEs; Before and After Preemption. 173
Cumulative NETSIM Bus Statistics; With and Without Preemption 174

Cumulative Network-Wide Bus Statistics; “With and Without
Preemption ....................................................................................... 174

xiv

Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10
Figure 11
Figure 12
Figure 13
Figure 14

Figure 15

Figure 16
Figure 17
Figure 18
Figure 19

Figure 20

LIST OF FIGURES

Automatic Signal /Eagle Signal BPS work plan .....................
Map of Washtenaw Avenue - East, Ann Arbor .....................
Example of The Wayne State’s Field Data Collection ..........
Simulation Network, Nodes and Links ....................................
Far-Side Bus-Stop Intersection / Detector Configuration .....
Near-Side Bus-Stop Intersection / Detector Configuration...
Phasing Configurations Along Washtenaw Avenue ..............
Network Average Delay Due to Volume Change ...................
Total Links Bus Delay Due to Volume Change ......................
Network Average Vehicle Delay .............................................
Person Average Link-Trip Delay .............................................
Average Total Bus-Link-Trip ..................................................
Bus Route 1 Mean Travel Time ...............................................
Bus Route 2 Mean Travel Time ...............................................

Average Delay For Advance Green / Green Extension
Preemption ................................................................................

Average Delay For Skip Phase and No Compensation ..........

Average Delay For Skip Phase Preemption With
Compensation ...........................................................................

Cumulative Network Delay ......................................................
Average Person Statistics .......................................................

Bus Route 1 Mean Travel Time ...............................................

XV

38

45

48

49

56

62

92

96

100

101

102

103

104

107

108

109

112
112

113

Figure 21
Figure 22
Figure 8-1

Figure 8-2

Bus Route 2 Mean Travel Time ...............................................
Total Link Bus Delay ................................................................
Far-Side Bus-Stop BPS Algorithm .........................................

Near-Side Bus Stop BPS Algorithm ........................................

xvi

Chapter 1

Introduction

Due to the lack of resources, there is a growing interest in the maintenance and manage-
ment of the existing transportation system. One consequence of this has been the emer-
gence of transportation system management ('1‘ SM) as a planning philosophy. TSM is a
process for planning and operating for which key objective is the conservation of fiscal
resources, energy, environmental quality, and the urban quality of life. TSM has been
defined to include a large number of project types; however, one type of particular interest
is the bus priority system. This is a system of traffic controls in which buses are given spe-
cial treatment over the general vehicular traffic (for example, bus priority lanes or preemp-
tion of traffic signals). Of particular interest in this research is the bus priority
(preemption) signal; BPS. It is a method of providing preferential treatment to buses and
other high occupancy vehicles (HOV) by altering the signal timing plan to favor those

vehicles.

The concept of bus priority treatment is not a newly introduced strategy. In fact, an early
experiment was conducted in Washington, DC. in 1962. In that study, the offsets of a sig-
nalized network were adjusted to better match the lower average speed of buses (Sunkari
et al, 1995). One or more of the following factors (acting singly or in combination) have
prevented the widespread use of bus preemption in the United States (US): (a) the absence
of a reliable technology to track the bus arrival and to initiate preemption; lack of an auto-

matic vehicle location and classification system, (b) lack of standards to determine

2

warrants for preemption, (c) the failure of these systems to strike a balance between ade-
quately providing for the needs of general traffic while concurrently providing sufficient
benefits to transit to make such systems cost effective (Jacobson 1993), and (d) lack of

sufficient commitment to the HOV philosophy on the operational level.

In general, providing preferential treatment for buses is expected to improve the perfor-
mance of buses and possibly of the other traffic on the bus direction. However, delay is
expected to increase for traffic on the cross street. In an attempt to reduce reliance on auto-
mobile travel, efforts have been made to make public transit more attractive by reducing
transit delays, providing more reliable transit schedules, and providing a level of service
that might make it competitive with private automobiles. When bus delay is reduced,
buses run on a more reliable schedule and their trip time is shorter. This makes transit a
more attractive mode of transportation and may increase bus ridership by diverting private
automobile drivers. This, in turn, will result in congestion relief and a reduction in exhaust

emissions.

The Intermodal Surface Transportation Efficiency Act of 1991 (ISTEA) provides the
framework for federal funding for transportation facilities over the next six years. ISTEA
stipulates that the US transportation network will provide “the foundation for the Nation
to compete in the global economy, and will move people and goods in an energy-efficient
manner.” One of the stated purposes of this bill is to reduce the number of single
occupancy vehicles, particularly in cities designated as non-attainment areas. The seven
county Detroit Metropolitan area, which includes Ann Arbor, was one of these designated

areas when this study was initiated.

3
One method of meeting this objective is to improve the quality of service to public transit
and high occupancy vehicles to make them more competitive with the automobile. The
Ann Arbor Transit Authority (AATA) has recognized this need, and initiated a project to
improve transit service by incorporating the technologies being developed under the Intel-
ligent Transportation System (ITS) program into their bus system operation. They
received a grant to develop and implement “smart card” technology in their bus fleet oper-

ation.

The smart card, which is commonly understood to be an integrated circuit-based, credit
card-sized portable data carrier, is fast becoming a preferred medium for ITS applications.
While there are many applications for the smart cards, the one of particular interest to this
project is the use of these cards to transmit a signal which can be used by a traffic signal
controller to identify the location of the HOV and to change the signal timing at selected

intersections to provide priority treatment.

1.1 Research Objectives

The objectives of this research are:

1. To determine the benefits of providing buses with an electronic signal preemption
device and to predict the changes in traffic performance and bus services caused by

implementing various bus priority schemes. The Washtenaw Avenue Corridor in Ann

Arbor, Michigan is used as the study location.

2. To determine the traffic conditions (e. g. volume, signal timing, percentage of HOV)

under which signal preemption will improve flow.

4
NETwork SIMulation (NETSIM) will be used to simulate different algorithms for imple-

menting signal preemption and to assess various bus preemption policies.

1.2 Bus Priority (Preemption)

One of the preferential treatments for buses is providing priority at traffic signals. Bus
characteristics are different than the general vehicular traffic. Unlike automobiles, buses
can not continue platooning through signalized corridors due to the random occurrence of
passenger loading and unloading volume and the resultant variable dwelling time. A bus
may skip a stop if there is no passenger waiting to get on or off. On the other hand, a large
number of passengers boarding and unboarding requires more time. These variations
make the bus arrival time at a signalized intersection uncertain. In addition, slow bus
acceleration and deceleration and the typical slower bus movement makes the bus unable
to stay in the traffic stream. In this case, the bus may not enjoy the full benefits of a coor-

dinated signal system.

There are two main techniques to provide priority treatment for HOVs at traffic signals.
These are passive and active detection and granting priority. Passive priority systems are
characterized by the fact that the flow of buses need not be recorded at a particular instant
in order to grant priority. Instead, the intensity of bus (or HOV) movements is deduced
from historical measurements of traffic flow. An active priority system is when the pas-
sage of an individual bus is detected and priority is awarded to the bus as a result of this

detection.

1.2.1. Passive Priority
This system is based on signal coordination and improved signal timing for all arterial
traffic to favor bus traffic. The following are methods for improving transit operation

USunkari et a1 1995, Allsop 1977, Nato 1976);]

Adjustment of cycle time. If signal cycle times are generally long, buses may have
to wait longer on a red signal. Reducing cycle lengths at intersections carrying appre-
ciable bus traffic can provide benefits to transit vehicles by reducing delayElowever,
short cycle times result in a decrease in capacity and can become insufficient to pass

all the traffic arriving at an intersectiori)

Splitting phases. Splitting a priority phase movement into multiple phases within a
cycle can reduce transit delays without necessarily reducing cycle length. By repeating
the priority phase within the same cycle, transit vehicle delay may be reduced at the

rntersectronfiiowever, there rs a delay penalty imposed each trme a phase rs 111111211qu

Area-wide timing plans. These plans provide priority treatment for buses through
preferential progression by designing the signal offsets in a coordinated signal system
using bus travel times.!This optimization method would have the objective of mini-
mizing passenger delay rather than vehicle delay. In addition, it would take into con-

sideration stopping at bus storfl

Gating (Metering Vehicles). The idea behind this method is to limit the number of
vehicles gaining access to a particular facility. Metering regulates the flow of vehicles
through a network by limiting the number of vehicles allowed into the system. Buses

benefit from this by allowing them to bypass metered signals with special reserved bus

6

lanes, special signal phases, or by rerouting buses to non-metered signals.

"1"

{luming prohibition. Where left turning vehicles at junctions cause congestion, it is
not uncommon to prohibit such turning movements even though the vehicles affected
may incur significant extra trip time.§xempting buses from such bans not only saves
them from delays due to diversion, but keeps them on those routes which are best for

passengers. !
___,J

1.2.2. Active Priority

Active priority is, sometimes, referred to as priority by detection or bus-actuated signals or
bus preemption signals. The benefit of active priority over the passive priority is that the
treatment is provided only when the bus is present. Efollowings are methods of active

priority treatment (Sunkari et a1 1995, Allsop 1977, Nato 1976)}

Green extension. This means extending the green phase beyond its normal setting to
allow the bus to pass the intersectioniitis usually limited to some maximum value. ‘
Phase extension is provided when the bus will arrive at the intersection just after the

end of the normal green period.

Phase recall (early start or red truncation). This priority treatment advances the
bus street green phase by prematurely terminating all other non-bus phases (and trun-
cating the bus red phase). This treatment is used when the bus arrives at the intersec-
tion during the red signal phase. This may be constrained by providing a minimum

green time for the phase to be prematurely terminated.

Phase skipping. To facilitate the provision of the bus priority phase, one or more

7

non-priority phases may be omitted from the normal phase sequence. In order to avoid
disrupting operations on the non-bus phases, some restrictions may be applied to this

treatment, such as no phases with heavy demand are skipped.

Compensation. One may choose to compensate for the time lost (skipped or cut)
from the other non-bus phases in the next cycle to limit the adverse effects priority has
caused to the non-priority traffic. Compensation for the non-priority phases involves
allocating extra green time to these phase to make up for time lost during signal pre-

emption.

Conditional versus unconditional priority. Unconditional priority is the provision
of signal priority each time it is requested (the bus detector or signal transmitter places
a call to the signal controller), after all other vehicular and pedestrian safety required
intervals are satisfied. Some professionals argue that since (unconditional) preemption
is disruptive to the cross-street traffic, it would be better to subject preemption to cer-
tain conditions. These selective conditions determine when or if the signal priority will
be granted to the bus. There are several factors that can be used; such as: is the bus on
schedule or behind schedule, bus occupancy, cross-street traffic conditions, and time

between consecutive preemptions (other conditions may also be used).

The above treatments are the most widely used forms of active priority. In this research,
the active priority system in detecting the bus and granting priority under certain criteria is
adopted. Several combinations of these various treatment schemes are tested. The
NETSIM computer model has been selected to simulate this process (details will be dis-

cussed later).

8
The following chapters present a review of the literature and past experience, data collec-
tion and requirements, research methodology, bus preemption signal algorithms, evalua-
tion of different BPS plans, evaluation of BPS under different traffic conditions, and

conclusions.

Chapter 2

Literature Review

One of the earliest known bus preemption experiment was performed in 1967 by Wilbur
Smith & Associates and the Bureau of Traffic Research in the Los Angeles Department of
Transportation (Benevelli et a1, 1983). Two intersections in Los Angeles were studied:
Broadway and First and Broadway and Second. In discussing this experiment, the authors
indicated that traffic signal delay constituted 10-20 percent of the average bus trip time
and that signal delay would be the easiest component of delay to reduce. The bus preemp-
tion was accomplished by having a person manually actuate the signal, so as to begin the
green interval earlier if a bus approached on the red interval, or to extend the green inter-
val if necessary to allow the bus to pass through the intersection. Bus portal-to-portal trav-

el time was reduced by 5 to 7 percent.

Several simulation models and field experiments with signal preemption have been con-
ducted in the U. S. and Europe since this early experiment. Most of these involved isolated
intersections, and only limited information is available on network level experiments.
Most of these projects were conducted in the 19703, a few of them were in the early 19805,

with very few recent studies in the late 19805 or 19905.

2.1 Isolated Intersections

During the late 19705, many papers were written on bus preemption with various strate—

gies. Vincent et a1, 1978, used a microscopic Bus Priority Assessment Simulation

9

lO

(BUSPAS) program to test five preemption control strategies. They examined (a) green
extension only; (b) green extension, red truncation, no compensation; (c) green extension,
red truncation, compensation; (d) red truncation, no compensation; and (e) red truncation,
compensation. Their experiments considered several traffic volumes, saturation flow rates,
and signal timings. Several bus detector spacings and placements were also considered.
For the three main priority control methods (a), (b), and (c), it was found that method (a)
gave limited benefits to buses (0-8 seconds), with little disbenefit to other traffic (less than
1 vehicle-hour/hour (veh-h/h)). Method (b) gave larger benefits to buses (4-24 seconds)
but also larger losses to non-bus traffic (1-24 veh-h/h). Method (c) produced smaller bene-
fits for buses (0-14 seconds) than (b), but also less disbenefit to other traffic (1-14 veh-h/

h). The above approach is somewhat similar to what was done in this study for an arterial.

Richardson et a1, 1979, developed and applied a new methodology for the evaluation of an
active-bus priority signal system which was installed at traffic signals in Victoria, Austra-
lia. Two new measures, perceived delay and budgeted delay, were introduced in their
study and were shown to have important implications in the evaluation of bus priority and
other transportation system management schemes. Perceived delay is a measure of the
psychological effect of time delay (i.e., the value of time savings is a function of the
amount of time saved). In this study, budgeted delay was defined as being equal to the sum
of the mean and the standard deviation of travel time (or delay). It corresponds to an upper
percentile point (for a normal distribution it would represent the 84th percentile point) of
the delay distribution. They found the consideration of changes in “budgeted delay” rather
than the mean delay results in a greater probability of justifying bus priority schemes.

They stated that it is possible to have better service even when mean delay increases,

11

provided that the reduction in variability of delay is of sufficient magnitude. Richardson et
a1 concluded that reevaluation of TSM schemes on the basis of perceived and budgeted
time savings would probably result in many of them being feasible.This concept has not
been observed in other research and, in this research, the classical delay measures will be

used.

Jacobson and Sheffi 1980, developed an analytical model of delay at isolated signalized
intersections with a bus preemption scheme. The analysis was presented for the simplest
case, i.e., two— intersecting one-way streets. The model treated the beginning time of the

’ green period as a random variable, the density function of which was developed. The
model also assumed a Poisson arrival process for the vehicles approaching the intersec-
tion. Four cases were analyzed: A) no preemption, minimizing total person delay; B) no
preemption, minimizing total vehicle delay; C) preemption, minimizing total person de-
lay; and D) preemption, minimizing total vehicle delay. The intersection performance in-
dicators were total person delay measured in seconds per hour, queue length, and delay to
both private vehicles and bus patrons. They showed that the preemption benefits can be
substantially increased by changing the underlying signal setting once preemption is
installed. It was found that the inclusion of phase durations in the design variables (case
D) significantly increased the benefits associated with preemption (17.3% with respect to

case A).

As a general conclusion (no numerical value was furnished), the benefits associated with
bus preemption were relatively small when the traffic flow in the preemption direc-

tion was much higher than the cross traffic flow and thus this direction already

12

experienced green for most of the cycle. Preemption was more beneficial where the

rate of arrival of buses was higher.

Twenty-seven priority treatment projects for HOVs were evaluated by Rothenberg and
Smdahl, 1981. Out of those, three included signal preemption treatments for buses. Two
were active preemption systems using on-bus emitters, and the other project utilized a
pavement loop to detect bus presence. The active preemption treatment produced bus trav-
el time savings in the range of 4 to 8 minutes, a 10-20 percent reduction. Bus reliability
was also improved. The passive preemptions produced comparable travel time saving
rates. In both cases, impacts on cross street traffic was not significant in most situations.
None of the preemption systems was reported to exhibit much direct impact on bus rider-

ship.

A macroscopic traffic delay model which applied a stochastic procedure was presented by
Radwan and Hurley, in 1982 to evaluate different bus preemption signal strategies at iso-
lated intersections. The model permitted the user to evaluate various operational strategies
provided for bus traffic. The model proved cross street passenger delay savings to be sen-

sitive to saturation headways between 1800 and 1980 vehicle / hour.

Roark, 1982 determined the effectiveness of bus signal preemption to be a function of the
cross-street traffic with the greatest potential on arterial roadways with little cross-street
traffic. He reported two problems associated with bus preemption signals: (1) platoons of
automobiles travelling around a bus to take advantage of priority operation and (2) bus
drivers who anticipate a green signal and approach the intersection at a high rate of speed.

Roark reported on several field studies that bus preemption reduced bus travel times and

l3

resulted in smoother traffic flow on arterial streets. He recommended four criteria for bus
preemption: (a) when total person delay (a function of cross-street volumes) is reduced,
(b) at least 10 to 15 buses are carried on the arterial during the peak hour, (c) a daily vol-
ume of at least 100 buses in both directions, and (d) the cross-street green phase can be re-

duced without conflicting with the minimum pedestrian clearance time.

A bus signal preemption algorithm was built by Smith 1985, for the New Jersey Depart-
ment of Transportation to be incorporated into the NETSIM simulation model. While five
bus preemption strategies were selected initially, due to budget limits, the evaluation was
reduced to just an algorithm for advancing or extending green while still maintaining a
minimum side street green. Smith reported that the algorithm was programmed into the
NETSIM model by the Federal Highway Administration (FHWA) and was tested by com-
paring the results obtained from N ETSIM simulation and the results obtained from a man-
ual implementation of bus signal preemption at one intersection. The algorithm was
considered to be a reliable estimator of the effects of using bus signal preemption at an in-
tersection. A nearside bus stop and a farside bus stop condition were considered in the al-
gorithm. The t-test showed no significant difference between the data sets of (measures of
effectiveness (MOEs) (average delay per vehicle and percent of vehicles stopping) mea-
sured and estimated at the 95% level. The preemption process resulted in savings of 6.2

vehicle hours and 9.2 passenger hours over the one hour peak period.

To study possible means of improving the movement of transit vehicles in Metropolitan
Toronto, the Transit Priority Study was established by the Metropolitan Toronto Roads

and Traffic Department as a three phase program. Bishop et a1 1988, addressed phase III of

14

the study which was to permit a transit based preemption system test on one or more inter-
sections on the selected routes. Several strategies were discussed: 1) green extension, 2)
red truncation, 3) window stretching, 4) red interruption, and 5) green truncation. The first
two strategies were selected for testing. The evaluation criteria were capacity improve-
ment, implementation capability, progression problems, safety problems and reduction of
transit delay. Two isolated intersections were tested for preemption, Queen (streetcar) at
Sherbourne and Sheppard Avenue West (bus) at Jane. For streetcar operations, it was de-
termined that the transit priority strategies of green extension or red truncation produced a
reduction in signal delay to the transit vehicles. The cumulative reduction in signal delay
per streetcar travelling in both directions was in the range of 8.7 to 10.7 sec/veh. The im—

pact to cross street delay was an increase of 0.3 to 10.6 sec/veh.

Davis et a1,1991, indicated that despite the fact that in ideal conditions a transit prior-
ity scheme would produce no reduction in network capacity, in reality, some loss of
capacity is likely to occur as a result of transit priority schemes. UK. Department of
Transport guidelines state that for a good scheme, capacity loss should be no more than
one or two percent. Total vehicle journey times might then be expected to increase by
three to ten percent. Poor priority schemes which produce much greater disruption need to

be modified or withdrawn.

There are a number of factors that have prevented the widespread application of bus pre-
emption in the United States according to Khasnabis et a1, 1991. These include the ab-
sence of a reliable technology to monitor the arrival of buses and to trigger preemption,

lack of standards to determine warrants, and inability of the system to prevent inordinate

15

delays to motorists travelling on the cross streets. Where a bus stop is located immediately
prior to the intersection the predictions of exact arrival times can be particularly difficult.

No effort was made to assess the adverse effects of preemption on cross street traffic.

Casey et al, 1991 indicated that currently signal preemption for HOVs is relatively un-
common in the U. S. He reported that a few cities do have preemption equipment for light
rail lines, including the Southeastern Pennsylvania Transportation Authority (SEPTA) in
Philadelphia, the Santa Clara County Transit District in San Jose, California, and the
Southern California Rapid Transit District (SCRTD) in the Los Angeles area. SCRTD also
had equipment installed for signal preemption on two bus routes. The system was taken
off-line fairly soon after implementation due to highway construction, but was to be reac-
tivated as soon as the construction was completed. They reported that two other agencies,
the Chicago Transit Authority and Broward County Division of Mass Transit in Fort Lau-
derdale, Florida were also discussing signal preemption as part of Automatic Vehicle Lo-

cation (AVL) systems.

Ingalls et a1, 1993 studied different alternatives for providing priority to HOV in the sub-
urban arterial environment. Different evaluation criteria of financial viability, geometric
feasibility, functional adequacy, and public acceptance for these alternatives were ana-
lyzed. Alternatives included signal priority treatments, continuous right-side HOV lanes,
continuous left—side HOV lanes, lane control for reversible HOV lanes, signal queue jump,
single occupancy vehicle (SOV) turn restriction, off-route alternatives, and special access
for HOV. Of these various alternatives, signal priority treatments which used advanced

technologies to minimize person delay at intersections showed the greatest potential to

l6

achieve the goal of bypassing congestion without unacceptable impacts to general purpose

traffic. However, no numeric values were provided.

Alice et a1, 1993 altered the Traffic Network Study Tool (TRANSYT-7F) model to repre-
sent the case of near-side transit stops in shared lanes. When used for optimization purpos-
es the transit-enhanced TRANSYT model tends to coordinate the intersections in such a
way as to make the transit load/unload operations occur mainly during the red phase. De-
spite some limitations, it was seen that delays and stops can be reduced considerably when

signal timings reflect the transit loading operation.

Chang et a1, 1995 formulated a model for an integrated adaptive control system with both
bus preemption and signal control functions. In the proposed model, absolute priority was
not given and minimum cross street green time was imposed. The model made use of real-
time algorithms instead of pre-specified strategies used by most conventional bus-preemp-
tion logic. The control decision for signal settings was based on a performance index
which incorporated bus delay, as well as passenger and vehicle delay. TRAF-NETSIM’s
outputs for an isolated intersection under different traffic conditions were used to test the
performance of the algorithm. They claimed that experimental results proved the superior-
ity of the model over the actuated control logic by NETSIM. However, since only a simple
myopic adaptive logic was employed in the model, they suggested that more enhance-
ments that employ information from both neural network prediction models and AVL

were needed.

Sunkari et a1, 1995 developed a simple analytical model to evaluate priority strategies,

which uses the delay equation found in the 1985 Highway Capacity Manual (HCM). They

17

have tested no priority, phase extension, and early start schemes. Stopped delay was used
as the field measure to validate the model. It was found that the model is reasonably accu-
rate in estimating the effects of bus priority at an intersection. However, it overestimated

delay for some phases.

2.2 Network and Arterials

In an early simulation study, Ludwick 1976, reported that an unconditional preemption al-
gorithm using the Urban Traffic Control System / Bus Preemption Signal (UTCS/BPS)
model on a network of quarter—mile route segments was used. The study provided a 25
percent travel time benefit to buses. However, the cross-street traffic delay could be ex-
treme, particularly at short bus headways. An algorithm limiting the preemption to a max-
imum of 10 seconds still provided a 20 percent bus travel time improvement with only a 7
percent cross-street travel time increase. It was found that far-side bus stops were far supe-
rior to near-side bus stops. Buses with frequent stops have greater potential for improve-

ment than express buses, especially if existing signal coordination is good.

In a demonstration project of signal preemption for express buses in Sacramento County
(Elias 1976), a bus preemption system was evaluated on a 3.8—mile section which included
nine signalized intersections operated as isolated, full traffic actuated signals equipped
with traffic signal preemptors. Two buses were equipped with transmitting units. Elias re-
ported a reduction in bus trip time of an average of 23 percent. Passengers benefitted by a
smoother and more comfortable ride with increased schedule reliability. There were no ac-
cidents caused by the bus preemption identified during a 3-month testing period. No ad-

verse effects were observed for cross street traffic (However, no data were presented on

l8

cross street delay.) Several benefits were reported: operating cost, trip time and depend-
ability were improved; fuel was saved due to elimination of starting, stopping and waiting;

and air pollution was reduced, as was the noise and wear-and-tear on tires and brakes.

Another bus preemption demonstration field experiment was conducted in Miami on the
Northwest Seventh Avenue corridor that has a 10 mile length (Wattleworth et al, 1976).
Five combinations of three bus priority treatments were evaluated: (a) a reversible, exclu-
sive bus lane; (b) a traffic signal preemption system; and (c) a coordinated signal system
designed to favor movement of express buses in the peak-period direction. They evaluated
the bus priority treatments by their effects on bus operations, traffic signal performance,
traffic stream, and transit operation. The provision of a preemption capability reduced the
average bus travel time by 22.5 percent from a before condition of 28.0 minutes. Buses
were able to clear the preempted intersection within the maximum allowable preemption
time of 1203. Slightly longer phase lengths were observed during cycles in which buses
arrived. The bus priority treatment increased the number of persons moved on Northwest
Seventh Avenue by 20 to 30 percent although buses constituted less than 2 percent of the

traffic stream.

Liberman et al, 1978 reported on a simulation study that used the Simulation of COrridor
Traffic (SCOT) model. This study evaluated a network in the Central Business District
(CBD) of Minneapolis under a fixed-time signal timing plan generated by SIGOP—II to
minimize person delay using a bus preemption control strategy. On each of two adjoining
parallel, one-way arterials, a contraflow bus lane has been implemented.The bus preemp-

tion control strategy could call for green extension, red truncation, the signal to cycle to

l9

reinstate the normal green phase, or the signal to cycle to reinstall the green phase after
satisfying other phase duration minimums. They indicated that the buses along the major
arterials benefited significantly, while those along the cross streets experience sharp degra-
dation in performance. The overall bus performance experienced improved service as
measured by a 12 percent reduction in the total delay relative to the base system. In the
peak hour a net reduction in delay of 26.3 passenger-hours per hour could be achieved. No

base value (or before value) was provided.

Salter and Shahi, 1979 developed a microscopic model to predict the travel times of buses
and other vehicles along a highway network that has different types of intersection con-
trols, with or without bus priority schemes in operation. Their model has the capability of
evaluating the effect of bus priority measures at priority and roundabout intersections.
Salter and Shahi tested the following highway and traffic situations: (a) a priority intersec-
tion where the nearside lane of the minor road is allocated to buses for different traffic
flow conditions and different lengths of priority lanes; (b) signalized intersections that
have two or three approach lanes where the nearside lane of one approach is allocated to
buses for different traffic flow conditions and different lengths of priority lane, and (c) a 2-
km length of bus route, which included three signalized intersections and eight bus stops
for differing traffic volumes and proportions of buses in the traffic flow. They reported that
the observed and simulated data were quite Close to each other and that the model was ad-
equate to represent vehicle behavior according to the purpose of their study. When bus pri-
ority schemes were introduced, travel time for nonbus vehicles was increased proportional
to the traffic volume. Salter and Shahi’s model is a general model to predict traffic charac-

teristics, but does not deal with specific strategies of bus signal preemption, like green

20

extension and red truncation.

Hubschneider 1982, presented a simulation study of an active priority system based on a
bus guidance and control system (BGCS). It is a computer supported system used in the
surveillance and control of a public transport system. All vehicles are supervised by a cen-
tral computer by means of wireless digital communication. A minimum green restriction
necessary for clearance and safety was used before the bus green period can begin. A mi-
croscopic simulation package, MISSION, was used to investigate the impact of different
systems of modules in a small network. He demonstrated that buses with higher needs for
priority (running behind schedule) can be treated preferentially, while the restrictions on

the non-priority traffic can be reduced by refusing priority to buses which are too early.

Urban Traffic Control System / Bus Priority Signal (UTCS/BPS) is a microscopic traffic
simulation model that was developed by the Federal Highway Administration (FHWA)
and was used to simulate the bus preemption system operation for various bus flow rates
and bus stop locations. Benevelli et al, 1983 conducted a study on bus signal preemption
using the UTCS/BPS model. They concluded, based on a benefit-cost analysis, that bus
preemption was justified for the 1.3-mile segment of Monument Avenue in Richmond,
Virginia. The benefits of bus preemption were found to be limited by the preemption algo-
rithm structure and the bus stop locations. It was found that multi-phase signals minimize
the benefits of preemption under the control algorithm, and as more signals on the arterial
were preempted, the benefits of coordinated signals disappeared and the vehicle delay in-
creased. A farside bus stop was found to minimize the negative effects of bus preemption

on automobile travel delay. Benevelli et al utilized SOAP and TRANSYT models to

2]

determine the phasing pattern and cycle length. It was found that the inability of the al-
gorithm to reestablish offsets once a signal preemption occurred may also have ad-
versely affected road user costs. The control algorithm also did not have the

capability to skip phases.

It has been argued by Casey et al 1991, that signal preemption disrupts traffic flow. Many
traffic professionals argue that signal coordination and progression are more effective
tools on heavily travelled arterials than preemption. It is difficult to give preference to bus-
es in the mixed flow traffic, especially under congested conditions. Casey and others could
identify at least four field tests of signal preemption in the U. S. during the 19705: Kent,
Ohio; Louisville, Kentucky; Miami, Florida; and Washington, D. C. In Kent, Ohio, equip-
ment was installed in three signals along a four-mile section of East Main Street. In this
study, the buses experienced higher average speeds and shorter delays at intersections.
The project eventually terminated for administrative reasons. Louisville, Kentucky imple-
mented 3M equipment on express routes, and bus travel time decreased significantly. In
Washington, D. C. the buses signalled their presence to the loop through an antenna
mounted in the undercarriage of the buses. Then, preemption would be granted as an ex-
tended green if there would be a net decrease in the overall passenger delay at the intersec-

tion. This proved largely ineffective. The Miami experience was discussed previously.

Davis et a1, 1991, reported the use of TRANSYT in the U. K. for bus priority in Glasgow.
This experiment involved the modification of signal timing plans in the city to optimize
the movement of people, rather than the more conventional passenger car units (PCUs).

For the purpose of calculating signal timing, the average occupancy of buses was

22

considered to be 28 passengers, with 1.4 occupants assumed for other traffic. This experi-
ment resulted in an increase in bus speeds of 9 percent, 8 percent and 7 percent during the
morning peak, off-peak and evening peak periods, respectively, with an overall reduction
of 16 percent in the time spent delayed by signals. Cars travelling along the bus route ex-
perienced a 5 percent reduction in journey time, while those travelling off bus routes faced
a 15 percent increase in journey time. Overall, however, journey times for cars on the net-

work did not change significantly.

They reported two other experiments implemented in the U.K. for bus priority at traffic
signals. They were BADGE and PUMMEL. BADGE provided only limited variation
from a fixed time plan to give priority to individual buses. Tests on the BADGE system
showed a reduction in bus delays of 15 percent, 10 percent and 13 percent during the
morning peak, off-peak, and evening peak, respectively. PUMMEL allowed greater varia-
tion from the fixed time plan, using TRANSYT to estimate resulting delay to nonbus traf-
fic. PUMMEL was found to be less effective than BADGE at reducing bus delays, with
savings of 11 percent, 2 percent, and 7 percent in the morning peak, off-peak and evening

peak respectively. Delays to other traffic were too small to measure.

2.3 Signal Technology

For the BPS to be operational, hardware that is capable of vehicle identification and loca-
tion is required. The lack of reliable technology is one of the reasons that has prevented
the widespread use of bus preemption in the US. However, there were several types of
technologies used in different experiments. The Opticum System was developed by 3M in

the US around 1976. It was used in the Sacramento County signal preemption project in

23

1976 (Elias 1976). The Opticum System was based on strobe light pulses at a specified
rate being received by detectors at the intersection. However, both the academic research
and the actual demonstrations found major shortcomings. About the same time, the Philips
Corporation of the Netherlands developed a product called Vetag. This was based on the
use of inductive (magnetically activated) detector loops in streets which are activated by
programmable transponders on moving vehicles (buses). In 1987-1988, Philips released a
new product called Vecom which was more sophisticated than Vetag. The on-board equip-
ment has the ability to receive, as well as send, messages, and has computerized control
with considerable storage capacity. Traffic engineers were reported as generally comfort-

able with the greater reliability provided by this equipment.

Davis et al 1991, discussed bus priority at signalized intersections as one of three Ad-
vanced Traffic Management Systems (ATMS) technology applications to transit rideshare
schemes. Transit vehicles can be identified by using automatic vehicle classification
(AVC) techniques, making use of inductive loops or piezoelectric axle sensors (Casey et a1
1991, and Davis et a1 1991). An alternative to AVC for signal preemption involves the use
of automatic vehicle identification (AVI) technology. Davis et a1 indicated that this tech-
nology enables vehicles to be uniquely identified through a communications link between
an onboard transponder and a roadside reader unit. Several alternative AVI approaches
have been developed including optical infrared and radio frequency systems. AVI can
therefore be used to detect transit vehicles for signal preemption. Davis et al reported that
by 1976, for example, research on AVI conducted at the U.K. TRRL had led to the devel-
opment of selective vehicle detection systems for bus and emergency vehicle priority at

signalized intersections. In Delft, Holland in 1971, buses between the Hague and Delft

24
were given local priority at signalized intersections using a simple form of inductive AVI
known as VIPS. This was reportedly successful in reducing travel times and delays. Traf-
fic signal preemption in the Netherlands was accompanied by the activation of an acoustic

signal to warn pedestrians and cyclists.

The Philips Vetag AVI system was implemented in Holland during the 19705 for automat-
ic tram control. The Hague commissioned an automatic interlocking system covering the
city’s tram network. In Hong Kong, the AVI technology has also been used to provide pri-
ority and identification functions for a light rail transit system. The equipment automati-
cally identifies each light rail vehicle (LRV) approaching the intersection and establishes
the intended direction of movement. This information enables the traffic signal controller

to provide the correct clearance and a signal to proceed.

Davis et a1 suggested that it may be possible to implement a scheme for traffic signal pre-
emption for other HOVs using AVI technology. AVI transponders would be distributed for
installation on vehicles registered to participate in a rideshare scheme. On-board comput-
ers and/or individual smart cards could be used to prevent signal preemption by registered
vehicles that were not carrying the required number of occupants. The AVI transponder
would become active only after the insertion of the required number of smart cards into a
reader unit. A vehicle smart card could be used as an AVI selective vehicle detector. The
onboard computers (OBCs) would contain a record of the vehicle’s schedule, including its
correct arrival time at each intersection and boarding point. If it is preferred, as the vehicle
approached a signalized intersection, the CBC would activate the signal in favor of the

transit vehicle only when a deviation from the required schedule is detected.

25

Smart cards are essentially miniaturized computers. Davis et al reported that smart card
technology has recently been applied to transit operations. Smart cards could provide
much of the data regarding scheduling of transit services which currently relies on histori-
cal trip data gathered by labor-intensive manual methods, leading to cost and time savings
and providing a more reliable base on which to plan transit services. License plate scan—
ners is another technology that could be used for selective detection. These are capable of
automatically reading the characters on vehicle license plates. The Dulles Toll Road in
Virginia in 1989 tested a 3M manufactured license plate reader. Since the character recog-
nition software was optimized for Virginia, the system was less successful in reading
plates from other states. Read accuracy for Virginia plates was around 65 percent, al-
though 3M reported accuracy improvements due to a system modification since the time
of these tests. However, it is unlikely that the technology will ever provide the perfor-
mance levels available from AVI. The French Elsydel company recently claimed accuracy
levels of 95 percent for its infrared license plate scanner.Yamamoto (1992) claimed that
the Japanese licence plate readers have 80 percent accuracy with one second image pro-

cessing and fuzzy logic and 70 percent accuracy during the night.

Davis et al identified infrared beacons or digital radio communication as other potential
future ATMS developments. This could be used to provide the unique vehicle identifica-
tion function required for priority signal control. These systems could have advantages
over AVI and license plate scanners in providing increased scope for the integration of
ATMS with ATIS (Advanced Traveler Information Systems). Another technology is video
image processing. This technology could be used to identify transit vehicles for signal ac-

tivation. An example of this is the DACimage system developed in France by Elsydel.

26

Classification by this technology is based on the features of the vehicle, such as its number
of axles and height. In the longer term, Davis et al reported that it may become possible
for a video image processing system to estimate or calculate the number of occupants in a

vehicle. Infrared heat-sensing technology is potentially applicable in this area.

2.4 Summary

From the literature, one can conclude that bus preemption signal projects at isolated inter-
sections were relatively successful in reducing delays for the main traffic stream. Delays
for cross-street traffic were not significant at low volumes, but became significant as the
intersection approaches capacity. The effectiveness varies with different preemption
schemes (strategies): green extension, red truncation, compensation and, no compensa-
tion. The benefits of preemption could be increased by changing the underlying sig-

nal setting once preemption is installed.

For arterials, the effectiveness of bus preemption was also found to be a function of cross-
street traffic. The greatest potential lies on arterials with little cross-traffic. Buses drive
smoothly on the main street, experience delay savings and more reliable schedule, but the
cross-street traffic may experience a delay that could outweigh the savings on the main
street. Signal preemption could disrupt traffic flow, and it was found by some investigators
that signal coordination and progression are more effective tools than preemption on

heavily travelled arterials.

The benefits of bus preemption were found to be dependent on the algorithm and the loca-

tion of the bus stop along the route. A nearside bus stop was found to be less desirable in

27

reducing negative effects. Results of conditional preemption have been more sound than
unconditional preemptions in terms of less disruption to non-bus traffic, providing safer
operation (by maintaining minimum green time for cross street vehicles and pedestrians),
and preventing the bus from receiving priority treatment when it is not needed (bus is on-

schedule).

Integration of the BPS process and vehicle identification facility with adaptive traffic con-
trol is a step toward the concept of interactive traffic control. Preferential treatment of bus
users is one of the promising strategies to accomplish this. The integration of preferential
treatment and adaptive signal systems is promising. However, research in this area is still

very scarce.

The technology is available to accommodate the bus preemption signal process efficiently.
License plate scanners as an AVI system seem accurate enough with an 80 and 95 percent
accuracy for the Japanese products and French Elsydel company, respectively. Smart
cards and onboard computers are even more advanced and accurate. Smart cards can be
used as a source of information for transit’s schedule time including its correct arrival time
at each intersection. It also can provide information about the transit vehicle’s occupancy.
HOV’s can use smart cards to emit signals to traffic controllers to provide them with pre-

emption.

Despite the fact the technology is available, bus preemption is not popular in the US. One
of the main reasons for its failure is the inability of the system to reestablish the original
settings as preemption is called. The literature lacks information on an up-to-date model

that could be used to evaluate the effects of bus preemption on the system rather than just

28

the corridor which has the preemption facilities. There is no comprehensive model that
has the capability of testing several bus preemption strategies (including skip-phase plan).
The UTCS/BPS model presumably had the potential for evaluating the effects of bus pre-
emption on the system. Although the user manual is available, the model is no longer in

use and current information on the model is not available.

Tables 1, 2, 3, and 4 summarize previous bus preemption simulation and field experi-

ments, at isolated intersections as well as for networks and arterials.

In this research, a network including Washtenaw Avenue and one intersection across on
each side will be evaluated under the bus preemption signal strategies. These strategies
will be comprehensive and selected in a way that minimizes disruption to progression and
coordination. TRANSYT-7F is used to optimize the network settings. The BPS operation
is simulated by the TRAP-NETSIM simulation model by using some features that are
rarely used. It will utilize the graphic animation as a way of detection and the different
(nineteen) time periods (time plans) to provide different BPS strategies. Since it is pro-
posed to test the smart card technology in an effort to encourage multiple occupancy vehi-
cle usage, it is proposed that carpooler may subscribe to this service and their vehicles
might be provided with smart cards, as a way of seeking priority. This plan will be tested

and its effects will be assessed.

To evaluate the different BPS plans, links, individual intersections, traffic directions, and

network-wide measures of effectiveness with and without preemption will be evaluated.

29

 

 

seam 23 a 5;,
95:3 .35.: of. -
663.2 29$ 30?. £5

623m gave Soc; .53 2:0; Lewcom

 

 

 

 

 

 

 

 

02¢.-th ecu. 25-862 - -mam No new EnocéoE? me Co wE>wm Hz was .m.O gmhmz LEEm mwfl
.20:
-686 :e. E 653 momsm 055:0on
Eco .mxaaemo: :26;me 8 03:25,. PM can 29on E 8
Eacw; 32905 tom .83 wEEm >206 Lowcommaa 82; 3.05 dd -8an enigma $9.
.8sz we; 28 REE“ £5 :2?» .cEEEe
Eocecon EOE mm; mam .Boc 052. one :55; was
$.08 05 52: SEE; :02: mm? 5526 FEB have 22:? o:
.eoESE :EEEBE 05 E 30: oEaE 2t :23, :88 52% coEom -85on van 8 8
263 Sat 25:25:. oz :55 33:32 083 mam Co Scenes 2E. :39 wEEEEE Boa—«5% 8383 owi
5.66:; 550 .mmm wEbsmi
E eoEoEoEE :83 Co b25305 832w a E 3:38 mien Egon— eoeowesm E 8
5: .8: E828 E: 36335 E wowcmco Co :ochEmcoU Egon eoEoBom ESE :30 :OWEEEM m3—
.oEaE $50 2 Ecoconmno £2 :5 m
5:: mgcocon 5:25 nouzeoa m 8562 - .0 red .m
SEES $50 2: 8 £2 comb: new .02 ..._..m .v
325 8 aces—on cows. 02% m 8502 - .U Hm rmd .m
.0595 85o 8 wéocenfln 2:: E3 .02 Hi rmd .N E B
momsn 2 fiancee BEE: 3mm _ 8522 - *.m.0 ._ m<m-mDm EocE> mm?
@639
3:08:80 338% 683 382 553V. a“;

 

 

Eustace... broth—

 

 

 

 

 

.meezuomcofi: e828— 3 855m sous—=85 ma:— m=e_>e.5 me FEES—i “a «3er

 

 

flcmaflmEQQ—COU "U GOmuflmGQQ—n—Avo 62 no.2 :OmuflUH—q-hh. BM— - .r—HM—

:fingxm :0000 . .md "0082*

 

 

.:0:00:
2 E0500§a=0 082-

.:0~::= 0.03

228:3 35:00 :0ch

0&2 5:0:

 

 

 

 

 

:50 mam 5:5 63:00 :00 :0330: 05 :03 .088 05 we been EHum—Z

03:90:: REESE- -095 0.: :0>oE 8.30: 3:08.898 05; Hz ::m dd #2”: .0. :0 wSEU moo—
.maoum :50
03:0:8 ms: :0 :03: dos:

855:0 9 :00:m::0 ¢0mo w:_:0o_ 5:0: 05 80:0. meE: .0: m:

00>» ab.;mz<~; -ma :0:3 :00::0: 0: :8 macaw :SN x209 .H>mz<m,_. .0 .0 00:< moo.
:88:

3:0EEOU £30m :0mD _0:o2 55.3 50>

 

 

E0830; bier—m

 

 

 

 

 

€035.80 .n 03.?

 

 

3|

:0.E:::0::_ :3. - Ad $5.325 30:53 -.m.>>

550005 :3. . Hum

00:025.; 000:0 - 9.9

00.28%,— 000:U . and "0032*

 

 

£03000 :0. 0: m6. :0m00:0E
>0.0: 00:5 30:0 0.; 203000,

 

 

 

 

 

 

 

We. 0: 5w. >0.0: :30: E :0: 9.0 ::0
0:00 00:0: 0 :00::0:0 00:00:20 :0: ....~. ..m.>>
000:3. :0: :0EE00E 0:03 0.03% :00 5.2038 :00:w :0 mam 0...; Jaw. rmd .0 :0 00:05 wwm.
..0:0:: :5 3000 0.00..0>0
:0.:0:::0:E 20:50 0: a. 0:05: - $.30: :.0c mam 0...: :0 :0:0E:m0 PM.
.:00:0..:0 003 .25ka - :00w 0 :0:0:.m:00 £3» .25ka ::0 .md 2%sz 5.8m mwo.
95:02: 000 :0 609:. 80:: 02 -
:00?
30:0 :0 609:. ::00c.:w.m:.- Hm .0 :0
oxeom 0: 00:. >0 :0>0m 0E: .0>0:,r- md $005053. .wo.
00.0w:< 004 E :0:0:::0U 0.00 PM. :0.:.::0
::0::.:00x0 50::00 0:... 0: 00m >0 :00::0: 0E: .0>0:: mam - *.m.O -05 .0092). .0 :0 ...0>0:0m. 0:0.
:050:
0:025:00 3.000% E0880; :95 .0:0.>. .8534. 20%
>:.:0.:n.

 

 

 

 

 

 

 

0:388:00:— :0:0.0m_ :0 00.:3m 20.....— mA—m 2.030.:— ..0 >..0EE=m "N 030,—.

 

 

32

 

 

.00:08:0::0: 0E0: wE

-:00::0 >_00:0>:0 0:03 0:00::0 :0:
-:0:000: 0: _0:08 0:: :0 >::_:08
0:: ::0 ._0_:0::0 0:: w:0_0 0:0::

 

 

 

 

 

 

 

 

.:0:0000: ::0 8 :0:_0>0 0:03 -:800:: >:08 0.0030 000::-_:_=8 h>mZ<MH
00:05:50: ::0:: ::0 :0800. 0:20-05: 0:_0:00Z .:0c_:0:.a PM: ::0 .:<Om
0:03 >85. 0::: 80:: 0:0003 003 m:m 0_0>_0:0 UR: :0 :000m .m.O .m:m\mU:-D _0 :0 :_0>0:0m mwm:
._0:::00 ::0 00:0.
.2023. 0:: :0: 0: DUOm 0::3 .:0::_:0=.q :03 0: 0:60:00 ::E0: H: 205%:
_0:08 0050.080 0 0: ZOEBE wEES: 000:: :0: 53:83:: .md ::0 QUOm :0:_0::00::: mm?
.08:_0>
050:: 0: .0:0_::0:0:: :0000:0:_ _0:08
.>w0:0::0 :0_::800:: 0500:.0 003 08:: _0>0:: 0::-:02 - w8:0_:
::3 _00: :0: 000: :3: .00_:0_:0: .0020 >_0>_:0_0: 0:03 0:0: 0:: 08000 ::0:m
080:0 0508 0860:: _0:08 0: H :0:0_:8_0 0:: ::0 :0>:00:0 0: :- - 0:08 :30 ::0 :0:_0m 03:
>20: .080:
:85: 00888 0: ::005 -:0w:000-0: mdm :0 :0_:0=:0m -
>: :0:0:0:0w :0_: 9:8: 8=8_::0 Qum— >: :0>0::8: PM: ::OOHm
0:: 0:0—38:0 0: :00: 0,03 HOUm 00:08:0::0: 0:: =0:0>O - m0 ::0 FOUm _0 :0 008:0:5 wni
.::080>0::8_ :0: 0.0::
-:0:0: :00:w 0>0: 0:90 80:50:: -
:0::0: 0:0 0:90-05: 020:0: -
:00: 003 88:: 0:3 9.:
-030 :0::800:: 0:056:80: >: 08:: _0>0:: 00: :0 :0_:0::0m - *md m:m\mU-r3 0:23:04 :2:
:83:
0808800 0:300: ::08:00::- :00: 6:02 :0:::< :00>
>:_:0_::

 

 

 

 

 

 

 

.0_0_:0::< ::0 03:03:07: :0: 00:55 :030885 was 05030:: :0 >:088=m ”m 030:.

 

 

1".
3

55005:? 00% . 93:06:03.8“ 526 - .m.U "8002*

 

 

 

 

 

 

 

 

 

.003 .mmm .0 300t0 0E $50850 5 Hz 0383 0:0
£3 205300 x300 20: mwi 0005000 30000800 2 $008 05% .m.O Boa—£2 .0 H0 tax—Sm mom—
20200090..
0030 1000 .Ed 000 1000.000 .0300 .Ed
-00: 0. 5:00 95:50. 022550 05 5 exam 0:0 .00N .00 Z 00 me>mm
00 H>mz<mh 00m: 300.: 2:: mOQ<m :05 0380000 $04 1523::— _0 00 £25 30—
.o0m _ >0 000.60.. >200 mam mOD<m .0 00 £25 50.
9:000:030 swarm .80>0m .002 .3 00000005
00 3032 0003000 $0500 0.; 06: .020: 050: 00007, $000 -
.m.0_0E0> :05 005:: >300 £5 E 00:00—02 exoc— -
00N_E_Eo 0003 $=0E0>0E 030$ §w 3 0030.05 000% gm - H>mz<mrr .0 .0 £25 .00—
0030“
3:25:00 £30m E0830; 00m: 6002 553x 30>
biota

 

 

 

 

 

 

 

$20.80 a 03.0.

 

 

:03025; 00% . Hip—0.30.5. 50:6 - .md "0082*

 

 

34

 

 

 

 

$.00 00.3% 08006:. -
00:00:08 $3-0m 3 00000005 3.0-529?
.0: 0:0 300:; 00.0 :0 3000.0 8. .0035 0030
0050000 0... :05? 00.0.8 :0w:0. 32%..” .0 00

003 0:0::00: mam 00.30.20: 02 0003 3 at. 25 000003.- 5:032:03 03.

0030000 0:03 $0050. 000008005

30:0 :0. 800:0 03300 07. - PM. 00:30 5.3
003 >0 8.: 25 000003. - *m..0 30:00:30 £02m. 03.

00:00
00:08:50 3.30”. 82500; 03.3 .0002 0053‘ 80>
09:05.

 

 

 

 

 

 

 

 

5.0.08.2 0:0 8.00502 00.. 00.005 0.0.0. ma:— m=0.>0.5 .0 50055.5 "0 030,—.

 

 

35
In this study, preemption algorithms for a six-mile arterial of thirteen intersections in Ann
Arbor, Michigan was used as a test site. Far-side bus stops and two-phase intersections are
common throughout the network. TRAN SYT—7F was used to optimize the network signal
settings. Offsets are reestablished in the cycle following preemption. The model possessed
the skip phase capability. There were several conditions placed on preemption to limit ex-

cessive preemption calls.

Chapter 3

Research Approach And Data Collection

The previous research has established that the benefits of preemption are negated if the
traffic signal system does not return to the underlying signal setting once preemption is
initiated. The inability of an algorithm to reestablish offsets once signal preemption
occurred may also adversely affect road user costs. The literature showed that there is no
computer model available that can simulate the bus preemption signal (BPS) operation
directly for several preemption strategies including skip phases, and then return to the

optimum signal setting. In this research, these shortcomings were overcome.

3.1 Alternative Plans

In the search for an appropriate computer model to study the BPS operation, several ele-

ments were considered. The model must be microscopic in nature so that it can track indi-
vidual vehicles, including buses, through the network. There must be a method to identify
buses in the traffic mix and their characteristics should be distinguishable from other vehi-
cles. The location of these buses with respect to the signal at any particular time must be

known. Fixed time signal settings must be changeable to accommodate bus preemption. In
addition, the model should be compatible with the traffic controller type that exists or is to

be installed in the corridor. The researched alternatives are discussed below.

36

37

3.1.1. Automatic Signal / Eagle Signal software-hardware interface:

It was known that the EPAC controllers of Automatic Signal / Eagle Signal and the
MONARC (Master Office Network Adaptive Real-time Control) system were to be
installed in the Washtenaw Avenue corridor. MONARC is a comprehensive computer
software package that provides centralized transportation management and control. Also,
it offers distributed area—wide on-street traffic control. It is a fully operational digital elec-
tronic data managements processor, receiving continuous real-time inputs from multiple
communication links. It generates status reports and failure reports, sensor reports, and

make adjustments to system traffic parameters.

Contacts were established with the Automatic Signal /Eagle Signal company to under-
stand the controller system and logic. The company’s headquarters in Austin, Texas was
visited to discuss the model to determine how to incorporate it into this study. The EPAC
controllers have the capability of processing a BPS operation. In order for the controller to
place the priority call, buses have to actuate the system’s detectors. This requires an inter-
face between the simulation model and the MONARC controller to simulate vehicle and

bus arrivals (see Figure l).

A microscopic simulation capable of generating traffic into the EPAC(s) was required.
Traffic volume data, turning percentages, signal timing and phasing, bus schedules, bus
stops, and geometric design would need to be coded into the simulation software. The
simulation model would then generate traffic arrivals which would be converted to
impulse signals into the EPAC controller. The controllers would, in turn, alter the timing

plans in response to the input. This process requires a hardware interface between the

38

Traffic Data

l

Simulation
—> Software

I

Hardware

 

 

 

 

 

 

 

 

Detector Input

V
EPAC(s)

T

Status Data

V

MONARC

 

 

 

 

 

 

 

 

 

Report Data

Figure 1: Automatic Signal / Eagle Signal BPS work plan

controller and the simulation model output from the computer equipment (personal com-

puters, PC).

The NETSIM model was selected for the simulation. The NETSIM source code was pro-
vided by McTrans. Although the computer / controller interface was made available by
Automatic Signal /Eagle Signal, after careful examination of the model, it was decided
that the structure of the NETSIM model was not well suited to the design of the detector

input interface required by the EPAC controllers.

39
3.1.2. THOREAU Model:
THOREAU is a recent Intelligent Transportation System (ITS) model developed by the
MITRE Corporation in November 1992. It stands for the Traffic and Highway Objects for
REsearch, Analysis, and Understanding. It is a microscopic and meso—scopic simulation
package. Extensive testing was conducted to determine the suitability of the model for the
bus preemption signal. Communication has been established with MITRE Corporation to

assist in evaluating the model and to explore potential enhancement of the model.

As a result of this assessment, several modifications were added to the model by the
MITRE Corporation. An understanding of what needs to be done to make the model capa—
ble of simulating the BPS process was reached. It was agreed that MITRE would enhance
the model to accommodate the BPS and MSU would deve10p the BPS algorithms. The
MITRE targeted date for implementing the enhancement extended beyond the MSU tar-
geted project completion date. Thus, while this effort is being continued, it was not suit-

able for this project.

3.1.3. TRAF-NETSIM Graphics and Simulation:

A third option was to enhance TRAP-NETSIM to provide BPS Operation. None of the
other models researched was found to match the strengths and capabilities of TRAF-
NETSIM. TRAF consists of an integrated set of simulation models each of which repre-
sents traffic on a particular environment (i.e., urban street, whether network or arterial,
two lane rural roads and freeways). NETSIM, which stands for NETwork SIMulation, is
one module of the TRAF family. It is a microscopic simulation model of urban traffic

(TRAF User Reference Guide 1994).

40

The model generates vehicles into the network randomly (poisson distribution) according
to a seed number coded in the data file. In this study, buses are introduced with uniform
headway according to the bus frequency. NETSIM applies interval-based simulation to
describe traffic operation. Every vehicle is a distinct object which is moved every time
period, and every variable control device (traffic signal) is also updated every time period.
A vehicle’s kinematic properties (speed and acceleration) are determined, as well as its
free flow speed, queue discharge headways and other behavioral attributes. Each time a
vehicle is moved, its position (both lateral and longitudinal) on the link and its relationship
to other vehicles nearby are recalculated. Vehicles are moved according to car following

logic and response to traffic control devices.

The current version of TRAP-NETSIM does not include the logic for bus preemption.
There were previous efforts by the FHWA to include this operation, but the work was not

completed, and therefore, was not embedded into the NETSIM model.

There are three issues involved in using NETSIM for simulating a BPS. First, the detec-
tion of bus arrivals at the intersection; second, interruption of the signal to respond to the

bus preemption call; and, third, the ability to test different preemption strategies.

It was found that the current model keeps track of every vehicle throughout the network
internally and does not provide this information as part of its output. However, by using
the graphical animation feature of the model it was possible to visually track buses along
the corridor (as buses are color labeled) and determine the signal status as the bus arrives
at the intersection. Also, NETSIM has the option of utilizing up to nineteen time periods,

each of which may describe changing conditions. These changing conditions are either

41

indigenous changes (internal to the system) or exogenous (external inputs prepared by the
user) such as changes in the signal timing, phasing, volume, and turning movement per-
centages, etc. With the combination of both graphical animation and different time periods

it is possible to simulate different BPS schemes.

The procedure is a) to detect the bus arrival in the vicinity of the intersection, b) determine
the signal status as the bus arrives, c) determine if preemption is to be awarded (based on
certain criteria to be established), (1) select a plan (different signal timing or phasing) to be
implemented, if any, and e) select the exact implementation time. These decisions are then
coded into the model and the system is simulated with these changes to secure bus passage
through the green light. Buses are monitored along the corridor in both directions (east
bound and west bound) and similar decisions are made at every intersection. Signal timing
plans are reset to their normal settings (offset, phases, and phase intervals) after every pre-

emption activation.

Some of the important characteristics of NETSIM’s time periods (T RAF User Reference

Guide 1992):

a. Each set of exogenous input data applies to (and remains constant during) one time

period.

b. Each time period is subdivided into a sequence of time intervals. Each simulation model
requested for a given run is brought in and out of the central memory once each time
interval. The time interval duration is typically set to the most common signal cycle

length in a study network. (It is set to 60 seconds in this study).

42

c. The duration of each time period must be an integer multiple of the time interval dura-

tion; 60 seconds.

In this study NETSIM’s signal control cards; signal phases, offsets, and durations (cards
35 and 36) may have to be changed in each time period to correspond to the BPS opera-
tion. Due to different cycle offsets and constrained by the time interval requirement above
(multiples of the cycle length) the beginning of a time period may occur in the middle of a
signal phase. However, NETSIM does not interrupt the signal cycle to adopt the new
change. Instead, the cycle is resumed as specified in the previous time period and the order
is carried out in the next cycle. Thus, an order has to be placed one cycle length before the
change is required. Also, it is worth mentioning that, in some cases, it may be necessary to
change both cycle offset time and phase duration to advance the green phase according to

the BPS.
3.2 Data Collection

3.2.] Background

Washtenaw Avenue in Ann Arbor, Michigan was selected as the test site. Ann Arbor is
located 43 miles west of Detroit and has a population of 250,000 (Ann Arbor Transport
Plan, 1990). Its population, as the largest city in Washtenaw County, is estimated at
115,000. Approximately, 30,000 of the 35,000 students enrolled at the University of Mich-

igan’s Ann Arbor campus live in the city.

Public roads and streets are under the jurisdiction of the City of Ann Arbor and the Michi-

gan Department of Transportation (MDOT). Public transportation is provided in the form

43

of bus services by the Ann Arbor Transportation Authority (AATA) and the University of

Michigan.

Growth and development of the city has led to increasing traffic congestion on major
streets, diversion of traffic into residential neighborhoods, and increasing conflicts
between University functions and non-University functions (Ann Arbor Transport Plan,

1990).

One of the recommended plans in the Ann Arbor Transportation Plan Update (Ann Arbor
Transport Plan, 1990) is providing transit-related improvements to increase capacity and
reduce congestion. In the field of transit, it was suggested that ridership should be
enhanced by improving services. Although the study recommended different ways for
enhancement, this research will be addressed to improving bus schedule reliability by sig-

nal preemption as a mean of encouraging automobile drivers to divert to transit.

The AATA operates twenty-two fixed routes transit lines in Ann Arbor and the surround-
ing communities. Ninety-three percent of all Ann Arbor residents are within one-fourth
mile of a route. Most routes operate with 30-minute service through the day, but the Wash-

tenaw route is one of two that operates with a lS-minute headway during peak periods.

Ridership has increased during the last few years to its level of about 4 million riders in
1990. In total, transit trips make up about one percent of all trips made in the Ann Arbor-

Ypsilanti Urbanized Area.

3.2.2 Network Selection

Washtenaw Avenue, east of the central business district has been identified as one of the
roadways that exceeds its design capacity (Ann Arbor Transport Plan, 1990). It is one of
the busiest corridors in the city. It runs from the west, crosses the CBD and continues to
the east through the city of Ypsilanti, see Figure 2. A major Ann Arbor Transit Authority,

east-west bus route runs through the corridor.

The eastern part of the corridor, between the Golfside / Washtenaw intersection on the east
and South University / Washtenaw on the west, was selected for this study for the follow-

in g reasons:

1. This particular corridor has been identified by the Ann Arbor Transportation Plan
Update study (Ann Arbor Transport Plan, 1990) as one of the roadways with a major

capacity deficiency.

2. Based on previous experience, it was decided that closely spaced and heavily congested
intersections (e.g., downtown Ann Arbor) are not good choices for signal preemption.
Furthermore, bus routes run in all directions (north, south, east and west) in the CBD

area, which makes it more difficult to improve overall service by implementing BPS.

3.2.3 Data Collected

The following data were collected to study the bus preemption operation along Washt-
enaw corridor:
A. Geometric Design: intersection geometry including number of lanes and lane configu-

rations and distances between intersections. Most of these data have been provided by

GL/xéltH

OUOBXIO

45

O
C(-

6:»
0
<2»
0

. \Y
Jams», q

NOOUBd

'3 s
5 NO"
5:"
0)

.-

HINBAdS

HI l, .I"..,I}}"F'

3018:1109

 

Map of Washtenaw Avenue - East, Ann Arbor

Figure 2

46

the Ann Arbor - Ypsilanti Urban Area Transportation Study Commission or deter-

mined from the city map.

B. Traffic Related Data: traffic volumes including morning and evening peak hour and
daily volumes and turning volumes (peak hourly volume are provided in Appendix A).
Part of these data were provided by the Ann Arbor - Ypsilanti Urban Area Transporta-
tion Study Commission and MDOT. Students from Wayne State University collected

data on the average and maximum queue length, and pedestrian intensity. Video

taping of several intersections was conducted to calculate the stop time delay (this data

was used for model calibration).

C. Signal Timing: signal phases and timing. The current timing plan for signals along
the State trunkline in the City of Ann Arbor was provided by the City and MDOT. Sig-
nal timing was collected in the field, and it was determined that different intersections
have different cycle lengths which prevents progression along the corridor. Since the
objective is to compare bus preemption against a “good” timing plan, it was decided to
maintain the same signal phasing but to optimize the cycle length, the green splits, and
signal phase offsets along the Washtenaw Avenue Corridor. This has been achieved
using the TRANSYT-7F computer model. The results of this optimization are utilized

in the simulation model.

D. Bus Related Data: bus schedule, bus routes, bus headway, bus stop locations, bus rid-
ership, and bus dwell time. Most of these data have been provided by AATA. How-

ever, bus stop locations were determined in the field.

47

The data were collected in the period of Fall 1993 to Spring 1994. From the data, it was
determined that the morning peak hour was from 8:00 to 9:00 and the evening peak hour
was from 5:00 to 6:00. Also, it was decided that evening peak hour is the ultimate peak

period. However, data were collected for both time periods.

Initial collection of queue data, speed limits, pedestrian intensity, and geometric features
was conducted in the Fall 1993 (see example, Figure 3). It was decided to use video
recording of traffic conditions at several intersections for model validation. This was con-
ducted in the morning and evening peak periods in Spring 1994. The video recorded data
were used to derive the stop time delay, the number of vehicles stopped, and to calculate
the average stopped delay time.These were compared with model output results (Kha-

snabis 1994).

3.3 Model Calibration

The data collected in the field and from MDOT, AATA, and the City of Ann Arbor for the
study network were coded into NETSIM. NETSIM’s link-node diagram for the study net-
work is shown in Figure 4. Nodes numbered between 1 and 33 represent actual intersec-

tions, while nodes numbered between 8000 and 8023 are dummy entry / exit nodes.

The network contains thirteen intersections along Washtenaw Avenue and one intersection
on each side along the cross street (if present). In this study, Washtenaw Avenue is consid-
ered the main street and all others are cross streets. The eastern part of the corridor (east of
Stadium Road) has different characteristics than the western part. The main street is wider

on the east. There are 2 lanes in each direction at Pittsfield, 2 lanes with turning pockets at

48

INTERSECTION DATA COMPILED FROM SELECTED
INTERCHANGES ALONG THE WASHTENAW CORRIDOR -

ANN ARBOR
INT ERSECTIONI: WASHTENAW - GOLFSIOE
DATE: SEPTEMBER 13. 1993
TIME: 7:30-8:30 AM
WEATHER: PARTLY SUNNY
OUEUE DATA:
WASHTENAWWB: LEFTTURN LANE: 0.1.1.1.0.0.1.0.0.1 AVE-0.5
CENTER LANE: 2.10.3.4.3. 5.5.3.0 AVE-4.4
RIGHT LANE: 2.5.7.4.3.8.3.3.2.5 AVE-4 2
WASHTENAWEB: LEFTTURNLANE: 3.2.4.1.3.2.6.3.1.3 AVE-28
CENTER LANE; 7.11.1.3.B.3.8.0.2.9 AVE-52
RIGHT LANE: 6.5.2.5.8.3.8.1.0.10 AVE-54
GOLFSIDE SB: LEFT TURN LANE: 2.4.1.1.3.4.8.3.5.3 AVE-3 4
CENTERLANE: 3.3.3.5.0.1.2.0.3.3 AVE-2 3
RIGHT LANE: 6.2.7.2.5.4.3.3.4.5 AVE-41
GOLFSIDE NB: LEFT TURN LANE; 1,1.2.0.1.2.1.3.1.2 AVE-14
CENTERLANE: 2.2.1.2.1,1.1_4,1.2 AVE-1.7
RIGHT LANE. 2.3.3.1.0.0.2.3.0.2 AVE-1.6
SPEED LIMITS. WASHTENAW . 40 MPH
GOLFSIOE - 35 MPH

RIGHT TURN INFORMATION: NO TURN ON RED - ALL APPROACHES

PEDESTRIAN INTENSITY: MEDIUM

LEFT TURN INFORMATION LEFT LEAD (LIGHT) ~ ALL APROACHES

CROSS STREET INTERSECTIONS
GOLFSIDE SB LIGHT AT PACKARD
GOLFSIOE NB LIGHT AT CLARK

Figure 3: Example of The Wayne State’s Field Data Collection. (Source: Khasnabis
1994).

49

0.0800: 5:: 0:: 00002 £33.07. 5.05055 "0 02%...

>02,me :93.

 

00:00:05.).

. U .ON— Esmfiwum

a \mNmy 1050.005
Q \@\ y 8m:
. \\ \\ g. <
@. ex s$ \ .
0::0>< .
30:00:00.5

e

.

fl 6. HQ.
0:...m:0>0o §\\ \@ .0 g,

. x o
085 a: .,
é a...

020.00

b.3032. .m Au

SO

Golfside, Carpenter, Huron Parkway, Yost. and Sheridan/Manchester, and 3 lanes at Sta-

dium Road. The western part, from Stadium to South University, has 2~lanes with no tum-

ing pockets.

Cross street traffic is relatively high at Golfside Street, Carpenter Road, Huron Parkway.
and Stadium Road. Arbor Land Mall lies on the north side of Washtenaw between Pitts-
field and Yost. Carpenter Road and Huron Parkway are controlled by actuated signals.
Golfside, Carpenter Road, and Huron Parkway are four-phase signal-control intersections;
two protective left turn phases and two right and through phases. The rest of the intersec-

tions have two-phase signals.

The model was calibrated against the average and maximum queue length measures col-
lected in the field, both manually and by video-camera recording. The simulation output
and field data were compared for several parameter values. Both evening and morning

peak hour conditions were studied. The model was calibrated until it reached a fair level

of conformity with field data (Khasnabis et al, 1994).

3.4 Sensitivity Analysis

The sensitivity of the model to several variables was tested. In the BPS process signal
green time is to be either extended (for the main street) or cut (for the cross street) in dif-

ferent time periods, as demanded by the bus preemption call.

Several intersections were selected to receive either a green extension or a termination of
cross street green. Fourteen time periods were utilized to analyze the sensitivity of the

model to the change. In the fourth and the ninth time period, main street green time was

51

increased by 10-seconds, and cross street green time was cut by lO-seconds. One upstream
link and two downstream links statistics were observed. These measures include vehicle-

link-tn'ps, total vehicle delay time, and average vehicle delay.

The model reacted logically to these changes. Generally, in the period where the signal
timing was changed and one period after, main stream vehicle-trips increased and average
delay decreased on Washtenaw Avenue, while the opposite occurred on the cross streets.
However, it is worth mentioning that these results were not uniform due to the random
vehicle arrival pattern, and the fact that the green time extension was selected independent
of the traffic demand or the location of vehicles approaching the intersection. The intent of
this calibration was to determine if the model behaves as expected. It was determined that
NETSIM is fairly sensitive to a change in signal timing. Tables 5A, SB, and 5C are pro-

vided as an example of this analysis.

The WSU group conducted a more extensive sensitivity analysis on several other vari-
ables (Khasnabis et al, 1994). These variables include a change in the green time, percent-
age of trucks on the main street, and presence of buses on the network. The study
concluded that N ETSIM is sensitive to these variables. A slight change in the input vari-

able leads to logical changes in the output.

52

.83....8 328:3? o5 ace? 5:2 293 :oEEeu .85 .«Ewte 2: a. museum...» _.

 

 

 

           

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

68.5.8.5 2a a a..." v 38.5.. u
«.2 3 3. 0:. _.~_ «.8 «.2 2.: QC «2 n2 ”.2 catacomfi
2 S 3 3 3 2. ca 3 2 Z 3 3: 52-505 as
h 2 c h on 2 a A o 2 e A:._._..fi>
0.2 3 mom v.2 we we q: «.2 5.3 3 3a 0.8 soZBm
a.“ 3 E o._ S. 3 3 3 no no 3 32 52-3, a-»
2 ca : m 2 w 2 m m o o “.51;
“ Hi “III
as S: E 3. E “.8 no «.2 3. 3 3 ”.2 a. guinea
3 mm 3 3 2 no 8 S no 3 3 S: 5.2;; as
o : m o S N o e v m n €9-29
"fll'l I Hi I
08.2 82K 08-3 83: MM 98-2 82K 82K 82K 82K 08-2 2-.
a: E. c: an mu 2 2 2 3g 2 9 : mm5:
88a 8.5m 85“. gram 85a 85a 85a 35.. Ste.— 33 8:8 85a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.Efimcowxm 595 muceoaméd at 55028:— ? EVE—Evok— mugoomé— he :33— a as $52 .253— amog u<m «Ema.

53

 

 

 

 

 

noZoom

  

 

   

52-5.5

 

   

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mmO—z

 

 

2 z a 2 an 2 : 2 3 2 : aE.-fi> II
in 2m 93 SN he MEN 3.. 3m 3m 3: «.2 SN 528m
3 an a.» mm 3.. 3 2 2 3 3 3 «.5 5.2-53 we
“55>

.Eemmeouxm 52¢ 3:30mé: at 5:89.35 “a 53:53.5 858mg: .3 .395 a ma mum—OE 253. .3m "mm «35.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fl 3. a: to. v.3 E E. v.2 E «.0 as «a la soioom
_ 3 8 3 No 3 do do 3 _.o 8 S 2: 52.55 W8
— _ _ _ _ v _ _ _ _ _ _ a525>
. a w t .
v.9 1 ma v.2 a: v.3 EN «.9 am 3m x: we =o>\8m

8 9o .3 Mo .3 no to 3 3 no no we =_2-fi> MEN
_ _ N _ n _ _ _ _ N _ €29

Li 1 MN

08-2 82K 32: 83: “MN 08-2 08-2 82K 82K 82K 82K :2

a; E. o: a“ we 2 on 2 E. 9 an : mmoz

85a 85a 8.5a eaten 85a 85m 85a 85a veto.— 85a 85a 8:8

 

 

5:0 59.0 mazeuomés at 55323:— 3 coca-=99:— muccoomé— .3 2:93— a as $.02 655m 5.5m % 5.52 "Um «EB.

 

 

Chapter 4

BPS Schemes and Algorithms

4.1 Bus Detectors

BPS operation requires a means of communication between the bus and a traffic control-
ler. Historically, this communication has been conducted by placing detectors in the pave-
ment that forms an electromagnetic field. These detectors identify the bus presence within
the vicinity of the intersection and communicate with the traffic controller, placing a call

for preemption. The controller then awards the preemption according to its built-in logic.

In this study, detection of bus arrivals and signal status were visually observed utilizing
NETSIM’s graphical animation. In the implementation stage, it is proposed to use smart
card technology to communicate between vehicles and traffic controllers. Although buried
detectors are not used in this study, schematic intersection configurations with detectors
were designed to develop the algorithms used in the research. The location of a bus-stop
relative to the intersection plays a major role in the BPS algorithm. The intersection con-
figuration for far-side and near-side bus stops are presented in Figures 5 and 6, respec-

tively.

Three to four detectors are needed at every intersection. Each of the detectors monitors the

bus arrival and progression at the intersection. The first detector is located at 400-500 feet

55

56

 

 

 

 

 

 

 

Bus Stop
""""""""""""" Stop Liri'e' ’(Déiéé'tii'r' 21);
g I Detector 2
Detector 1

 

 

Figure 5: Far-Side Bus-Stop Intersection / Detector Configuration.

 

 

 

 

 

 

 

  
 

a: Detector 3
§ 5 Bus Stop
8' . Detector 2
t,

 

Detector 1

Figure 6: Near-Side Bus-Stop Intersection / Detector Configuration.

57

(ft) upstream from the stop-bar. Its purpose is to detect the bus arrival in the vicinity of the
intersection and to assist in checking the traffic status. The second detector is located at
200 ft upstream and its purpose is to detect the bus progression toward the intersection and
to predict the signal status at the time the bus reaches the stop bar. The third detector is
used only in the near-side bus-stop case. It is placed just after the bus-stop station and its
purpose is to indicate the bus departure from the bus-stop. The fourth detector is at the
stop bar. Its function is to verify that the signal preemption scheme has been successful

(the bus has left the intersection).

4.2 BPS Schemes

As mentioned earlier, several combinations of the existing BPS schemes are possible. The

following are the schemes tested in this study:

(a) green extension, red truncation, no substitution (inhibit),

(b) green extension, red truncation, substitution (if necessary),

(c) skip phase, inhibit, and

((1) skip phase, substitution (if necessary).

Some of these plans work better than others at different intersections and under different
traffic conditions. Sensitivity tests were conducted, and the most suitable plan(s) for each

intersection were determined. In addition, a signal preemption plan conditioned on the bus

running behind schedule was tested.

58

4.3 BPS Logic

The following constraints were used in testing the effect of different BPS strategies:
(1) no preemption is allowed during two consecutive cycles,

(2) the minimum green time for any signal phase is ten seconds, and

(3) the maximum extension or advance of the green signal phase is ten seconds.

BPS algorithms and flow charts for different strategies were constructed to be imple-
mented as routines into the main computer program. This was initially developed to be
used with the THOREAU model enhancement alternative plan that was examined earlier
(refer to Chapter 4). As the bus arrival is detected in the vicinity of an intersection the fol-

lowing algorithmic steps are employed:

- The first check is to assess whether preemption has occurred in the last cycle. If yes, then

preemption is not permitted. If no, then proceed.

- If this is conditional preemption, is the bus on schedule? If yes, then preemption is nor

allowed. If the bus is behind schedule or this is not conditional, then proceed.

- Does the bus arrive on red? If no, there is no need for preemption. If yes, then preemp-

tion might be possible.

- Is time available for preemption? (i.e., how many seconds are needed to secure the bus

passage on a green light?) If more than 10 seconds are needed then preemption is not

59

allowed (unless this is a skip phase plan). If 10 seconds or less are needed, is the cross
street minimum green condition satisfied? If yes, then preemption is provided. In case of
the skip-phase(s) option, one may choose to skip a phase(s) if it provides the bus passage

successfully; minimum cross street green is to be completed before preemption.

- Select the suitable plan; advance green or green extension. Action is to be taken accord-

in gly.
- After preemption is granted a compensation or no substitution alternative is selected.

Flow charts that describe the detailed programming steps for both far-side and near-side

bus-stops are shown in Appendix B.

Chapter 5

BPS Simulation Results and Analysis

5.1 Study Cases:

There were six bus signal preemption cases studied in this research. These are:

1. Base case: N o Preemption. The optimal existing conditions were simulated and no spe-
cial treatment was given to the bus. This is the reference case against which all other

cases are compared to assess the impact of BPS.

2. Case 1: Green Extension, Red Truncation, No Compensation. The green signal phase
was either extended or advanced. There was no compensation given for the cross street

or the phases which had been reduced.

3. Case 2: Green Extension, Red Truncation, With Compensation: Compensation was
given only for phases that were reduced and were in high need to make up for capacity
loss during preemption. The need for compensation was determined based on average
vehicle delay, queue length, and number of vehicle-trips subsequent to preemption.
Compensation was provided only when cross street delay increased W
queue resulting from preemption could not clear in the cycle immediately following pre-

emption.

4. Case 3: Skip Phase, No Compensation. When the green extension or red truncation pol-

icy were not sufficient to let the bus pass through a green signal, phase(s) was (were)

60

6]

completely skipped for one entire cycle, i.e., green phase was extended for one full cycle

length. No compensation was provided in this case.

5. Case 4: Skip with compensation. As in case 2, compensation was given based on need.
A few intersections which have low cross street volume did not experience high delay
due to preemption and the queue was completely cleared in the next cycle. Therefore,

there was no need to compensate at these locations.

6. Case 5: Selective plans. Based on the results obtained from the first four BPS plans, the
most suitable plan for each intersection was selected. It was anticipated that the BPS
process could result in higher delays than the original signal settings, since preemption
causes the signal to deviate from its optimal timing. Thus, the most suitable plan(s) for

each intersection was determined to be the plan(s) that did not cause excessive delays.

7. Case 6: Conditional Preemption. In this case, the bus progression against its scheduled
arrival time at different stations was compared and the selective preemption plan, i.e,

case 5, was awarded only when the bus was late.

These seven cases were simulated, and signals were changed at specific times to accom-
modate the BPS operation. Most signals have two-phases and 60 second cycle length
except for three signals; at Golfside (120 second, 4-phases), at Huron Parkway and Car-
penter (Actuated, 4—phases). Preemption was not provided at the two actuated signal inter-
sections. Most locations have a typical two-phase signal with permissive left turns.
Golfside, Yost, and Stadium have different phasing movements and configurations. These

phasing configurations are shown in Figure 7.

62
— ’_I H l-__.
l
A. A Typical Two-Phase Signal
Yost Street

712
W

Washtenaw Avenue

I
2

B. Yost Street Two-Phase Signal

3
,2§
—‘§

3

Avenue

Stadium Road C. Stadium Road Two-Phase Signal

Golfside Street

__Jf+f

Washtenaw
Avenue

34
‘7 it

D. Golfside Street Four-Phase Signal

Figure 7: Phasing Configurations Along Washtenaw Avenue.

63

5.2 Analysis of BPS Time Period Specific Statistics

To understand the overall vehicle behavior resulting from preemption, each of the preemp-
tion strategies was simulated, and MOEs were collected every minute. NETSIM generates
cumulative network statistics as well as link and movement specific statistics. Cycle (or
time period) specific statistics were derived from the cumulative statistics. The type of sta-
tistics (link or movement specifies) that fit each intersection depends on the particular sig-
nal phasing of that intersection. For example, movement specific statistics were collected
for intersections with protected turn movements, and overall link statistics were collected
for typical two-phase intersections. The total number of vehicle trips, the total delay in
vehicle-minutes, and the average delay in seconds per vehicle were collected for each
cycle. NETSIM, also provides network-wide bus statistics and bus link statistics as part of

its standard output.

Statistics for two cycles before and two cycles after preemption plus the preemption cycle
(a total of five cycles) were collected. The first two cycles show the normal traffic behav-
ior without preemption, and the last two show the traffic behavior immediately following
preemption. Statistics were collected at the end of every sixtieth (60th) second. However.
since the cycle length is either 60 or 120 seconds and many cycles have an offset larger
than zero (the cycle does not begin and end at the beginning of an analysis period), pre-
emption may take place and its effect may be partially observed during the preemption
time period (third time period) and partially in the following time period statistics.
Depending on intersection conditions, the effect of preemption may be observed for sev-

eral cycles.

64
The three primary MOEs used are Vehicle-Trips, Total Delay, and Average Delay.

NETSIM defines these terms as follows: Vehicle-flips are the number of vehicles that

have exited the link during a specific period of time, Total Delay is the difference between
the free flow travel time and the actual travel time for all vehicles that exited the link dur-
ing a specific period of time. Vehicles that are in the link at the end of the analysis period
are counted in the time period as they depart the link. Average Delay (Seconds / Vehicle)

is a derived formula computed as = Total Delay (Veh-Min) * 60/ Veh-Trips.

5.2.1. Case 1 Preemption:

During the 45 minute simulation period, there were a total of eight preemptions involving
green extension or red truncation. Preemption time ranged from 3 seconds to full preemp-
tion (10 seconds). Each preemption was analyzed by studying the above mentioned MOEs
for two cycles before and after preemption. The full results of case 1 preemptions are pre-
sented in Appendix C. Tables 6, 7, and 8 present examples of case 1 preemption results for

three different intersections with different signal phasing.

Table 6 shows the results of preemption at a typical two-phase intersection (South Univer-
sity and Washtenaw). The first two time cycles represent the average vehicle-trips and
delay before preemption. Preemption took place in the third time period. The east-west
direction green time was extended and north-south direction green time was prematurely

CUI.

As a result of preemption, an increase in the number of vehicle trips and a decrease in
delay along the main street, accompanied by a decrease in vehicle-trips and an increase in

delay for cross street traffic would be expected. However, since the main street green time

65

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o..N o... N.mN N... o... 20288 .23.
2. mo w... ..o N... 2.2-5338 2.8.. 552 2-8
N m N N N 85.?
odm ow. ix 3% Sum £3.89 .23
no N... o. ed to EESEEB 28.. 5:8 or;
i. . N . . 35.-.?
:3 3 N... «N 2 .5289 .28
0... no a... m... ... 2.2-2.5128. 289mm. SNN
a w : w a m.......-..0>
ed 0. : 2 am 3.. 22%. .28
m. hm N. 0.. Wm 32-233330 eczema“? o;
o. E o. : o. 8.5.503
Bum"... mNnmue VNumne mNumue NNumue
$8st .2858 8:8...m-oo 8:853 moc8om-oo $082.85 8825 .5.
Bron. notch. cot»... eaten. 8:9. .0 8.332
be... an 08F 5.. 2...... Em .25... EN 2E... a.

 

 

 

.658: .93. m_ 538.5 2.: .BEEuoE fl
3......— oEE. FEE. .mu:euom-m .3 2.0 m. 6-2. .8...m 89.0 $2.85-.” .8 135...”... m_ 59.0 95......
.ofifi £52 .3 .52.... Z 5:98.35 .5... 5.2—E85 32:81 a 9. 83355333 3 «Eu...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Yo. m.» NM. ..: Qw 20283 .830
a. N.N ON in o.— A=.2-;o>v .500 9.38. 5.52 Fem 2
: o. N. N. N. animus L
1% SN mom cam ..: 238$ .25
3.. QN md .6 ON 3.2-53 .330 2.25 .mmm c-..
m. m 3 Z n 3:553
a... a... 31% a... .i
3 .3 N... m... N.N E 3.. N... o... .3 .8289 ER.
8... n. N. o... M... E w... o... N.N 2 2.2-53 ER. .88.. .83 em 1
c m. N c. m. : m m. m. m. 8....-fi>
m .. m .. m . m .. N. ..
.fifl. Nam... .wfifl... .uwm... .wawmv 8.2.8... as... a:
8.5.. 2...... 5m votes. 2...... 5v cocoa. 2...... Em voted 2...... Em cocoa 0::... .2 .0 3.3.8.2

 

 

 

 

.28.. .8... a. £58.... a... 82.58.... a. 8.8.. 2...... 2...... .8. .5.
888 :3. 25 m2 828 820 EE... .5 8.58%... d... 888 5.... S... :8... 8:83. .8
858m... 8 8.52 5:833... S... 8.2.58... .o .38.. a a 8.8.533 "N 2......

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2... ed... non QR 5.? ad... va odm .30 o... 5.5. m4... ~.~m Namm o6... 20>?va .330
S E q. a... N... o... 2. 2 2. o... S. 2. .N 3. ..m 5.2-5335 655m 582 .7...
. o. o. N .. m . .. v o v. m m w m 8......9
v.3. a... 0.8 3.... 3.. 2.. 3.: a... 2.. 3.2 3... o... a... 3... .2. 20288.330
«N o... N.N. nNN .N S N.N QNN 3. w... 3N ..m m... a... N... 5.2.8332. 655m :58 ..-:
o. c .. o o. a o 3 z o m. N . N. N. m......-..8>
jun" T "t” H JI
n... no. 8.... SN 2. 3.... N... 0.9. .38 3.. ..N.. SNN E. a... ”SN 23288.530
7 on ....N N... m... N..N 2. o... ed. Q8 5.. a... n... no 3N 3h 9.2.33.330 2:50.30 :-.
6
m .N m. v .N v. m .N n N NN N. 0 ON 3 m.......-..8>
3.. N... o... 0.8. a... v... n... Q... .No 32 2.. N8. 9N2 2... an. 20288.330
no. S... o... 3.. 3.. ..m EN 3.. ..m 0.... S... S w... o... ... 32-53380 852.88? ..-mm
.. w. o x mm m : a m n a . v N. N 8.5.505
"I” h ! u” h

N. ... .. N. .. .. m ... .. m ... 1. M. ... A

.mumne oNumuk Run"... 26".. MNunuh .
mucooom-om. mwcooomém. mucooomém. mucooomém. mocooumém. $352883. 5.38.0 VE...

not... 8.5.. 8.3.. 8.3.. 82.3.. .0 3.22.3.2

2...... 5:“. 2...... 5:5". 2...... BE... 2...... 2.83m 2...... 6...”.

 

 

 

 

 

 

 

 

 

 

 

.55.... .5... a. 55.8.... 5... .8558... a. 8.8. «5.... .55...
8.588%... E .5 8.. 5E; .8. .2... .85 880 5.58%.... .5. 5855.2 8.. a a. a $9.... .85
5...). .2 8.5.2 5:853... .5... 5.558... .8 .58.. a 8 825.53.35.33; a 2.5

 

 

68

was extended for only 3-seconds and traffic flow fluctuates randomly, the effect of pre-
emption was not very significant for any direction. For example, there was no increase in
west-bound vehicle-trips and no decrease in cross street vehicle trips as a result of pre-
emption. The later may also be attributed to the right-tum on red movement allowed.
While delay decreased during preemption (period #3) for the west bound direction, it did
not decrease for eastbound direction, since its delay is already low and fluctuates signifi—
cantly. Northbound delay was significantly higher as a result of preemption. Traffic returns

to its normal conditions during the fifth time period (2-cycles after preemption).

Table 7 shows statistics at Stadium Road, which also has a two-phase signal. During pre-
emption (3rd time period), the eastbound green signal was advanced for 10 seconds, cross
street (north bound) and east bound left turn green signals were terminated 10 seconds

early, while the west bound right turn has a continuous green arrow.

Although, more vehicles exited the east bound link during preemption (14 compared to 7
and 11), delay did not decrease. However, vehicles experienced a reduction in delay in the
following cycle (4th period) as a result of fewer vehicles being stopped. Since vehicle
arrival is fixed, and since more vehicles completed their trip during the preemption cycle,
there were not as many vehicles in the link in the following cycle (only 5). Westbound left
turning vehicles experienced a decrease in vehicle trips during preemption (11 compared
to 13 and 15). As a result, vehicles that were stopped during preemption, in addition to
vehicles arriving in the fourth period, left the link in the following cycle and experienced a
higher average delay (18.3 compared to 7.8 and 14.2). Traffic returned to normal status

after the first cycle following preemption.

69

Table 8 shows statistics at Golfside Street which has a four-phase signal. During preemp-
tion, main street (east-west) right and through traffic green signals were advanced by 10
seconds, main street left turns were terminated 10 seconds early, and cross street (north-
south) signals remained normal. As a result, eastbound left turning traffic experienced a
significant reduction in number of vehicle trips (7 compared to 12 and 14) with a signifi-
cant increase in delay that was carried on for several cycles, because the left turn lane was
already at saturation flow. However, west bound left turning traffic was not affected by
preemption since its vehicle arrival and discharge rate is very low (2 to 3 vehicles per
cycle). Therefore, all vehicles could exit the link before their green time was prematurely
cut. Furthermore, main street (east-west) right and through traffic experienced a slight
increase in their vehicle trips with a slight reduction in west bound average delay during
preemption and the following two cycles (3rd, 4th, and 5th period). However, only east

bound right turning traffic experienced a reduction in delay.

5.2.2. Case 2 Preemption:

By analyzing case 1 preemption results, it was determined that compensation should be
awarded only at Golfside Street under the green extension / red truncation preemption
plan. Thus, case 2 preemptions were exactly the same as case 1 preemptions, except that
compensation was provided at Golfside Street. The results are shown in Table 9. Green
time was extended for 10 seconds for main street right and through and cut from main
street left turns (3rd time period). To compensate, in the following cycle (4th time period),
lO-seconds were taken from main street right and through and were added to main street
left turns. A total of six cycles; two cycles before preemption, a cycle during which pre-

emption occurs (third time period), a cycle during which compensation occurs (fourth

70

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_.NN 3m N.Nm N.NcN o? Se o.NN Nam moo N.N”. wow oNo 9o N9 2% N.Nm N.Nm o.Nm 238833
9N 0.2 2. _ 2 o N : m 3 no 3 o.o 2: I. N.N 3 _.m 2.2-2.9128 65.8.2 :9
N : N _ 2 o N : m _ : e o 3 m m w m 85-2.9
an: 0.8 QNN 3.2 gm 0.8 $2 0.: N9. 8.: S... ns 35 3s :8 N32 9% an 22885.5
3N N.N n: «N o6 N.N. N.NN N.N o.N N.NN QNN 2: mi 08 3 m9 0.: N.: 33-5358 .anm.m :-N_
: v N_ 2 o N_ a 2 a o 3 3 c m. N N N2 2 8529
”.8 2:. we; 1:. we. NSN an N.No NEN Nam 0.9. ”an we. 3. N.NNN N9 3:. ”SN £38858
N.N o.oN New «m 3: 98 N.N QmN «mm 3 0.2 N.NN N._ «.2 n? 3 EN 9% 3.2-;9153 2.8mm .2
N a 2 n NN N_ v «N M: m NN N N NN N_ o oN 3 852%
«.2: N62 as. $3 N.N: o 22 N.N: can 3: a: ”.8 9N2 N.Na N.N: 9N2 22 n: 238958
5 SN N.N _.N_ 0.3 o q: o? 3 EN EN 2 B: New 5 ”.2 SN 2 Ezéozioo 258.3 :-mm
_ cm L v N 2 o m NN N : an m m NN _ v Nm N 85-55
# u
N; N 4 m .N 4 m .N r_ m .N a m .N 4 m .N A
mmumu... Rang. oNHmuh Sum"... mam"... mmumnh.
mucooowéfi mucooomdm— mucooomém— 3:88-02 mucoUoméN— mucooomém. mmocoéootm 8:025 xcfl
cocon— uotom cocoa cocoa cocoa no.5; No 85322
08F 55m 08F NEE 05:. 553 05F BE. BEN 28% 08:. EE

 

 

 

 

 

 

 

 

 

 

 

:389235 .5: :ezamsanU ecu 529:3...— ue ::8m 5 ma mozmzfimdfigmqaaazddfluufla & 035,—.

655m .3. a 5:025 2.: .Nezeasm .322. 5e .25 Em
2: 5 $3859.50 98 mgeuomé— .3 .50 95 2:5. :3 95...": 62% En: 1338033. .32qu—
Ev 25 En 2E. E 30 5:... 23 3.88%.: .5. u~o=a>c< «.5 h. a m 99.8 39.5 Eu: .2 .3552

 

 

71

time period), and two following cycles, are presented in Table 9.

Although west bound left turn green time was cut by 10 seconds, traffic did not experience
any decrease in the number of vehicle trips or any extra delay, because of its low traffic
volume. However, east bound left turn vehicles experienced a decrease in vehicle trips (7
compared to 14 and 12) and a slight increase in average delay (256.4 compared to 246.8
and 227.3) due to preemption. During the compensation period (fourth time period), more
vehicles exited the link (18 compared to 14 and 12) as a result of the 10 extra seconds
added to the green time. Despite compensation to the main street left turn phase, east
bound left turn delay continued to increase in the following cycles because the left turn

lane was at saturation before preemption occurred.

Main street right and through traffic experienced an increase in vehicle trips and a slight
decrease in average delay time during preemption. Vehicle-trips decreased and the average
delay increased during the compensation period (fourth time period), since 10 seconds
were cut from the green time. This direction returned to normal conditions in the follow-

ing cycles (fifth and sixth).

North and south bound signal phases were not changed, and any changes in their statistics

were merely due to random traffic variations.

5.2.3. Case 3 Preemption

Some of the bus preemption calls in case 1 and 2 were not awarded because of the lO-sec-
ond maximum preemption time constraint. Since a skip phase option was used in this case
(case 3), there were more opportunities for bus preemptions to be awarded. There were a

total of ten preemptions; five skips and 5 green extensions / red truncations. The time

72

period specific statistics for all preemptions are shown in Appendix D.

Under case 3 preemptions, traffic followed the same behavior as found in the first two
cases, in terms of increase / decrease in vehicle-trips and decrease / increase in average
delay. However, the effect on traffic behavior of the phase skipping preemption was more
significant than the previous preemption plans. A phase (or more) was completely skipped
and thus, no vehicles could exit that link (except for right turn on red). At most locations,
stopped vehicles for which green phases were skipped could exit their link in the cycle fol-
lowing preemption. Vehicles at Golfside Street had to wait more than one cycle to clear,

due to traffic volume close to the saturation level.

5.2.4. Case 4 Preemption

By observing case 3 preemption statistics, it was determined that skipping a phase at Golf-
side Street is the only case that warrants compensation. All other stopped vehicles clear
the intersection in the cycle following preemption without compensation. As a result,
there were a total of ten preemptions; six green extension / red truncation and four skip
phases, three of which included compensation. The time period specific statistics for all

preemption occurrences are shown in Appendix E.

Although compensation was provided for the skipped phases, traffic could not recover
from the adverse effect during preemption. Golfside Street’s east bound left turn statistics
remained disadvantaged for a very long period. By the time it started to recover another
preemption took place, and thus the delay continued to increaser towards the end of simu-
lation time; from 110.9 seconds/vehicle at time equals 5:07 (Table E. l) to 407.7 seconds/

vehicle at time equals 5:37 (T able E.10). However, when the phases were skipped from

73

Golfside Street, the adverse effect lasted no more than two cycles after preemption. This is
because cross street traffic volume to capacity ratio is less than that of the main street east

bound left turn.

5.2.5. Case 5 Preemption

For each previous preemption case, before and after statistics were collected. The overall
intersection statistics for the three periods before preemption and the three periods after
preemption are summarized below each table shown in Appendices C, D, and E. Five
periods before and after preemption were considered for Golfside (with compensation) to

capture the effect of compensation.

Vehicle-trips, total delay, and average'delay were calculated for each preemption strategy.
Strategies with minimum adverse effects were selected as the preemption choices for

strategy 5. Traffic behavior (queues and delays) were visually observed using NETSIM’s
graphic animation to further assess preemption impacts on intersection MOEs. As a result,

the following strategies were selected as the most suitable plan for each intersection:

Intersection 11 (Golfside): Cases 1 and 2 preemptions.
Intersection 2 (Yost): Cases 1 and 3 preemptions.
Intersection 3 (Pittsfield): Cases 1, and 4 preemptions.
Intersection 5 (Sheridan): Cases 1 and 3 preemptions.
Intersection 6 (Stadium): Case 1 preemptions.
Intersection 7 (Brockman): Cases 1 and 3 preemptions.
Intersection 8 (Austin): Cases 1 and 3 preemptions.

Intersection 9 (Hill): Cases 1 and 3 preemptions.

74

Intersection 10 (South University): Cases 1 and 3 preemptions.

As a result of these selective plans, there were ten preemptions; 7 case 1 preemptions, 2
case 3 preemptions, and 1 case 4 preemption. The results of this preemption plan (and the

case 6 preemption plan) are discussed in the next section of the paper.

5.2.6. Case 6 Preemption:

When the bus schedule was compared with bus progress through the network, there were
eight preemption occurrences; 4 case 1 preemptions, 3 case 3 preemptions, and 1 case 4

preemption.

5.3 Intersection and Link Overall Statistics

Since the maximum number of time periods allowed by NETSIM is nineteen, it was possi-
ble to simulate up to 45 minutes. Overall traffic performance at every link, in all direc-
tions, and for every intersection over the simulation time were summarized. Statistics over
the simulation period with and without preemption were compared for the first four cases
of preemption. Intersection statistics were obtained by adding all intersection inbound link
vehicle trips and total delay. The average delay was then calculated as before (Tables 10,

11,12 and 13).

Over a period of 45 minutes, the number of vehicles exiting the network (vehicle-trips)
under preemption should not be much different than that under no preemption. The only

difference might be due to the difference in the last cycle or two.

75

 

 

 

 

 

       
  

  

 

 

 
 

 

 

 

 

 

 

 

 

  

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

__ NE o2 — _.NN N.NN l_ o.NN SIN _ 3N 3E QNN N.NN _ .8 0.8 _ 258$ 33
= N.NNN QSN _ N.NNS N33 _ N.Nof Ex: _ N.NN: N.NN: :86 N.NNNN _ N.oNNc 0:; _ €2-53 38
5% NNNN 82 N34 NNS 8N4 5N4 NNS 4N3. EON cNNN BEANS
m N53035:: v Caucasus m 5508.335 N cowoofioucm — comuoomuous : N55083:: _ =a..5>O __
N2 .5 N.:_ 3: EN EN o.N_ 4d 4.5 22 N.NN N.NN _ 20288 33
§N N.NN 232 N.NNE 9o: 32 «Na N.N”. oNflN EVEN N.:N_ N.:...N _ €25,123
Na 8a SN NNv NNv E E 5 NNo NNN N_ 35.-.;
m - mm v - E m - S N - 2 _- Q : - S _ 9:5: 5.5m
New 9% _ q: q: EN EN Na. 2.9. _ 238$ 38
N9 _ N.NNN NSN _ <va N.NNV _ N.NN o8 _ N.NNV News. _ to? v: 32-;9123
N? or _N_ NNN - cNN Ln 3 N_ ates»?
m - om N - M: _- N_ : - 2 _ .258— 5.52
to N.N 3% NS N._N NdN N.NN N.NN N.NN can N.N: N. z: _ 2238 EB
N._NN N.NNN N.Nva QNNE QNNN 4.an 35: «No: N32 N82 N.NNVN N.NflN _ 32.62 53
No: 8: $2 NNN_ 82 E: vNNN NNNN Ne: 82 NNN_ NNN_ _ 35-59
m-c 7m: m-v LFNNU _.IN :.N _ BER—Eam—
N.NN N.NN N.No .N.NN _ 2; Ni _ cial Nm NS N.NN N.NS N.N: _ 20289 3.5
4.9% 3% N.NNON N.N3N_ 0.an N.NNN _ :5 0.20 _.N:N 3E: N.NNNN N.NNNN_ EEEEEB
No: as Na: 4N2 NNmrz lWNE 0N: ON: 6% NNM N42 :2 _ " SEEP.
N-N N-N N-N N-_ w— T: :-NN _ 2.5583
.23 so? 55, so? _ 5;» so? _ .23 so? _ .23 so? 55> 45$»

 

 

 

 

 

 

 

 

 

 

 

 

.8338... .355» 25 53> 853m 3:: as... N 85 x: 2%...

76

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

Né. N.N”. N.N: N.: 3.: N.N: NE N.N_ N.N. v.2 ,2: NE 20289 2200
.NoNNN .oNNNN ©va N.NNN noe- NNoN N. SN N.NNN NSN N.NNN N. SN c. E €2.55 38
SNNN NNNNN Ne: 82 No: 82 22 22 $2 NNN_ oNNN ENN 85.2.2
023-5—53qu 9 52585:: a cove-.5525 w 558525 N 550585 w 5508.55— =<...5>O

 

w.w_

 

 

n.5-

 

   

 

 

 

 

   

 

 

2028328

 

N.Nv _ o.NN

£2.55 55

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

—_#—

F—lflfl—T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

H JINS _ No [1% i l _ I! Hat-N29»
o - 5N w - NN - 6:25 5.5m
o.N_ oi 0.8 Ni 3: 3: NS 0.8 2: 2: 2538 Ex:
N.NN N.NN N2; $2 4.2 0.2 N.N. oi N.N: 32 252-20% 225
N2 N2 NNN NNN NN NN No No 2N oNN mac-2.2
F FEE o - mm m - 0N N mm c - vN 9:5: 552
NSN choN N.|Nll N.N _.: N.N: N.N: of N.N a N.NN N3 2389 5.5
0.9. NSN 2.2 N.NS _ 3.2 52 _.No_ N.NS Nd:- Nde = 32-6555
f oNNIII CNN NNN Now EN EN N5 N-Nm :2 :e Math-20>
i an o - S w - a N- w r o - N A .2513— “mam—II.
Q8... N.N; q: N.N2 3: N.oN 3; 0.2 N.NN _.NN N.N 3 2388 EB
32 N.NE $2 v.2N N.N: v.8. N.NNN N.NNN N.NoN o.8N €2-55 Eva
_- NoN 3N NNN 92 N3 NS NS «No NNN_ 4N2 35-29
_ 33 SN N-N N-N N-N. 0N 25.5.83
_ 5.? so? 52, =5? 5.3 so? 52, so? 5;» so; 5;.» so?

 

 

 

 

 

 

 

 

 

 

 

5:525 .2 2%..-

 

 

77

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9: 0.0— wfiw Néw wdm WON QMN QVN mumm- w.wb wdo @- co Azoaoomv %w_ofl_
Thom Odom wdmwo 0% o. Ew— m. _Vw— Md _ : @wa N.wme v.mmwm vfimmc o. w _ mo AE—Ztcox/v Nam—DO
meow Sam 89... Nomv co 3» mm — v mmmv va mm; vav 3% Oman math-hex,

m :Ofloow—OHE .v :Ofioowuoucm N 5500335 N :Ofioomuoucn — COUOOW—QQH M — N56955:.— =5LQ>C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

  

 

 

 

 

 

 

 

  

 

 

  

 

 

_
_
_
_
w: <2 3: N.:_ EN EN N.N_ N.N_ 32 22 N.NN N.NN _ 5258559
N.NN N.NN _ ENE N.NNS N.NN. $2 _ N.NN N.NN _N.NN_N N.vN_N_ N.NN: N.N: _ €2-55 555
NN NN NNN NNN NNN NNv N_N N_N NNN NNN NNN N_ 35.2.2
N -NN v-2 N-_N N -N_ _-N_ : -N_ _ 258.52%
N.N: N.N: N.NN N.NN N.NN N.NN N.N_ N.: N.NN N.NN <NN <Nv _ 5259 Eva
N._N N.NN N.NNN N.NNN N.NNN N.NNN N.NN N.NN N.NNN N.N? N.N:- 3._ €2-53 559
a: x: NNN NNN oNv New or _N_ NNN NNN NvN N_ 35.55
N -.NN v -N_ - JNVNN N -2 l L - N_ : -2 _ 25.5 582
. N.N N.N <NN N.N.N _._N I N.NN N.NN N.NNI Nle N.NN N.:_ N. z: 5289 550
= N.N_N N.NNN _ v.53 N.NNN_ N._NN N.NNN _N.o§ N.NN: N92 N.NN: <5NN N.NN: €2-5>:N5o
No: No: _ NNN_ INN: 22 r Ker—TNNN NNNN NNN_ 82 NNN. NNN_ Late-25>
— Nb ”TN mmw N-N N-N :L acammamam
N. N.N N.NN N.NN N.NN N.N: N.N_ N.NN N._N N.NN N.NN 3.2 N.N: 5288 .359
N.NNN N.NNN N.N—NN N.NEN N.NNN N.NNN N.NS N.N; N.NN: SE: N.NNNN N.NNNN €2-53 Eon
_. No: NE NNN: NNE 9.2 82 S: NN: $2 NNN_ NNN. 22 ace-25>
N-N N-N N-N N; 2-: :-NN 5555.83
53 .30? .23 so? 53 so? 55, 1:5? .23 so? 53 .50?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

.5558... .855» E; .23 8552 NE: 5:: N 3.5 H2 N.NN-N

78

 

I!

 

l‘

 

[I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N.NN N.NN I: N.: _ 0.2.. NS _ N.N_n N.N. _ <N_ N.N. _ N.N. NE 20288300
.SNNN .NNNNN N.NNN N.NNN _ 3.2. N.NNN _ N.NNN N.NNN — N.NNN N.NNN _ <NNN 3P 33-5355
TmNNNN uNNNNN N3; 8: .N: 82 E: NN: R2 :2 NNNN ate-:9

avg-x5382 9 880885 N 558225 w 5:83.35 N. 5883:: o sonooNcBE =<.._o>O
N._N 3N N.N_ N.N. 3N N._N ENRBNKENQ
N.NN N.NN N.Nv N.Nv <NN N.NN ENE-fiono
NN NN NE N2 NN NN 852$
2 - 8 I 95.5 5.8m .
<:| N.N: 20288 35 ,
N.NN N.NN N.N: N.N2 v.2 N.N_ <2 o2 _ N.N: Nd; 23-6338
NN_ NN_ NNv NNv NN I NN 21- NN _ NNN NNN 85.-.;

E - on N - mm m - em 1 - mm o - VN :55: 5.52
N.NN N.NNN N.N N.N <: N2 N.N2 N.N: <N <N N.NN N.NN 2028835
<2: N.N2 N.N2 N.N2 N.NN: N.N2 N. _ G N.NNN €2-53 Eon
NNN NNN N: E NNN NNN 3 SN ate-:3
NM: I. N N - N- N -N k: figs—NW”.—
N.2 - N.NN Ni 3; 3an flN N.N <N ENQBNVIENQ
NSN N.NN N.N2 <8. N.NNN N.NNN N.NNN N.NNN €22-53 Eon
NNN 2% ii NNN NNN" NNN NNN_ vNNCN ”SE-£5
N-N I N-N N-N N-N 28583 ,
.23 so? -55, so? as, sol; 5.? H5?

62.5.50 .: «zap

 

 

 

 

 

79

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S. 0...: on” n; NR :2 SN 3m 2: a: 38 3o _ Eoioomisuvfl—
3% 38 9E: Ems. n8: 3%. n82 98$ 38v 3m; 3:: $3,; §2€>v§5 __
$2 $2 8% 23 £8 88 8mm 8% 92% $3 8% $3 _ 85-59
W 55922.: v 5883:: m 5885:: N 53088:: _ couoofiefi : 8:00.285 _ __a.._m>o
3: $2 0.8. 32 N.NN SN ad oé 2: Q: 3: N8 _ 2028933
EN EN 75.: can: _ Se 32 _ Sm Em :3: 2527.2: 2% _ 32-59128 __
% $ 5 E mm am me E 8m 8m NS 5% _ 35-4.5
m -8 v -3 m - a N -2 _ -2 : -2 _ 25cm 58m
3: q: 9% 9% 3m v.2 S. q: Em .3 3n 3:. _ £0288an
3m gm 2% 3% 22m 3% we .2 32 ”gm 3% ES _ §2$>V§ua
2: m2 2m an g N? 8 8 80 0% 8V 3v _ 85-5%
m -NN v -2 m -8 N -2 _.P :.|c|_ _ 2:5: gfiz
ma 2 So «E. E can fix 2&4 0.3 fl; 3: ilie _ 2028933 _
no: :2 :8: Ev: _ 38 3S _ 3.3 we: _ 2% on; was v.22 _ €2-23?an __
E: E: 22 29 ”8. $2 82 one _ a: wLE: El Rm _ 35-5:—
m-o v-m m-v N-m _.m :L _ 65:553—
wfi SN $5 n2 v.2 ”.2 EN EN 3w 0.3 OS 2: _ 2028835
Ex 3% SEN «$2 $3 3% 33 NM? £2 23; own: «.3: _ 3.2-53an
as 3: £2 22 mm: 9.: 82 82 «m2 o8. go so _ 85;?
mé vé m-~ N; T: : -mm _ 2.558% U__
5.3 .505 .23 so? .23 _ so? as, 505 5.3 so; 55 4:0?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5.358...— .aeéB Ea 53> 85:5 3:: Bee m 3.5 “2 2...?

80

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 
 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5? 0.3 2: 2: 3. 0.2 3. mi 3: a: 0.: 0.2 2388 Eva
.wowf gm: 0:: 32 Sam 38 33 EN 38 22 3% 32 32-5% 53
33 2% $2 $2 3: as S: E: 22 so 28 88 SEE>
DEB-3.5332 O— COME-0&8“: a COEQOW—OEH w somuoomuoucu N. :Ofloofious O :OCoOmuoucu =<ahv>o
«.8 20289 Emm-
oe ma _ mam 3m _ 0.: a: _ 3.2-£5 35
mm mm m: m: 8- 3 . [ u_T u mag-50>
E - _m m - 3 m - R - 3:5: 5.5m
as nfi 3m 92 E: 3: q: _.: q: 2: £288 35
mg 53 E: 82 ofi ofi S; ”.3 ME: 32 25-5% 35
E v: 8m 8m am am Nm Nm own an mach-50>
w - cu m - mm c - X L. v5.3. 5.52
82 3: <8 3 g :m mam £389 35
5 gm 0% VS .2: 2: 0.? a: 3% $8 Aug-EC Eon
mrlmom 9H: 3% L cmm ommflfi 8m 3m 8w m9- mane-50>
Ufl S-Nm o-o_ w-o 5-x o-h 9.25qu
8% 3:- 92 0.2 n: we .3 n: v.2 3N ca o6 20288 55
$2 3: 3,2 03; mo: 2: 2: 32 3: mm: €2-55 35
as aw we- 5- E- 8,- cmv 3- So :5 85-4-5
58- S-o o-w w; To a; 28583
.23 so? 55, so? 5.? so? 55 so? as, so? as) so? 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

62.525 .2 2.3

 

81

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

   

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9: 92 :8 9; _ 9a 93 _ 9x 9% N0: 9: v.8: 93 20208120:
:8 8% 9%? 8%:- _ 982 92.: _ 9:9 982 Be? 8m; 9%: 83; 22.5% :20:
:8: 2.2 88 2% “8:.- 88 88 88 aa- wvmm EN 8% mace-50>
n 8:08:25 v 8:08:25 m 8:08:25 N 8:08:85 _ 8:08:25 2 8:08:25 =a.._0>c
we 2: 98: 9.8: 9: SN 9:: 0.2 98: SE 98: .8» 202080 :20:
3m 2m _ 3: 9%: _ 32 92: _ :8 98 _ 9%: 8%: _ 3%: 2% €2.88 :20:
% % E. E. :90. mm: 8 8m 8 I8 ate-£5
n - mm v - E m :N m - a: _ N: : - S ::8: 58m
3: 9: 9% 9% 9% 9% 9: 9: 8m :8 9% 33 £088: :20:
8m :8 2% 9% 9:: 9va 92 «2 9:2 85 98v 3% £2.58 ::0:
ma: mm: 2m 2m 8 NR :3 8 Re 8% 8w 3v mach-:9
n - mm v - m: m 8 m - SF : m: : - o: ::8: 5.52
no 2 fl 3% 9% 9% 9% 98 mm 9% 9:- 98: 9:: 20:88: :20:
9%: :2 9%: 2:: 3% 8% _ v.8: 9%: 3% 9m; 88: 922 25.5355
8: INS: Eel:- 22 a: $2 _ 8%: 0%: 2: 5: go So math-50>
m-o v-m m v m-m : N 2-: 5:58:85:
8m 8m 9% 9% 9E 92 9 a SN 8: 9% 2m: E: 20289 30::
9%:- 8% 8%: 9?: 9S 3% :3 88 ~82 83; Sam 98: €2.58 50:.
22 N8: :2 SN: 8: 8: am: 82 88: 3: ca :8 85--ng
n-v m N N; :.8 05.8803 __
55> :8? .23 so? so?

 

 

 

 

 

 

 

 

£22.82: 32.23 05.. 53> 8:828 0...: .39: .v 85 an: 03.:-

82

 

 

 

   

 

 

 

  

 

 

 

 

 

 

 

 

 

 

  

 

    
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

__ 9S 9% = 92 92 92 92 92 92 92 92 92 92 20282 262
= .22 222 52 9:2 9222 92m 92m 922 922 9.2 :02 9 E #752203 302
222 2:2 22 $2 22 22 22 22 82 .8 :8 252$
Ovmauv—hoguoz Ow COmuooevfi: O COmuQthvu—u— w :OmuoomuouCH N. cOmuUDmHUHP—u O COMHQOWHOZH— gk? =<uLO>O
98 92 9mm 92 _ I 202081202
__ 92 92 _ 9: 9mm _ 92 92 — 222-5% :25
mm I29 9: m2 2. 3 H 29Efi$
2 - :m a - 2 w - R J - 2:“: 2:8 2
92 92 92 92 92 92 9: 9: 9: 9 2 20282 :25
__ __ 92 92 :2 92: 92 92 92 92 982 32:22:22125
2 _ 2 _ 8m 2m 2 om mm mm 9% 29 25-9.5
o: - om a - mm lw - cm 9 - mm MTV 0 - vm 3:52 5.52
99m 98m 9m 9m 9: :2: 92 :2 9m 90 2.9m 92 20208125
__ = 92 92 9; 92 9:2 922 92 9E 2.2m 98m = 2222032202
I m2 1 2m 8m 93 o? own 28 3V 39 I M? 29:29
o: - mm j o - E w - o n - w o -M :55: 89m
2% 929 92 92 92 92 92 92 98 92 9a 9a 20282 Eon;
__ =MW 92:2 92: 922 922 2.2: 9:: 9:2 9:2 922 922 222-63 302
Ev a? 2? 2; 2; 2% 22 so So 25-9.5
2202. 9-0 0-x w-h 5-0 c-m ESGQHQB
23 so? so? 52, so? 23 so? as) so? 23 so?

 

 

 

 

 

 

 

 

 

 

 

 

 

62.5.80 .2 22¢

 

 

 

83

As far as average delay is concerned, most intersections experienced only minor differ-
ences over the 45 minute period. However, there are three key intersections that contrib-
uted significantly to the overall network statistics, because of their high volume and delay.
These intersections were Golfside (intersection 11), Carpenter (intersection 1), and Huron
Parkway (intersection 4). Although, no preemption was provided at Carpenter and Huron
Parkway because their signals are actuated, their statistics vary significantly between no

preemption and preemption cases.

Signal preemption had, in general, an adverse effect on both the Huron Parkway and Car-
penter intersections. The reason might be that due to preemption at an upstream node.
more vehicles were released into the link from the main street than the intersection could
handle. Due to this flux of vehicles, progression was disrupted by the extra vehicle arrival
time and volume. As a result, some of these vehicles could not clear the intersection in the
green time, thus causing extra delay. Since these locations were already near saturation
and the cycle length runs longer than two minutes, delay was significant at these locations.
Compensation and skip phase preemption plans proved to be poor alternatives for traffic at
Golfside Street, while the green extension /red truncation plan had no adverse effect in the
long run. N etwork-wide overall statistics under preemption were slightly worse than under

no preemption, and case 4 was the least favorable plan as shown in Table 14.

The average total delay experienced by only those vehicles travelling along the corridor in
an east-west direction was also examined. For every preemption case, the average delay
(in seconds per vehicle) at each link for east-bound and west-bound uaffic were added and

compared with average delay without preemption, as shown Table 14. These figures

34

indicate the total average delay that vehicles travelling from the first entrance in the corri-

dor to the last exit (east or west) would experience.

Table 14: Average Delay Over The 45-Minute Simulation Period

 

 

 

 

 

 

 

 

 

Base Case iCase 1 Case? Case 3 rCase4

w. Bound I 418.5 409.0 416.6 W 421.1

E. Bound | 305.0 306.3 310.8 330.5 343.3

Total Delay | 44.6 46.5 46.3 45.7 47.3

 

 

 

 

   

Note: The W. Bound and E. Bound delay represent delay along the entry route, while
total delay is delay per intersection.

The delay was higher in the off-peak direction (west bound) for both the base case and the
preemption cases because progression on Washtenaw Avenue in the evening rush hour
favors east bound traffic. Vehicles travelling west bound benefited from preemption, since
the green time was extended or advanced, and thus their travel delay was reduced, as in
cases 1 and 3. However, compensation for phases prematurely cut or skipped increased
travel time in both direction (case 4). Since main street traffic volume is relatively heavy,

this increase in delay outweighed the delay reduction gained during preemption (case 4).

East bound through vehicles were always disadvantaged by preemption regardless of the
plan used. The traffic volume in that direction is higher than west bound volume (Appen—
dix A), and as a result each time preemption was awarded their progression was disturbed
and delay was increased at the downstream node. It appears that for the heavy traffic direc-
tion progression is crucial and preemption increases travel time. The east bound through

traffic experienced the highest delay under the skip phase preemption plans, since this plan

85

involves the greatest signal disturbance.

The total network delay presented in Tables 10, 11, 12, and 13 includes vehicles travelling
in the east, west, north, and south directions. The network-wide delay under preemption
was higher than without preemption for all cases, because the network without preemption

was optimized and preemption deviates the optimum.

5.4 Cumulative Network Measures of Effectiveness (MOEs)

The microscopic traffic behavior for every link and at every cycle has been presented and
discussed in the previous sections. In this section, the network cumulative MOEs; overall
vehicle statistics, person MOEs, bus route M0135, and total bus link MOEs for all six pre-
emption cases are discussed and compared with the basic no preemption case, for the total

simulation time.

The first bus enters the network (from both directions) approximately 8 minutes after the
start of the simulation. Therefore a significant portion of simulation time (8/45) has been
processed before the first opportunity for bus preemption. Also, it was observed that net-
work delay increases at the beginning of simulation as the network becomes loaded with
vehicles. Therefore, it was decided that it is more reasonable to collect statistics after the
network reaches steady state condition. It was determined that at time 5:23 the network
reaches steady state with two buses from each direction in the network and conStant

delays.

Table 15 shows the cumulative network statistics for the steady state period (between time

5:23 and 5:45). As defined earlier, vehicle-trips are the number of vehicles that have

86
Table 15: Cumulative Network Statistics; With and Without Preemption.

 

 

 

 

 

 

 

 

 

Veh-Trips Veg-6111213 Min/BZE-éfips

Tia Pfeemption a 3469 E13 :83
Preemption Case 1 3426 165.8 2.90
Preemption Case 2 3435 165.7 2.89
Preemption Case 3 3408 170.57 3.00
Preemption Case 4 3398 178.45 3.15
Preemption Case 5 3424 167.83 2.94
Preemption Case 6 3449 170.65 2.97

 

 

 

 

 

 

 

 

 

 

completed their trip and exited the network from any entry point to any exit point (not
including vehicles that are still in the network). It is clear that the no preemption option is
the best plan (minimum delay) for overall system delay. This is no surprise, since the sig-
nal timing has been optimized and any signal preemption causes the signal timing to devi-
ate from this optimum. Preemption cases 1 and 2 (green extension /red truncation, with
and without compensation) are the Options that produce the lowest increase in delay to the
system (2.89 and 2.90 min / veh-trip), since they involve the minimum disturbance to the
system. The skip phase plans create the highest system delay (3.00 to 3.15 min/veh-trip).
NETSIM assumes an average occupancy of 1.3 persons per automobile and 25 persons per
bus. The bus occupancy figures were compared with actual bus ridership provided by the
Ann Arbor Transit Authority, and the number was close. Therefore, the NETSIM occu-
pancy default values were used to assess the impact of BPS on person MOEs, in terms of

number of trips, miles travelled, travel time, and total delay time, as shown in Table 16.

87

Table 16: Cumulative NETSIM Person Measures of Effectiveness; Before and After
Preemption.

 

 

 

 

Person
Trips

  

 

Person
Mile

       

   

Travel Time
(Person—Min)

Delay

(Person-Min)

 

    

Avge. Delay
Sec /tn'p

    
         

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

No Preemption 14921 5255 20264 12042 48.4
Preemption Case 1 14911 5227 20350 12174 49.0 '
Preemption Case 2 14918 5232 20409 12221 i 49.2
Preemption Case 3 16662 5659 22758 13769 49.6
Preemption Case 4 16506 5557 23230 14398 52.3
Preemption Case 5 14974 5267 20624 12383 49.6

NETSIM provides person statistics on a link-by-link basis (no network statistics). To
assess the cumulative network person MOEs, the link statistics were summed for the
steady state period (between time 5:23 and 5:45). The average person delay was calculated
as the total person delay divided by the number of person trips. Although the average
vehicle delay is indicative of the average person delay, the way each one was measured is
different. The average network vehicle delay is measured for only vehicles that have
exited the network, while the average person delay is calculated based on summing the
delay at each link, and thus includes the delay to persons still in the network at 5:45. The
average person delay ranges from 48.4 sec/trip (for the no preemption case) to 52.3 sec/
trip for case 4 preemption. Person delay measures followed the same trend as the vehicle
delay measures; the no preemption case had the lowest delay and the skip phase with com-
pensation case had the highest delay. Bus headways were not small enough to have a

major influence on the network person statistics.

Table 17: Cumulative Network-Wide Bus Statistics; With and Without Preemption.

88

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Mean Person
Route Bus Travel- Travel- Person Travel-
Tr1ps Time Time Trips Time

(Bus-Min) (Sec/Bus) (Min)

Original 1 2 68.1 1361.6 50 1702.9

Conditions 2 3 60.7 1214.2 75 1518.3

Preemption 1 2 64.4 1361.8 50 1609.2
Case 1 2 3 60.5 1208.1 75 1511.3 II

Preemption 1 3 66.4 1326.3 75 1659.2

Case 2 2 3 - 60.4 1208.0 75 1510.8

Preemption 1 3 69. 8 1296. 1 75 1746. 3

Case 3 2 3 60.6 1210.9 75 1515.0

Preemption 1 3 64.4 4 1288.0 75 1610.8

Case 4 2 3 60.7 1213.5 75 1517.9

Preemption 1 3 69.9 1301.4 75 1747. 1
Case 5 2 _ 3 59.3 :4: 1184.0 = :5 1481.3 ||

Preemption 1 2 68.7 1342.3 50 1716.3

Case 6 2 3 61.8 1234.6 75 1544.6

 

 

 

 

 

89

Bus statistics were collected in two ways, route based and link based statistics. N ETSIM
provides cumulative network-wide bus statistics per route. There were two bus routes in
the network, both using Washtenaw Avenue; route 1 (west bound) and route 2 (east
bound). NETSIM statistics are provided only for buses that exited the network (no consid-
eration for buses in the system). Therefore it was necessary to collect bus statistics on a

link basis to monitor the bus progression within the network.

As expected, skip phase preemption produces lower bus delays than the other plans, since
having the bus pass through a green signal is almost guaranteed. However, when a signal
phase is skipped, extra vehicles along the main street also take advantage of the extra
green time. These vehicles accumulate at the next downstream link. As a result, these
vehicle may form a long queue at the next down stream intersection and may not be able
to clear the link within the fixed green time. Therefore, a bus arriving at that link, which
might have originally faced a green light, may not be able to pass within the fixed green
signal, especially when preemption is not allowed at the particular time or location. This
phenomenon wa observed on the graphical display, with the result that despite the provi-
sion of BPS, bus mean travel time was only slightly lower than without preemption. In
some cases, travel time was equal to the no preemption case (route 1 of case 1, and route 2

of case 4) or even slightly longer (route 2 of case 6).

The total bus link statistics (bus-trips, travel time, and delay time) were summed to form
Table 18. The average bus delay was then calculated as: average delay (Seconds per bus-
trips) = total delay * 60 / bus-trips. The total bus-link-trips shows how far the bus has trav-

elled along the network. In the 45 minutes simulation time, buses travelled the greatest

90
Table 18: Total NETSIM Bus LINK Statistics; With and Without Preemption.

 

 

 

 

 

 

 

 

LinggtaBus- Travel'lime DelayTrme A332?

Trips (Mm) (Mm) (sec/B-Trip)

No Preemption I I__—_
Preemption Case 1 I I 79.6
Preemption Case 2 I I 78.9
Preemption Case 3 I I 79.1
Preemption Case 4 I I 76.1
Preemption Case 5 I I 77.6
Preemption Case 6 I I 82.9

 

 

       

distance (67 bus-link-trips) during preemption cases 3 and 5, and travelled the least during
preemption case 2 (64 bus-link-trips). However, the lowest bus delay occurred during pre-

emption cases 4, and 5.

Preemption case 5 (selective plans) has reasonably good MOEs for both buses, persons
and overall vehicles. That is expected since these selective plans (case 5) were chosen so
that the adverse effect of preemption (in terms of excessive delays and long queues) were
minimized. Although case 5 puts a limit on certain kinds of preemptions at certain loca-
tions, the bus gained more benefits than in any other plan (except case 4). That is because
when excessive delays and long queues were permitted to occur, as a result of BPS, the
whole network was disadvantaged including the buses. If a second bus arrived at the same
intersection from the other direction the bus would have a high chance of experiencing

extra delays and a lesser chance of passing the green light, without a need to stop.

Chapter 6

The Dynamics of BPS

The impact of different preemption strategies on Washtenaw Avenue under the existing
traffic conditions was discussed in the previous chapter. In this chapter, the effectiveness
of BPS under changing traffic conditions is analyzed. The sensitivity of BPS to traffic vol-
ume, main street to cross street volume ratio, traffic mix (percentage of carpools), and ran-

domness of vehicle generation were tested.

6.1 BPS Sensitivity to Volume

Traffic volume throughout the network was varied from 20 percent less than the original
volume to 20 percent more, with a 10 percent incremental change. These different volume
cases were tested with and without preemption, for a simulation period of 45-minutes. The
case 5 preemption plan (selective preemptions) was applied. The results are shown in
Tables 19, 20, 21, and 22. Also, the overall vehicle statistics and the total bus-trip-links

statistics are shown in Figures 8 and 9, respectively.

Table 19 and Figure 8 show the network cumulative statistics with and without preemp-
tion. The overall traffic was better off without preemption because, as discussed earlier,
preemption results in a deviation from the optimum signal settings. The adverse effects of
preemption on overall vehicles MOEs at low traffic volume was less than the adverse
effects at high traffic volume, because progression is very crucial at higher traffic volumes

(as discussed in Chapter 6).

91

92
Table19: Cumulative Network Statistics; With and Without Preemption.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Veh-Trips Veg—clflaoyurs Min/[1)]:llil-Xl‘rips
+20% Volume 3753 327.25 5.23
With Preemption 3750 327.06 5.23
+10% Volume 3619 243.917 4.03
With Preemption 3584 254.08 4.25
Base Volume III 3469 163.43 2.83
“nth Preemption 3424 167.83 2.94
-10% Volume 3212 104.62 1.95
\Mth Preemption 3194 106.15 1.98
-20% Volume 2800 82.03 1.76
With Preemption 2815 82.74 1.76
6
I No Preemp.

 

 

 

With Preemp

 

 

 

 

 

Average Delay (Min/Veh-Trip)
00

 

 

s :\

O
1

20% 1 0% Base -1 0% -20%
Percent Volume Change

Figure 8: Network Average Delay Due to Volume Change

 

 

93

Table 20: Cumulative NETSIM Person Measures of Effectiveness; Before and After

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Preemption.

Person Person “(g/gsgjfm Delay . Av ge. Delay

Trips Mlle Min) (Person-Mm) Sec / Person
+ 20% Volume I__ 32519 23883 90.3
With Preemption I 15793 5479 32309 23729 4 90.2
+10% Volume I 17316 5813 28676 19430 67.3
With Preemption I 19078 6199 31595 21605 67.9
Base Volume I 14921 5255 20264 12042 a 48.4
With Preemption I 14974 5267 20624 12383 49.6
_—————I__ 15167 7629 I 33.6
With Preemption 13563 4787 15219 7733 | 34.2
— 20% Volume ' 12058 #72731 12621 5975 ' 29.7
With pmemptionI 12779 4274 12719 6021 II 28.3

 

 

 

 

At a very low traffic volume, deviation from the optimum was not as critical since the dis-
advantaged traffic (cross street traffic) is low. Furthermore, the main street traffic may gain
some benefit during preemptions even though it may loose these benefits due to the loss of

progression at the downstream intersection.

Under very high traffic volume, many intersections either reached saturation or became
over saturated. Although preemption might have provided some benefits for the main
street through traffic, the same traffic would most likely be stopped at the downstream
node since the links were already over loaded. Thus, any gains for through traffic during
preemptions were most likely lost at the downstream intersection. The increased level of

congestion is apparent in the recorded number of vehicle trips as the volume is increased

94

in increments of 10%. When going from 20% to 10% below the base volume, the vehicle
trips increased by 13%. The respective numbers for the remaining volume increases were

7%, 5%, and 5% respectively.

As far as person measures are concerned (Table 20), preemption had little effect. How-
ever, at a very low volume rate (-20%) person delay under preemption was lower than no
preemption. This is due to both the priority given to buses and to the fact that at low vol-

ume the bus passenger percentage in the traffic mix increases.

Bus statistics show that generally bus travel time was shorter and delay was less under
preemption (Tables 21 and 22). The bus mean travel time was shorter under lighter traffic
volume and buses traveled longer distances within the 45 minutes simulation period.
Under heavier traffic volumes bus route 2 (east bound) benefitted from preemption, while
this route did not benefit under lighter traffic volume. However, route 1 (west bound) gen-
erally benefitted from preemption. Except for +10% volume, the average total bus-link-
trip delay time (Table 22 and Figure 9) was less under preemption. At low volume, using
preemption, buses could travel longer distance than at high volumes. Bus travel time and
delay decreased with the decrease in traffic volume. However, at very low volume (-20%)

the delay and travel time leveled off.

 

  

 

95

Table 21: Cumulative Network-Wide Bus Statistics; With and Without Preemption.

  

 

   

 

 

   

 

   
 

___.1._____
Total Mean Person
Route Bus Travel- Travel- Person Travel-
Trips Time Time Trips Time
(Bus-Min) (Sec/Bus) (Min)
1 2 66.4 1811.9 50 1660.8
2 2 65.5 1396.2 50 1637.9
1 2 63.9 1753.3 50 1598.3 II
2 2 62.8 1353.7 50

1570.8

 

   

 

   

 

   

 

 

 

   

  

 

 

 

 

    

 

With

   

 

  

Preemption

2

Volume 2 2

With 1 2

Preemption 2 3
Base Case 1 2 =

Volume 2 3

3

3

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

- 10% 1 3 65.5 1308.6 75 1637.5

Volume 2 3 57.0 ' 1138.8 75 1423.8

With 1 3 62.4 1247.4 75 1560.8
Preemption 2 3 56.9 1 137.6 75 1422. 1

- 20% 1 2 64.6 1315.8 50 1615.8

Volume 2 3 56.9 1138.3 75 1422.5

with 1 3 63.3 1264.2 75 1581.3 I
PreemptionII 2 3 58.9 1264.2 75 1473.3 II

 

 

96

Table 22: Cumulative NETSIM Bus Statistics; With and Without Preemption.

Total Bus-

Link-Tl'ips (Min)

Travel Time

Delay Time Avge. Delay
(Min) Sec / B-Trip

 

 

 

 

 

 

 

 

  

 

   

 

 

  

 

 

   

 

 

 

+ 20% Volume I 57 131.9 97.8 102.9 1
With Preemption I 57 126.8 92.7 97.6
_——I 61 121.2 84.5 83.1 I
With Preemption I 61 124.0 86.3 84.9
Base Case Volume . 65 127.4 86.2 79.6
With Preemption 67 129.3 86.6 77.6
- 10% Volume 66 122.4 81.3 73.9
With Preemption 66 119.3 78.0

 

- 20% Volume

   

     

    
    

 

   

With Preemption

 

 

 

 

Average Delay (See/Bus-Trip)

.3
N
O

 

...L
o
O

 

I No Preemp.

 

 

 

on
O
l

O)
O

.b
O

N
O
l

 

O
1

   

With Preemp

 

 

 

10% Base

Percent Volume Change
Figure 9: Total Link Bus Delay Due to Volume Change

 

-1 0% -20%

97
6.2 BPS Sensitivity to Volume Ratio

In this test the volume and the main street to cross street volume ratios were varied to
determine the sensitivity of BPS to these changes. In NETSIM, traffic volume is coded
only at the entry nodes and not at each individual intersection. Thus, it was not possible to
change link volumes by a constant increment for the whole network. However, changing
main street and cross street volumes is feasible for a single intersection. In this test, BPS
was simulated under different volume ratios for a typical two-phase signal at a two-lane
(in each direction) isolated intersection. All preemption plans of green extension / red

truncation, skip phase, and skip phase with compensation were tested.

Traffic volume ratios were selected to be 2:1, 3:1, and 5:1. It was determined (using com-
mon sense) that BPS for volume ratios less than 2:1 is not reasonable and for ratios higher
than 5:1 will, most likely, reduce delay. Main street volume was chosen to range from

1000 vehicle per hour (VPH) to 2000 VPH, and the corresponding cross street traffic vol-
ume was calculated. The average turning percentages was set at 20% from the cross street
and 7.5% from main the street. A five minute bus headway was chosen for both directions,
so that the effect of bus presence, and thus preemption, on the network overall statistics is

not negligible.

The following volumes and ratios were used in this study:

Ratio Sxmhnl MW
Upper 2:1 U2:l 1750 / 875
Middle 2:1 M2:1 1500 / 750

Lower 2:1 L2:l 1000 / 500

98

Upper 3:1 U3:1 2000 / 667
Middle 3:1 M3:1 1500 / 500
Lower 3:1 L3:l 1000 / 333
Upper 5:1 U521 2000/400
Middle 5:1 M5:1 1500 / 300
Lower 5:1 L5:1 1000 / 200

The reason that the main street volume for the upper 2:1 ratio was 1750 VPH instead of
2000 (as suggested earlier), is because under a two-phase signal and a 2:1 volume ratio the
intersection was over saturated, and the queue continued to accumulate on both the main
and cross street directions throughout the simulation. Thus, it was determined to reduce

volumes to 1750 VPH: 875 VPH (2:1 ratio).

All nine volume cases were tested under no preemption, preemption (green extension / red
truncation and skip phase) without compensation, and preemption with (skip phase) com-
pensation; a total of twenty seven cases. For all 5:1 ratio cases, because their cross street
green signal time was already at its minimum (10 seconds), the only preemption strategies
were skip phases with and without compensation. A maximum simulation period of 55
minutes was achieved. Overall vehicle, person, and bus MOEs were evaluated. Also, over-
all intersection vehicle statistics for four cycles before and four cycles after preemption
(including the preemption cycle) for different preemption plans were studied. The results
are shown in Appendix F. A summary of these results are shown in Figures 10, 11, 12, 13,

and 14.

99
6.2.1 BPS Overall Statistics

Figures 10 and 11 show the overall vehicle delay and the average person delay for high
volume (main street volume of 1750 VPH or 2000 VPH), medium volume (main street
volume of 1500 VPH), and low volume (main street volume of 1000 VPH) ratios. Vehicle
delay generally increased with preemption. The adverse effects of preemption (in terms of
delay) were very significant at the lower volume ratios (2:1), becoming insignificant at the

upper volume ratio (5:1) and for low cross street volume.

The preemption with compensation plan was better for very low main to cross street vol-
ume ratios (2:1). Because of the high percentage of the cross street traffic, losing green
time during preemption had a significant impact on delay. Compensating for this time loss
is beneficial. However, compensation (for skipped phases) at very high volume ratios (5: 1)
added more delay to the intersection. In general, the 3:1 ratio is the border line, above
which person and vehicle statistics favor preemption with no compensation, and below
which preemption might not be favorable and if it is provided, compensation would be

warranted.

As far as bus statistics are concerned, they generally followed the same trends, as shown
in Figures 12, 13, and 14. Bus travel time and delay reductions were relatively more sig-
nificant at low volume ratios (2:1 and 3:1) and less significant at the higher volume ratio
(5:1). At the high volume ratio, main street green time is naturally much longer than the
cross street green time, and the bus has a better chance of facing a green light as it arrives
at the intersection. Thus, the number of bus preemptions needed would be less than with

lower volume ratios. Unlike the vehicle and person statistics, the cross street volume rate

Delay (Mln/Veh-Trip)

Delay (MInNeh-Trlp)

Delay (SeclVeh-Trlp)

1 .60
1 .40
1 .20
1 .00
0.80
0.60
0.40
0.20

1.60
1.40
1.20
1.00
0.80
0.60
0.40
0.20

1 .00
0.80
0.60
0.40
0.20

 

 

 

1(X)

 

 

 

 

I No Preemption if
I Preemption it
1:] Compensation

 

 

 

 

1750:875 2000:667 20002400
Main to Cross Street Volume Ratio

A. High Volume

' I I No Preemption
7" }};}.,_.;__..__, I Preemption
1:1

 

1500:750 1500:500 15002300
Main to Cross Street Volume Ratio

B. Medium Volume

I No Preemption
I Preemption

El

 

 

1 0002500 1 000:333 1 000:200

Main to Cross Street Volume Ratio
C. Low Volume

Figure 10: Network Average Vehicle Delay

 

Delay (Sec/Person)

Delay (Sec/Person)

 

101

 

 

 

I No I Preemption
I Preemption ‘1;
El Compensation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 750:875 2000:667 2000:400
Main to Cross Street Volume Ratio

A. High Volume

 

 

   

i173}? I No Preemption
5.; I Preemption _ 3;:
33235:; ElCompensation

 

 

   
 

 

 

 

 

 
 
  
 

 

 

 

15002750 1 500:500 1500:300
Main to Cross Street Volume Ratio

B. Medium Volume

 

 

' ' on. em
f_::}?% IPreemptlon' I",
El Compensatlon

 

. .\' .5'

 

 

 

 

 
 

 

 

 

 

Delay (Sec/Person)
N
o

 

1000:500 1000:333 1 000:200
Main to Cross Street Volume Ratio
C. Low Volume

Figure 11: Person Average Link-Trip Delay

Delay (Sec/Bus)

Delay (Sec/Bus)

Delay (Sec/Bus)

 

 

102

 

1.;1 I No Preemption
I Preemption
El Compensation

 

17502875 20002667 20002400

Main to Cross Street Volume Ratio

A. High Volume

I N9 Preemption
9IPreemption
D

 

1 5002750 1 500:500 1 500:300
Main to Cross Street Volume Ratio
B. Medium Volume

I No Preemption
I Preemption - _

 

1 0002500 1 0002333 1 0002200
Main to Cross Street Ratio

C. Low Volume

Figure 12: Average Total Bus-Link-Trip

175 9;; . w 2 M
' ' 9’ 9" INo Preemption
IPreemption
uensafio

—l

h‘. -
A.
.\

Mean Travel Tlme
(Sec/Bus)
a

 

100
75
1750:875 2000:667 2000:400
Main to Cross Street Volume Ratio
A. High Volume

ms
962 -“
\ ,

I No Preemption
I Preemption
. nosatin

Mean Travel Time
(SeclBus)

 

1500:750 1500:500 1500:300
Main to Cross Street Volume

B. Medium Volume

.5
\l
01

I No Preemption
I Preemption

_L
01
0

§

 

Mean Travel Time (Sec/Bus)
u RS
01 01

1000:500 1 000:333 1 000:200
Main to Cross Street Volume Ratio

C. Low Volume

Figure 13: Bus Route 1 Mean Travel Time

104

 

 

 

I No Preemption
IPreemption.
£2353; El Compensation

 

 

 
 

 

 

  
   
   

 

Mean Travel Time

............

 

 

 

 

 

1 750:875 2000:667 2000:400
Main to Cross Street Volume Ratio
A. High Volume

 

V 513f.._}:.._,gggf..§;j.;; I No Preemption I‘
' :2 z >_: .~ -, '- I Preemption, f *7}?
' ‘ 1:57;” DCompensatIon "

    
  

 

 

 

 

 

Mean Travel Tlme

 

 

 

 

 

 

1500:500 1500:300
Main to Cross Street Volume Ratio

B. Medium Volume

1 500:750

 

I I No Preemption ff!
-. A. I Preemption
-’ DCompensation

 

73150 ., .
3 .. ‘

 

 

 

 

 

d
N
0"

 

.5
o
O

\l
01

 

 

 

 

 

 

 

Mean Travel Time (Sec/B

1000:500 . 1000:333 1000:200
Mam to Cross Street Volume Ratio

C. Low Volume

Figure 14: Bus Route 2 Mean Travel Time

105

has no impact on bus statistics.

The bus average travel time and the average total bus-link delays under the compensating
plan were higher than with no compensation for high traffic volume (main street volume
larger than 1500), and low volume ratio (2:1 and 3:1). Using the compensation plan cuts
some of the main street green time (the direction where the bus runs). As a result, main
street high volume traffic became congested and long queues are formed. As the next bus
arrives at the intersection, it would experience a delay due to the delayed traffic and may
have to join the queue that was made longer by compensation. Since green phase timing is
proportional to volume, low volume ratio would have a considerable cross street green
time. When compensating, the main street green signal may be cut as much as the cross
street green phase (if the phase was skipped). Compensation for the low volume ratio

cases may increase delay for the main street traffic, and thus for the bus.

6.2.2 Before and After Analysis

For every preemption that took place, four cycles before preemption and four cycle after
(including the preemption cycle) were considered. However, since bus headway is five
minutes (random) and the cycle length is one minute, there was a good possibility of over-
lapping between two successive preemption before and after statistics. Therefore, to sepa-
rate the effect of every preemption, only preemptions with no overlap were studied. As a
result, four preemptions were selected for the analysis; two green extension / red trunca-
tion at around 5:13 and 5:24, and two skip phases at around 5:35 and 5:45. Overall vehicle
statistics were calculated for the periods before and after preemption for each of the above
four cases. For cases of 5:1 volume ratios the only possible preemption plan was skip

phase, since cross street green time was already at the minimum (10 seconds).

106

Figures 15, 16, and 17 show a comparison between before and after overall intersection
delay for the two preemption cases of each of the green extension / red truncation, the skip
phase with no compensation, and the skip phase with compensation preemptions, respec-
tively. Preemption #1 and preemption #2 in the graphs refer to the first and the second pre-

emption of each. The detailed data is provided in Table F.6, Appendix F.

Figure 15 shows conflicting results between preemption #1 and #2. While the first shows
that under the green extension / red truncation preemption policy, overall intersection
delay increases with preemption, the second shows the opposite. Low traffic volume cases
(L2zl and L3:1) were exceptions to the first preemption, since cross street volume is also
low and thus, preemption might not increase delay. Low main street traffic volume would
have a greater chance of clearing the intersection within the predetermined green signal,
and thus extending the green time to facilitate the bus passage through the intersection
may not benefit vehicles, other than the bus. At low volumes, cross street right turning
traffic may have enough gaps to turn on red and thus reduce excessive cross street delays.
Therefore, preemption might be beneficial when employed at low volume intersections.
Also, the upper 3:1 case was an exception to the second preemption. This might have been

purely due to random traffic fluctuation. However, the impact was small.

Since two preemption cases of the same type gave two completely conflictin g results, the
effect of green extension /red truncation on overall intersection delay was inconclusive.
The explanation might be that it is a function of traffic arrival randomness. The first green
extension / red truncation preemption took place at the beginning of the simulation (at

5:13) where there was minimum disturbance due to any other preemption, while the

107

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4 ‘ i .’ ..

 

 

 

 

120

100

l l l

O O O O 0
co (0 V N

(us/V939) llama afieJeAv

l38'l

l33'|

LIEW

LIZW

L380

tran Z#3Jd

l38'l

L331

LISW

lIZIN

I-38n

Lian l-#91d

Main Street to Cross Street Volume Ratio

ion.

Average Delay For Advance Green / Green Extension Preempt

Figure 15

0
co

II'

§\

gifil‘f'fi

(O

O
V

108

\)¢ I].

1.1““;

IIIIIII IIII IIIIIIIIIIIII! IIIII

O
N

(games) [also eBeJaAV

1‘).

 

IIII'II IIIIIIIIII

7‘. .
:9;

Q

0

L391

1391

men

men Z#91d

l39‘l
£381
L331
139W
LISIN
lIZW

139”

men |-#91d

Main Street to Cross Street Volume Ratio

Average Delay For Skip Phase and No Compensation

Figure 16

1

 

 

 

 

«I»,

 

 

 

. . .. .c.‘. w -. ~ .

 

 

 

 

 

_ , » . :~‘ . »/. _ . A. §. _.
~ ; . >_ I .

 

 

 

 

 

 

. .. _ .. .\-,,' . ._ , , , ,. , ,.,. ._- .\ "

 

names) Aeiea OBBJOAV

         
    

:91
:81
:21
29w
:ew
:zw
:9n
zen
2m were
:91

:81

39W
58W
38W
39“
38f]

380 L#91d

Main Street to Cross Street Volume Ratio

IOI‘I

With Compensat

ion

Phase Preempt

Ip

Average Delays For Sk'

Figure 17

110

second one happened later in the simulation (at 5:24) after other disturbances may have

occurred.

Figures 16 and 17 show that skip phase preemption, with and without compensation, was
not beneficial to the overall intersection statistics. Delay generally increased with that type
of preemption. However, skip phase preemption was more successful for high volume
ratio cases (5:1 ratios). Compensation was not a decisive factor for low volume cases.
However, it influenced intersection delays negatively at high volume, low ratio cases

(U 2:1 and U3:1). Compensation either increased the disbenefit or reduced the benefits.

6.3 BPS and Carpools

In an effort to reduce the usage of single occupancy automobiles and encourage drivers to
switch to multiple occupancy vehicles, a unique idea was proposed for use in the imple-
mentation stage of this project; providing signal preemption service to carpools. The car-
pool choice could be more attractive for people than buses, if the necessary incentives
were provided. One of these incentives is carpool signal preemption. However, loading
the network with so many carpools that signal preemption would be called every cycle
would be a great disturbance to traffic flow. Therefore, the effects of carpool preemption
as a function of the percentage of carpool users in the network was tested, using the case 5

preemption plan (selective preemptions).

Cases of no carpool, 5% carpools, and 10% carpools with and without preemption were
simulated. The effect of the presence of carpools in the network and carpool signal pre-

emption on bus trip delay were also tested. Carpools were introduced only at the east and

111

west ends of the corridor. NETSIM assumes an average private vehicle occupancy of 1.3

persons per vehicle and an average carpool occupancy of 3.5 persons per vehicle. In order
to maintain the same number of users along the main corridor, the appropriate conversion
factors were used and the main corridor traffic volume was adjusted accordingly. Thus, the

higher the carpool percentage the lower the network traffic volume.

Figures 18 and 19 show the effect of 5% and 10% carpools on the system, respectively. As
the percentage of carpoolers in the system increased the average vehicle and person delays
decreased (without preemption), because network traffic volume was reduced. A maxi-
mum of only ZO-minutes of simulation time was achieved, since there was a preemption
call at almost every minute and the maximum time periods allowed by NETSIM is nine-
teen. When 5% of the main street traffic were carpoolers using preemption, there was an
insignificant effect on vehicle and person delays, although absolute network delays were
significantly reduced. However, adding more carpools with preemption into the network
(10%) increased the overall vehicle and person delay rapidly, because there was a carpool
calling for preemption almost every cycle at every intersection, and thus traffic optimiza-
tion and progression were greatly disrupted. Despite that, buses generally continued to
gain benefits from the frequent preemption calls by buses and carpools, since they almost
always found either a green signal or an already placed preemption call before they
arrived at the intersection. These results are shown in Figures 20, 21, and 22. Detailed

results are shown in Appendix G.

112

 

 

 

I No Preemption
2-50 . __ ~ _ , , _ IPreemption

 

 

 

 

Delay (MinNeh-Trip)
a:
O

 

 

 

 

 

 

 

 

 

 

1.00 -
0.50 —
0.00
0.00% 5.00% 1 0.00%
Carpool Percentage
Figure 18: Cumulative Network Delay
,5, 45 INo Preemption
3 40 — IPreemption ’
% 35 — 9 * .
‘3, 30 _ - - "A T
g 20- - -

0.00% 5.00% 1 0.00%

Carpool Percentage

Figure 19: Average Person Statistics

Mean Travel Time (Sec/Bus)

113

 

1350 _ No Preemption

 

 

 

Mean Travel Time (Sec/Bus)

 

 

0.00% 5.00% 1 0.00%

Carpool Percentage

Figure 20: Bus Route 1 Mean Travel Time

 

 

 

 

No Preemption 5*"
I Preemption

  

 

 

  

 

 

 

0.00% 5.00% 1 0.00%

Carpool Percentage

Figure 21: Bus Route 2 Mean Travel Time

114

 

90 I No Preemption

I Preemption

 

 

 

80

l

 

 

 

70-

60-

50~

40—

Average Delay (Sec/Bus)

 

30-

0.00% 5.00% 1 0.00%

Carpool Percentage

Figure 22: Total Links Bus Delay

Thus, a certain percentage of carpoolers in the system might be beneficial, since it reduces
the number of vehicles on the streets and reduces network delay. But, a high percentage of

carpools, such that a preemption call is made every cycle is non beneficial.

6.4 Test of Random Vehicle Generation

It was clear that in many instances vehicle generation at the entry nodes, vehicle arrivals,
driver’s behavior (cautious, normal, reckless), and turning movements, which were all
randomly assigned by NETSIM, played a significant role in the network measures of

effectiveness. The network characteristics are randomly selected based on a random

115

number seed coded into NETSIM. The model’s default number seed was used in the pre-

vious simulation runs. However, to test the effect of randomness on network MOEs, with

and without preemption, a different number seed was selected. The network was first sim-

ulated without preemption and then the case 5 preemption plan was used. The results of

these simulation runs are presented in Tables 23, 24, 25, and 26. A 45-minute simulation

period was reached.

Table 23: Cumulative Network Statistics; With and Without Preemption for a Differ-
ent Random Number Seed.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Delay Delay
veh'Tnps Veh-Hours Min/Veh-Trips
_ =L __ =
Default No Preemption 3469 163.43 2.83
Seed No. Preemption 3424 167.83 2.94
= It i if

Second #No Preemption 3446 147.69 2.57
Seed No. I Preemption 3431 149.38 2.61

Table 24: Cumulative NETSIM Person Measures of Effectiveness; Before and After
Preemption for a Different Random Number Seed.

 

 

   
 

 

Travel Time
Person Person (Person-
Tnps Mrle Min)

#

14921 5255 20264

    
       
 
 
 

Avge. Delay
Sec / Person

 

No Preemption I
Preemption I

 
 

Second

14974
14131

20624

   

   

 

Seed No.

 

Default
#— __'"'
--

   

14248

 

 

 

 

 

 

 

116

Table 25: Cumulative Network-Wide Bus Statistics; With and Without Preemption
for a Different Random Number Seed.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

II Total Mean Person

Route Bus Travel— Travel- Person Travel-

Trips Time Time Trips Time

(Bus-Min) (Sec/B us) (Min)

II

Default No 1 2 68.1 1361.6 50 1702.9

Seed Preemption 2 3 60.7 1214.2 75 1518.3
=r—-——_:==

Number With 1 2 69.9 1301.4 75 1610.8

Preemption 2 3 59.3 1184.0 75 1481.3

Second No 1 2 64.6 1481.0 50 1613.8

Seed Preemption 2 3 62.4 1191.0 75 1559.6

Number II With 1 2 62.1 1429.5 50 1553.3

ll Preemption 2 3 61.1 1164.5 75 1527.1

 

 

 

Table 26: Cumulative NETSIM Bus Statistics; With and Without Preemption for a
Different Random Number Seed.

 

 

 

 

 

W
Link-Trips (Min) (Min) Sec / B-Trip
Default fi"I No Preemption i 65 127.4 86.2 79.6
Seed No. 11 Preemption II 67 129.3 86.6 77.6
Second No Preemption 64 127.0 87.2 81.8
Seed No. Preemption 64 123.3 83.3 78.1

 

 

 

 

 

 

 

 

 

 

 

117

As shown in the tables, different random numbers generated a difference in statistics for
the reference case (no preemption) that ranged from around 2% to 10%. However, the
benefit or the disbenefit from preemption was less than 5%. Furthermore, in comparing the
effect of randomness on the change from no preemption to preemption, Table 23 shows
that vehicle delay was 3.9% worse using the default random number and 1.6% worse
using another random number; a difference of 2.3%. Person MOEs (Table 24) show that
preemption made a difference averaging from a 2.5% increase using the first random num-
ber to a 0.2% decrease using the second random number, a difference of 2.7%. These
results indicate that the effect of preemption on the vehicle and person delays found in this

study may fall within the normal traffic fluctuation.

As far as bus statistics are concerned, bus route 1 travel time was reduced by 4.4% using
the default random number, and by 3.5% using another random number; a difference of
0.9%. Also, bus route 2 travel time dropped by 3.5% and 2.2%, respectively; a difference
of 1.3%. Total bus-link-trips varied from 2.5% to 4.5% reduction in delay for the default
random number and the second random number, respectively; a difference of 2.0%.
Although, the second case reduced delay more than the first case, the bus did not travel a
greater distance (in terms of total bus link-trips) within the simulation period. Even con-
sidering random variations, the bus still gains some benefits from preemption, although it

might not be very significant.

The preemption tests studied in this research under different traffic conditions and using
different preemption plans resulted in small changes in the network statistics (in terms of

vehicle delay, person delay, bus delay, and bus travel time). Most did not exceed 5%. The

118

randomness test showed that some of the network statistics varied more between two sim-
ulation runs using a different random number seed than they did between preemption and
non preemption. Although randomness influenced these results, not all the changes
described were the result of randomness. The changes in delays and vehicle trips associ-
ated with different preemption plans and under different traffic conditions (discussed in

this research, chapters 6 and 7) were the result of the selected preemption strategies.

Chapter 8

Conclusions

The literature reviewed failed to identify an up-to-date model that can simulate various
BPS strategies, that is comprehensive, and is capable of restoring the original signal set-
tings. These shortcomings are cited as reasons for the lack of implementation of BPS in
the US. However, the use of NETSIM’S graphical animation capability provided the flex-
ibility to test several BPS plans and to restore the optimal signal settings after preemption

is granted.

Based on the results of this study, it is clear that BPS provides little benefit to the corridor,
with the volume and bus frequency (15 minutes) characteristics of Washtenaw Avenue.
Optimization of the network traffic signals and progression provide the least delay in the
network. Preemption, which deviates from the optimum setting, created an increase in
vehicle and person travel time and delay. As the frequency of preemptions increased,
delay increased in the network. The green extension / red truncation plan resulted in less
vehicular delay than the skip phase plan, since the later provides more disturbance to pro-

gression.

The maximum benefit that a bus gained under any tested condition averaged 80 seconds
out of a l380-second travel time (6%). This benefit (80 seconds) for any single bus trip
could be lost or gained if a bus randomly missed or caught a green light at a signalized
intersection that has more than a 60-second cycle length; e.g. Golfside, Huron Parkway, or

Carpenter Streets. The overall benefits gained by buses from preemption were not

119

120
sufficient to counter the delay to other vehicles in the traffic stream. Therefore, when con-
sidering the costs as a result of preemption, the BPS process was not beneficial overall. To

justify the provision of BPS, bus headways would have to be less than 15 minutes.

The best BPS is the one that combines various treatments for different intersection (case
5), see Table 27. The green extension /red truncation plan results in the least increase in
delay. The skip phase plan results in a significant increase in delay at intersections with
high cross street volume and low main to cross street volume ratios. Compensation was .
generally inappropriate since the main street volume in the study corridor was relatively
high. The success or failure of a specific BPS plan is, primarily, a function of signal phas-
ing and traffic volume. Thus, the most suitable plan for each intersection in a corridor
should be selected so that the benefits of BPS are maximized. For the study corridor, this
means using green extension /red truncation and skip phase plans at Yost, Sheridan,
Brockman, Austin, Hill, and South University, green extension / red truncation and skip

phase with compensation plans at Pittsfield, and green extension Ired truncation plan at

Golfside intersection).

It was noted that when preemption took place at a highly congested intersection (at satura-
tion), preemption effects continued for several cycles. Sometimes, the effect (delay) accu-
mulated to the end of simulation (link reached over saturation).The presence of a single
heavy volume intersection in the network can skew the overall network statistics. The

weight of these intersection MOEs was very significant in the overall network MOEs.

121

Table 27: Summary of Statistics For Several Preemption Plans.

 

 

   

 

 

 

 

 

 

Avge. Veh. Avge. Person Avge. Bus Avge. Bus
Cases Delay Delay Travel Time Delay
Min/Veh-Trip Sec / Person—Trip Sec / Bus Sec / Bus-Trip
No Preemption l 2.83 48.4 1361.6 79.6
Case 1 | 2.90 49.0 1361.8 79.6
Case 2 | 2.89 49.2 1326.3 78.9
Case 3 | 3.00 49.6 1296.1 79.1
Case 4 I 3.15 52.3 1288.0 76.1
Case 5 | 2.94 49.6 1301.4 77.6
Case 6 | 2.97 50.6 1342.6 82.9

 

 

 

 

 

 

By testing the sensitivity of BPS to volume change, it was found that the effects of BPS on
delay to the general vehicular traffic were not significant at very high and very low vol-
umes. Bus travel time and delay decreased with a decrease in volume up to a certain low
point and then leveled off. In general, the 3:1 main to cross street volume ratio is the bor—
der line, above which person and vehicle statistics favor preemption with no compensa-
tion, and below which preemption might not be favorable and if it is provided,
compensation is warranted. Providing preemption for intersections with 3:1 or higher
ratios, and cross street volume of 500 VPH or less did not generally result in an increase in
delay to the general traffic. Preemption increased delay for volume ratios less than 3: 1 (see

Table 28).

Testing the before and after MOEs at an isolated intersection showed that the green exten-

sion Ired truncation preemption plan was inconclusive; beneficial at one time and non

122

Table 28: Summary of The Overall Statistics For The Volume Sensitivity Test

   
 

  

 

  
 

 

 

 

  

 

  
 

 

 

 

 

 

 

  
 

 

 

 

V1223}: 1:22; $3101; AvlthZILeh. Av g1; ::son Avgziafjlus

Min/Veh-Trip Sec / Person-Trip Sec / Bus—Trip

1500: 750 No Preemption I 0.53 = 25.5 — fi
(2:1) Preemp, No Comp. I 0.79 40.0 38.8
Preemp. W/ Comp. I 0.65 32.3 39.2
. No Preemption I 0.37 fl=FT== 43.1
Preemp., No Comp. 0.39 17.6 37.8
Preemp. W / Comp. 0.38 17.1 38.1
MI No Preemption I 0.30 13.0 40.2
I Preemp., No Comp. I 0.31 13.5 37.1
I Preemp. w7 Comp. I 0.32 13.7 __ 37.4

 

 

 

 

beneficial at another. The results were dependent on the vehicle arrival pattern. However.
skip phase preemptions (with and without compensation) were beneficial at a high volume
ratio (main street volume: cross street volume = 5:1). Compensation was not a decisive

factor at low volumes, but resulted in a higher delay at high volume.

There appears to be advantages for providing carpools with preemption capability up to
between 5 and 10% of the main street traffic volume. Carpool services provide benefit to
the network, if they replace some of the private automobiles and thus, reduce main street
volume. When 5% of the main street traffic was replaced with carpoolers with preemption
capability, network vehicular traffic delay was not increased, people’s travel time and
delay were slightly reduced and bus statistics were generally improved. However, when

increasing the number of carpools into the system, the bus travel time and delay continued

123

to benefit, but the network vehicular delay and person delay were significantly increased,

due to the frequent interruption of the optimum signal settings by preemption calls (Table

29).

Table 29: Summary of The Impact of Preemption on Carpools.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Avge. Veh. Avge. Person Avge. Bus Avge. Bus
Case Delay Delay Travel Time Delay
Min/Veh-Trip Sec / Person-Trip Sec / Bus Sec / Bus-Trip
Base Case I 2.39 40.8 12.084 82.1
W/ Preemption I 2.39 40.6 1200.9 81.6
5% Carpools 2.26 38.1 1136.0 78.8
W/ Preemption 2.27 38.0 1113.0 78.9 I
10% Carpool II 2.13 35.9 1110.7 78.6 I
W/ Preemption 2.42 39.8 1114.0 76.8

 

 

 

 

 

 

 

In any corridor there is likely to be random fluctuation in the traffic demand, and this vari-

ation may be as large as the measured effect of BPS. Although NETSIM’s time period

specific statistics provided a microsc0pic picture of what happened before, during and

after preemption, the effect of vehicle arrival pattern was significant and may mask some

of the preemption effects. Testing a different random number seed showed that most of the

changes in network statistics and the effects of BPS found in this study corridor were

within the range of variations resulted from merely changing the random number seed

(Table 30).

The primary recommendation for the Ann Arbor Transit Authority is that the provision of

BPS for buses only under the current conditions is not worth the costs. However,

124

providing limited carpools with preemption capabilities should be tested, this may provide

lower overall delay.

Table 30: Comparison Between The Results of a Different Random Seed Number
and Case 5 Preemption ‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Avge. Veh. 11:13; Avge. Bus Avge. Bus
Random Case Delay Dela Travel Delay

Number Min/Veh- Sec 7 Time Sec / Bus—

Trrp Person-Trip Sec / Bus Trrp
. . l

Default No Preemption I 2.83 48.4 1361.6 79.6
Number Seed Preemption I 2.94 49.6 1301.4 77.6
Second No preemption I 2.57 463 1481.0 81.8
Number Seed Preemption I 2.61 46.1 1429.5 78. 1

 

   

For further research, it is recommend that a model be developed that has the capability of
automatic bus detection and the flexibility of changing signals automatically, according to
the preemption plan, instead of the visual detection using graphical animation. The
NETSIM and THOREAU models have good potential for such an enhancement. The a1 go-
rithm for both models has been developed in this research, Appendix B. The application of
these algorithm will be of greater benefit if the model possesses the capability to select the
most appropriate preemption plan for every intersection and to optimize the network sig-

nal timing plan after each preemption.

Appendix A

Network Peak Hourly 'D'affic Volume

Appendix A 125

Table A1: PM. Peak Hourly Volume Along Washtenaw Avenue

 

J -

 

 

      
   
   
   

Intersection Direction Left Through Right
Intersection 11 East Bound 359 1445

 

West Bound 107 947 196
North Bound 151 100 314
South Bound 344 341

 

 

   

Intersection 1 East Bound
West Bound
North Bound
South Bound

 

 

 

 

 

Intersection 2 East Bound

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

West Bound
North Bound 0 0 135
South Bound 0 0 210
49
West Bound 92 1419 0
North Bound 106 0 117 I
South Bound 390 22 76
Intersection 4 East Bound 147 1645 52
West Bound 266 1167 168
North Bound 134 371 248 I
South Bound 248 552 197 II
Intersection 5 Eas—TBound i 38 1709 102 I
West Bound 104 1370 24 1
North Bound 42 32 95
South Bound 40 23 33

 

 

    

Intersection 6 East Bound
West Bound 774 0 665
North Bound 0 0 810

 

 

 

 

 

 

Appendix A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

126
Intersection Direction Left Through Right
Interse=ction 7 East BoTTd *0 1031 97
West Bound 19 646 0
North Bound 44 0 26 I
Intersection 8 W42 I102 -f 37
West Bound 16 651 23
North Bound 18 25 12
South Bound 14 41 ll
Intersection 9 East Bound 0 1250 70
West Bound 0 660 27
North Bound 107 137 136
South Bound 28 94 27
Intersection 10 East Bound 12 537 21 N
West Bound 130 406 1 1
North Bound 26 66 76
South Bound 22 53 11

 

 

 

 

 

 

 

 

 

 

Appendix B

BPS Algorithm

(Flow-Charts)

Appendix B 127

Has Preemption Occurred

uring One of the Past N Actuatio Yes No Preemption
r During The Last N Seconds? This Cycle L.@

 

 

 

 

 

No
Detector 1 NO Detector 2 N O
Actuated? Actuated?
I Yes
Recordo'Iime 0 Yes Record Time of
Actuation, t1 -—> Actuation, t 2

N0 sThis Conditizrk Yes
eemption? /

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

t . No
s Bus Behrn> - No Preemption _..@
Schedule? This Cycle

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I Yes
Con gestionTime No
Limit between Detectors
1 & 2 Elapsed /
Yes Yes Detector 2
N0 Detector 2 Actuated?
Actuated? N 0
I Yes
. Contin e O tside The
BUS 15 Record Time Ofl‘—__ Progrgm u
Stuck. Actuation, t2
No Pre-
emptron I
Subtract l-Actuation
From Detector 1

 

 

 

Figure B-l: Far-Side Bus-Stop BPS Algorithm

Appendix B

128

 

 

tor 2 to Signal (t2+X)

Calculate Time From Detec-

 

 

 

 

  
   
 
 
  
       
 

  

<Bus Arrives N00
On RED

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

   

 

 

 

 

I

 

 

Subtract One

Actuation from
Detector 2

 

 

Figure B-l: Continued.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Subtract One
Actuation from
Detector 2
15 mm No Return to
~b- Non Preemp. N0 Detector 4 Yes
Control Actuated?
ax. Green Record
Is Bus Too Reached “me Of
Late For N0 Actuation, ‘4.
Extensron ' IYes
Extend No B t .
Complete Min I'SCC' x ensron
gross Street I Subtract One
Extend Green reen Actuation
tto Scar Bus or * From
0 axrrnum
. Advance for Detector 4
Extensron XX Seconds
or into the I
Green

Appendix B 129

Is Time
Available .for E3 gfc‘gflp‘t" Non e
Preemptron Control

I Yes

Has Cross Stree
or Other Phase)
reen Be un

No

 

 

 

 

 

 

 

     

 

  

Yes

    

 
  

Complete Minimum Green Then
Advance For Mrnrmum of
Seconds, (10) seconds After
Detector 2 Actuated or mto
Green

  

 

 

 

 

Extend Green to X Seconds
After Detector 2 Actuatron

I

Subtract One Actuation

 

 

 

 

 

 

 

 

 

 

From Location 1 <
N0 Detector 4 Yes
Actuated?

 

 

1

Record

ax. Green Trme of
N If Reached Actuation
0 t4 .
JYes I

153885“ [No Extension]

 

 

 

 

 

 

 

 

 

 

 

 

Subtract One
a Actuatron
0 From
Detector 4

Figure B-l: Continued.

 

 

 

130

Appendix B

 

635 2:. co Em
E. swap—E. 635800

 

 

 

+

 

SEC. 520

 

 

0&9 mwconoom PU Ea

 

+

 

descend mm
82m 3389:. BE.
:25 853me 8.25

2F as, 886$

 

 

 

+

H0 .30
SE... 580 “ovum
$ch 033025

+

N sin

y

 

 

 

 

 

 

 

 

 

 

230
0 cam
: :85.
oocm>v<
0::—:50

 

 

 

 

 

 

 

 

 

@893 850 2: c8 23396 ”N San
8:268 6223 58:88 £25538 02 L :«E
”mam—a wage—E 2: we 25 88:0

 

 

230

0 new

z :85.
02650

~

 

 

 

 

08E. 52.0
oft 9 m5 BE

 

 

 

»

 

826
$80 2:. 9
£96 enuncou

 

BEE—80 "Tm 9.sz

 

 

 

.

 

m0 fleecexm
68C. condo “ovum
:32 833:3

 

 

2&0
o “Em
; 29:3.
6:22.80

 

 

 

+

N 55

v

 

 

 

 

+

 

 

_ :wE

 

 

 

+

 

‘

 

@893 550 05 5» 83:33 ”N ENE
8533 893 58:88 625635 02 ; 5E
”mama 336:8 2: mo 25 88:0

 

oZ

\Bccofim coom/
/=ooc0 mam \

8%

 

131

Appendix B

55:82 8: new as 62882 as. 2:3...—

mu £28323
00 Ban. Becca

 

 

 

 

 

 

 

 

 

8% v
Efiwoi ‘IA $2833....
of. 222:0 2:550 02 m 888qu

 

a;

 

@T _ cocoouon Eoi
c2823?— 825.6

 

 

 

2

363833..
We 08:. 908%

8% +

seoeaoa.

 

 

 

 

 

. 228 /
A NOWOHQOHQM— \

+

8% 8%
8m «E N 6% H

808200 :833 :85
/ 2:5- cosmowcov

8%_ ?

   

      

oZ

 

 

:2an
8E oz

.285
2 8m

 

bHOu Ovuoa 0 Z

 

 

 

2&0
25. comanoi oz

 

  

 

 

 

 

 

@ll. .6 2:5. Eooom

 

 

 

 

9 628234

 

 

 

 

 

 

 

 

89+

moivonom
oz 5 o m: m
$25888;
@ 8% 3:25:80 25. 3 Z
1.
NH .comumsueos S .cOflwaoa‘
mo 8:: Booed mo 22E. 908m
8% 8% b

 

 

@8822 V _ A 883334 I A pecans/V.

 

@tzlA
@1

 

2&0
£5. :oumEooE oZ

 

 

62+

wmncooom Z 28: 2F watsfl co
8% 83833. 2 28m 00 8:0 watso

@8580 52588.5 83

 

QIIQIL

 

Appendix B 132

t (9

Calculate Time From Detectors 2
to Signal (mX)

111 The Bus Subtract One No
Arrive At The Inter— OActuation From Rgfigttgaw
section at YRED Detector 2
Yes
Detector 3 Yes Record Time of
Actuated? >_’ Actuation, t3

‘”°'@

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

  
 

Is Time
Available for
Preemption

 

No

 
   
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Is Bus too Yes Is Bus Too No
Late For Early for
Extension Advance
No Yes .
Complete Mrn
Cross Street
Extend Green Green
to Clear Bus or I
Exxfigilglum Advance for
XX Seconds“
I or into the
Green
Subtract One
Actuation from

 

 

 

 

DCICCIOT 2 Subtract One
I Actuation from ’CS)

Detector 2
N o

 

 

 

 

Detector 3
Actuated?

Yes

 

 

(5

Figure B-2: Continued.

 

don—£250 "Ntm 95w?—

 

 

 

 

 

820 2:. 95 :O

0 8:83 XX 8m

:80 8:83.

:85

meeoowm 82: 6:5

-N o... D C8 802m 820

Goa-U “005% 3:2 vac—@800
:32 w:2xm .

 

 

 

 

 

 

 

 

 

 

 

 

m :98qu

m 88280 88m 8883.

88m 8883. 0:0 88:5
0:0 885%

 

 

 

 

 

» >

mgooomN 820

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

4 A8533 38m “0:266 8:08mN
H 8% 8m 9.22m
::wom 820
B 02 38:: 850 :8 :282xm
t 02% 880 8: oz 820 8%
a K mgr—RMO/
oz /:o:ano.£ m_\ 8%
8% a
N :98qu 032824 08E.
@ll EB: 8832‘ 83885 :0 8:83
2:0 Sunbeam oz Aim 22:1 2%

 

 

 

Appendix B

 

134

Appendix B

 

 

_ 88080 88m

8833: 0:0 :8b8m

 

 

a

 

 

:00:0 95 :0
80883:: N 8:00:09 :05: 888m

%X :o 8:80m XX a :0”: 0088<
:20. 80:0 82552 083800

 

 

 

8880
0::: :oZ

 

9 580m

8A

 

 

@

 

8%

80830.:
m .8826

8%

OZ

88:80 "NA: 0::wE

 

35:00 88:8
-05 :02 9 88¢

 

 

 

oz»

9|A2§§< 05:. V
8% 888005 E

 

>

 

 

 

: 8:00:09 80:”:
8830< 0:0 888m

 

 

 

  
  

  

8: 28:52

08:. fl

»

8888

8:88:02 3::

.5 288:3: 0:0.

88:88 8830’: N
80.5 :0:.:< 8:80m
NX :0": 80:0 8:0:xm

 

 

 

 

Appendix B 135

 

Has Green No

Yes
I Been Extended

Choose one of the following plans: a

Plan 1: No Substitution, maintain phase sequence
Plan 2: Substitute for the other phase(s)

l i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Plan 1
‘ Calculate Mafii
Y , Street Green Tlme
Contmue Extended, GE
Through
End of C t1 C I
C cl e on nue yc e
y I to The Cross
, Street
Add GE to The
Green Time
Continue
Through
End of
Cycle

 

 

 

Figure B-2: Continued.

Appendix B

No

 

D -tector
Actuated

 

Plan 1

 

 

 

 

136

 

 

f

Choose one of the following plans:
Plan 1: No substitution, maintain phase sequence
Plan 2: Substitute for the other phase(s)

 

 

 

ctuations A

@etectors 1, 2
or 3

 

 

      

Yes

 

Subtract One
Actuation
From Detector

 

 

 

4

Continue
Advance
Throth

End 0 Cycle

 

 

 

 

 

Yes

 

 

Return to
Cross Street

Green

 

 

 

Figure B-2: Continued.

r

Plan 2

 

 

 

 

 

Calculate Cross
Street Green Time
Cut, CT

 

 

 

 

Proceed With The
Phase Sequence Until
The Truncated Phase
is Reached.

 

 

 

 

Add CT Seconds to the

 

 

Green Time.

 

l

 

 

Continue Throu h The
End of The Cyc e

 

 

é

Appendix C

Case 1 Preemption Results

137

Appendix C

Qw—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.0: 8:023:88 00:8
0.00 0. : 0 202.03: 00:8
0: 0:: 005-005
v.0: m.w Nd: _.: Qw E0>\o0mv QED
0.: 0.0 0.0 0.0 0.: €2-08 00:0: 0:58 0:02 0-00
:: 0: l 0: 0: 0: 005-0%
0.00 0.00 0.00 0.00 2.: 20280: 00:0:
_ :0: 0.0 0.0 :.0 0.0 2:20:03 00:00: 00:00 0000: 0-:
:
m: m E Z 0 0Q::.:.-:_0>
:38. :88. :98. :88. :88.
0.0 0.0 0.0 0.0: 0.: E 0.0: 0.0: 0.0 0.0 202800 00:0:
0.0 0.: 0.: 0.:. 0.: 0.: 0.0 0.0 0.0 0.: 0.3.53 00:0: 0550 0003 0-0
0 0: 0 0: 0: :: 0 0: 0: 0: 005.000;
N: .: 0 : m : N: : N: :
Sumuh : u mfimuh : u H n 0883000
000800-00 wwwomow? 000800-00 0w??? 00:00:00w0- %0 00m 00:82:: V:0:
88m 08:. 5m 8.8m 08:. 5v 88m 0:5. Em 88n— 0EF EN 88: 0:::L. :0: mo 00:308.):

 

 

 

8:80. 08”.— 0: 838:5 8: .88; 05E. 83,—. 2:. 0: 88:0:xm 83:0.— ée. a? 88%
E02 8:0 :2 000:5 08:0 £8:— ::0 888%.: \ mm. 8.5m 5:2 .5...— :00:0 8:080: :0 8::
68.3 .0 80:52 8:80:00:— ._e.m 8:982.— 3:581 : 8 838.05% 3.0 030,—.

 

 

138

Appendix C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.0 0.0 3:00:88: 00:8
0. :: 0.0 202-003 00:8
00 00 00:00.08
_ 0.: 0.0: 0.00 n SH 0.0 eczowmvsg
_ 0.0 0.0 0.0 :.0 0.0 93-0.5038 0.500 .052 0:00
0 0 N N 0 005-09
0.00 0.0: 0.00 3.0 0.00 202800 E8
_ 0.0 0.0 0.: 0.0 0.0 2:2-00>0§8 0.500 580 0:-:0
: I: 0 I: : aEfi II
0.0 u 0.0 5. H40 0.0 0:0va :8 .l.
_ 0.0 0.0 0.0 0.0 2 9:209:28 0:80.000: 0:-00
. 0 0 :: 0 0 005:;
_ 0.0 0.:: 0.0 0.0 3.: 20280058
_ 0.: Z 0.: 0.: 0.0 3:20.858 05:00:83 0:-0
0: E 0: :: 0: 005-005
_‘ 0000"? 03¢ :00"? 00:3. 0008
80800.00 000800-00 000800-00 000800-00 000800-00 08020800: .5825 05.:
Ston— notom totem cocon— vorom 00 00.50022
20:. :00 20F 5:. 2:: ::0 20:. 000 05F a:

 

 

 

 

 

 

 

 

 

.2500 .000: 0: 00:08.5 0:0 630502.: 0: 00:00.:
25.0 0.:—z. 0008000 0.: .5 0: :0.20 .850 820 \ 00580-0 .50 805.00: 0: :85 05.0.: 08.00
0:0:: .0: .2502 02888.0: .00 000058.... :0 :33: a 00 80003033 ad 030,—.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

~00 0.3 3_o_:_o>\oomv .0200
0.:.00 0.:00 202.09: 00:8
0::. 000 00.5.09
0.00 0.0:. 0.00 0.00 0.00 0.00 0.0:. 0.00 0.00 0.0 0.0:. 0.0:. 0.00 0.00 0.00 202800028
0.: E 0.0 0.0 0.0: 0.0 0.0 0.0 :.:. 0.0 0.0: :.:. 0.0 :.0 :.0 2:208:28 .0550: 0:02 :-0:
: 0: 0: 0 :: 0 : :: v 0 :.: 0 0 0 0 005-08
:..:.0: 0.00 0.00 0.0:.: 0.0:. 0.0:. 0.00: 0.00 0,00 0.00: 0.:.0 0.00 0.00: 0.00 0.00 202890200
:.00 0.0 0.0: 0.00 0.0 0.0 0.00 0.00 :.:.: 0.0: 0.00 :.0 0.0: 0.:: 0.:: 2:208:28 .0080: 5:00 :-:
0: 0 :: 0 0: 0 0 :.: :.: 0 0: 0 0 0: 0: 005-005
P 1 hr
0.:. 0.0:. 0.0:0fi000 :.00 0.:.00 0.00 0.0:. :.000 0.0:. :.0:. 0.:.00 0.00 0.0:. 0.0:.0 202890300
m 0.0 :.00 0.00 00.: 0.:0 :.: 0.0 0.0: 0.00 0.: :.0: 0.0:. 0.0 0.00 0.00 93-003028 000002000: :-:
0 00 0: :. :.0 :.: 0 00 0 0 00 0: 0 00 :.: 00.5-08
0.00: 0.:.0 0.0 0.8: 0.00 :..:0 0.0:: 0.0:: 0.00 0.00: 0.00: 0.00: 0.00: :.00: 0.00 Eo>\8m;0_on
0.0: 0.00 0.0 :..:.: :..:.0 :.0 :..:0 5.0 :.0 0.0: 0.:.0 0.: 0.0: 0.: :.: 2:209:28 000002003 :-00
fi:. 00 0 0 00 0 :: 00 0 0 00 : :. 00 0 @529
me:mkqmeqmp_:me:
:00u,0 00:0".0 00:0".0 00:0",0 0000...?
00:88-3: 00:88-00: 00080060: 00:80:06.0: 000800-00: 00253885 02825 0:::
00.000 00.00:: 00.50 00:00 00:8 :0 005002):
08:. :EE 05:. 6:5“: 0:5. BEL. 08:. 958m 0EF 0::":
C
.m
m .0550 :08 0: 00:62:: 0:0 00:952.: 0: 00:00.: 25.0 00.:. 02080
m. .0: 0.: :5 2.: 05.0. 00.: :20: 32:0 0:0:: : 00880.0: :0: 002.33. 0...: .0 .0 a 05.8 .800 ::0:
A

.:: .3502 50820:... 00.:: 0000502.: :0 0000:: 0 0.: 80005005803553.0383 "0.0 200.0

140

Appendix C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.00 0.0: 8:028:80: 00:8
0.0:. :.00 202-003 00:8
00: ::.: 00:02.;
0.0: 0.:.: 0.0: :..0 0.0 28:80: 0::8
_: 0.0 0.0 0.: 0.: :.0 25.003008 0:58 0:02 0-:.0
m: c: o N: :.: 00::.:.:.-::o>
= H l!“ H
0.0:. 0.00 0.0:. 0.00 0.00 2880: 00:8
0.0 0.0 0.:.: :.0: 0.0 202.09: 0::8 00:50: ::0: 0-0
_ N: m ON N: c 00::.:.:.-::o>
"I H
:98. :88. :88. :88. 10:8.
0.0 :..0: 0.0 0.0: 0.0 0.0 :.0 0.0: :..0 0.0 £580: 00:8
0.0 0.0 0.0 :..0 0.: 0.0 :.: :.:. 0.0 0.0 2:208:28 0.58 :00; 0-0
:. :. 0: 0: 0 0 :: 0: 0 0 000508
= 0 .: N: .: m .: 0: .: 0: .:
mmumnh. H u :mumuh H n H n 0026380 m:
000800-00 0mmom00.W0 00880-00 mwmomom? 00:0:MUW0-%0 .0 8:825 0:0::
8.50: 06:. 50 cocoa: 0:0:. 5:. 00:80: 08E. Em 8:30: oEF ::cm 8:000: 080:. :0: .:o 02:03:):

 

 

 

 

 

 

 

 

 

 

 

655m :03: 0: 5:625 0:5 axiom 0E5. BE... 2:. 0: ::0—Evin:— uotom :..0: 55 39:8
::0—2 ::0 :2 39:8 0020 59:...— :::U 858.3 \ mm .028 ::0—2 .5..— 520 8:33.. .3 0::—:.:
.8.3 .0: :85: Z 5:80:85 .5...— =o=aE~2m 0.: 2:8: :0 0:: 09505383? :..0 030,—.

 

 

:.. :0 :..00 6:028:80: 00:8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

141

 

 

 

 

 

 

 

 

:..00: 0.00: £2.09: 00:8
000 000 00:00-09
0.:.0 0.00 :.:0 0.0: 0.0: 20280: 00:8
0.0 0.:. 0.0 0.0 0.0 25.003008 00000 5:50 0-:0
0 0 0 0 0 ackfio
0.00 0.00 0.:0 0.00 0.00 £02800|§8
0.:. 0.:. 0.0: 0.: 0.: 25.5338 0:80 5:02 0-00
0 0 0: :: :: 00:00-08
:.0:. 0.00 0.00 0.00 0.00 202000: 00:8
0.:.0 0.0: 0.0 0.0: 0.0: 2:209:28 0:08:80: 0-:.
00 00 0: 00 00 00000-09
:..0: :..00 0.0: :.00 0.0: 202000: 00:8
0.:. 0.0: 0.0 E 0.0 :=:2-::0>:0:8 0:08:83 0-0
00 00 00 00 00 00:00-09
00:0".0 00:08. 000$ 00008 0000",:
09588-00 09:88-00 09:88-00 09:88-00 09:88-00 000003885 02825 0:0:}:
::oton: totem: totem: totem: notch: 06 0800002 . .
08F :00 08F 0::. 20:. ::0 08:0 0:0 25.0 :0:

 

 

 

 

 

 

 

 

 

 

 

Appendix C

0.50:: :08 0: 00:82:: 00:: .3058: 0: 00:00.:
20:. 0.0.; 00880.0 00 :5 0: 6.2: 32:0 08.6 : 00580-0 .5: 08:23: 0: :85 05.0:
009.8 ::0: .m eon—5:2 5:888:— 30: :.0:.—E8...— .:e :58: a ma 8503:0833 8.0 030,—.

142

Appendix C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:..0: 0.0: 3:029:80: 00:8
:.. :0 0.00 2:20:05: 00:8
0:.: :.:.: 8:00-005
:1. Nd: N.m: wd ON: 20538: 00:05
:_ 0.: N.N Wm w: o.m 3:37:05: 00:00 ncsom 5.52 cém
MH m~ C— : m~ mQtF-r_o>
9mm 9% :.0:. inn 0.3 Eo>\o8: 00:00
0.0 0.0: 0.0: 0.0 0.0 00:20:05: 00:8 0000: 0000: 0-0
3 ON M: O: m 00:00:.-so>
130:. :0qu :58. :0qu ESP
0.0 0.0 0.0 0:0 :.0 0.0 0.0 0.0: 0.0 :.0: 20280: 00:8
0.: 0.: :.0 0.:. 0.0 0.: 0.0 0.0 0.0 0.0 2:20:05: 00:8 00000: 0003 0-0
0: 0: 0: 0: 0: 0: 0 0 0: 0: 0000-005
0 .: 0: .: 0 .: 0: .: 0: .:
0:.”0 u 0 :.:0 n 0 000 u .0 0:.“0 u ,0 u n 000002000000:
00:88-00 0288.00: $508.00 00:88-8 muwamom- NW0: 5:885 0::5
BCDAH DEF SW GOP—Dam DEF 5V COT—0m DEC; “Hum BCDAH DEF USN COP—Dn— DEW—u Hm— Mo WDHDWNDE

 

 

 

.0000: 0000.: 0: 00:08:: 00:: 00:00.: 0:000 00:00 00.0 0: 00003000.: 00:00.: .000. :3 :80:0
::0: :.:: :2 :88 0800 :30...— :::0 01:88.3 \ mm 8008 ::0—0,: .5... :2052000 50.5 .8 0:::
000-0: .0 0000.02 00:08:00.: 00: 00:00:80.: 00 :38: 0 00 80003033805“: :00 200.0

 

 

143

Appendix C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:.0: 0.0: $028000: 00:8
0.0: 0.: 2:20:03 00:8
:0 00 00:00-03
m0: i0“ 0. :0 0.00 0.0 0:880: 0.0.0
0.0 0.0 :.: 0.: :.0 2:209:28 0:500 00002 0:00
0 0 0 0 : 00.00-08
0.00 0.0: :..00 :..:.0 0.0: 0:000:80: 00:8
0.0 0.0 0.0 0.0 0.0 23.003008 00000 .0000 0:-:0
: : 0 : F 0000.005
0.0 0.0 0.:. :..: Hm 0:02.000: 00:8
= 0.: :.: 0.0 0.0 0.: 25-008028 00000000 0:-00
_ 0: 0 0 0 0 00:00-09
0.0: :.0 0.0: 0.0: 0.00 0:02.000: 00:8
_ 0.0 0.: 0.0 :..0 0.0 23-003028 000000.03 0:-0
0: :: 0: 0: 0: 0000-005
00.0”.0 00.0.0 000 0008. 00.0.0
.00... .00.... .00... .00... 0.00.0.0 .5
0.0:. 0:00 000:0 00:. 000:0. 0.0 0000.0 000 000:0 0...:

 

 

 

 

 

 

 

 

 

 

.0000: 000:: 0: 00:80:: 00:: 00005000.: 0: 00:00.:
0:0:. 0.0:.—.0. 00:38.0” .3 ::0 0: Ami: 3008 009:0 \ 00:38.0 00.: ::0—0:308: 0: 50.5 $9.8 3008
::0—2 .3 08:52 5:88:00:— 0:.:_ 5:30:00...— .:0 :08: 0 00 8:03:83 "0.0 030,—.

 

 

3: i: 3.26289 Eon

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.0 0.00 €2.53 38
8 9: 8.5.03
0.0N 0.5. 0.3. odm Q: 2.0588025
_ 0.0 0.0 3 0.0 0.0 E2-fi>§28 2:50 552 0.00
m 0 0 0 o 09.5.;
1 Q0 oé 0.8 0.2 0.00 20280328
3 no 3 no I §2$>§28 9:50 5.50 0-00
N N m N 0 35-55
0.0 90 0.2 no 9.: 2589 Ex.
m 1 N; o._ m.m o; _.N E25355 oczomaam 3:
S 2 2 o 2 85;;
0.2 0.2 0.8 0.0 v. : 20288 38
ed 3 E 0.0 gm 25-530200 252383 0-0
2 2 m : 2 80.9-0.5
Sane 0000"... 0000"... Run 03$.
02800-8 08800-8 08800-8 08800-8 02580-8 00052.85 8.525 0::
8.5.0 8:8 8.5.. 85.. 8.55 082382 . . .
2:2. 50 25,0 50. 2:? Em 25... EN 2:: E

 

 

 

 

 

 

 

 

 

 

 

 

 

658m «mam m_ 5309.5 2.: .8EEutm m_ groa—
259 Rs... 02800.0 3 .5 .0 $.20 .85 880 \ 3580-0 .8 8:55 00 :85 SE:
895m :..0: .a hon—.52 5:828:— .5.._ 539:8; .3 258% a ma mozmzfimgg "ad «Ea...

Appendix C

Appendix D

Case 3 Preemption Results

30 5.05 55252558 Eco

0.80 0.050 22.5% £8
£0 .00 8.5-0.5

 

 

145

Appendix D

 

0.00 0.8 0.00 0.2 3.0 0.3. 5.5 30 5.00 0.8 3. 5.9. 30 0.00 :5 252890250
0... _.o S S. 93 $.m ed 2: m4. m.m ow wé ma w.m 3w A52é0>v >200 .chom 5.52 :-c—
0. a 0 5 t 0. o 0 0 0 0 0 0. 0 0 8:02.09
5.02 5.05 20 0.5: 0.8 0.2 5.3 5.: 5.2: 5.3. 5.3 0.00 0.05 0.00 55. 296090200
mdm 2: to pd Qam ad— w.m Md— >.m~ _.m 06 02 ad w.w W: A:_Z-:o>v >2an .Ucaom 520m :-:
a 5 : 0 2 2 0 0 a L. 0. 0 2 5 2... an.“ £5-00...»
0.:. 3.2 5.2:. E... 0.50 5.30 0.2 5.2 5.5 2: 0.00 0.5 2.0 .00 3: 202880500
QM «Em Qoo 3.. ma— ~.mo he 3. od «.0 ma— odm o._ o6 oém A52é0>v >300 vasomamwm :-—
§ 0 mm 2 v a 2 m 00 o _ mm 2 N S 2 80.05505

_

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.2. 5.00 5.5 as 0.8 0.5: 30 0.00 5.5 0.8 5.2 0.2 050 5.3 5.: 202885200

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5.5 5.9. 5.5 0.5 0.50 3 5.0 0.00 5.5 2 to... no 2: 0.50 0.5 25.530050 05.5983 :-00
0 50 o 5 50 0 o a o p 0 00 N 0 50 _ mpfiéo>
fix .0 4 m 0. _ 4 a F ._ m .0 4 m .0 ._
25"... Sunny zany 85¢ Sunny
8:88-8— Scoooméfi 3:88-8— 0ucooom-o~_ 8:08de 0853885 cognac 505..
8:8 vote“. no.5; cocoa. note; .6 022032
25 5E 55: 5.80 2:: BE... 55F 28.0 55:. Er.

 

 

 

 

 

 

 

 

 

 

 

.95 a 5:925 25 63:53.5 m_ 3.5m Em 0.0:.—am 3.5893 325m 0. 8.25 .—. a. m $5.8
52.0 =5: 2:. 62550 0. 8...: c3 $1.5 52.0 :52 02850-: E .5 .0 8...: m a .0 $.25 .85
89.0 .3 LEE—.2 $59858:— gem 529:8...— ._e 2:93. a 3 853.05% 3.: «En—s

146

Appendix D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5.0: 0.0.. 352525505 38
5.00. 0. S. €2.52 525
n: 0: 55.-.;
v.3 m w N.m_ _.: ow Eo>\oomv >209
a; N.N 9N 0mm 0; AE—zéoz >200 950m 552 0-3.
: 2 Q 2 fl 055-59
”i h H LI‘ ” fi
5.50 5.50 5.50 5.00 _.: 25805 5.50 H
3; 5.0 0.5 S 5.0 552-5% 38 2.55 gm 5-0
2 m E Z N. mth-:o>
:38. :08... :38. :38. :38.
me wd m.w WE N.N. 5.5 Q? NE 9w mg. 20288 .339
cd m4 m4 0.». M: 5.0 wd 9m N.N n; €02-53 >200 950m 53>? 9m
5 2 5 2 2 : 0 2 2 2 85-59
m A m A m A m A m 4
Sum HP 8% HF 20m" P 3% n F H u mmocoéootm
05:88-8 09.8098 accounméo 055805.00.U mcmwowm- We c2885 x54
Eton 08:. 5m cocoa 2:5. 5V vow—om oEF Em veron— oEF ccm Eton— oEF 52 .50 3.5032

 

 

 

 

 

 

 

 

 

 

 

 

.555: .050. m_ .5525 25 535.52.. .0 53.5.. 2.5 BE... .5... 0.3
58.5 :52 5.5 52 52.5 08.0 EB... 26 05.585.5— \ 5m 525m :52 .5..— tsm 055m 5 025
.83: .5 .2552 5.83.3.5 .5..— .55558... 5 .58: a 2 833045395905 .3 255...

0.50 5.00 0252-3855 0250

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.05 500 32-553 528
:0 55_ ace-55>
5.50 5.50 5.: | 5.: 5.2 125285535
5.5 5.0 S 0.0 5.0 25:-5055350 55:55 £555 050
5 0 5 5 5 any-55>
L . . . . .
5 50 0 55 5 50 0 55 5 55 295555 Eon
: 5.0 _.5 5.5 5.5 0.0 €2-55>:5_5o 55.55 5:52 0-50 5
I 5 5 5_ 5 5 I awe-55> I
n 5.00 5.0 0.50 _.: 5.50 252855025
1 5.: _.0_ 5.55 5.0 5.55 052-595028 55:50:30.5 0-5
00 50 50 _0 50 5.55-59
0.: 0.0_ I; 5.5 5.00 252855 5255.
0.5 5.5 5.5 0.0 0.: €_2-55>:5_5o 555555553 0-0
50 00 50 50 50 ace-55>
005$ 005$ sane 50mm"... 53¢
55:83-8 55:88-8 5v=OQom-oo 5258900 56:83-8 5505358me 52885 5.55
pom-80. ©0qu totem poten— votum 50 52553—2
Buc- Em 58C- 55 2::- Em DE:- ncm 08C. 555

 

 

 

 

 

 

 

 

 

 

 

 

 

.2555 58m 5.. 5:525 5.5 535552.. 5_ 55:5.—
5.....5 2:: 55885.0 05 =6 5_ 05-25 .855 58.0 0 35.5.5 55585-0 5355 55: 555.6 55.0.:
552.5 552 .0. .5552 85825.5 .5... 55:55.8: 55 5.585 a 55 85553533 "0.5 5.55.0

Appendix D

148

Appendix D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.: 5.5. 3.5252555. .25..
5.5. 0.0. 2.2-55>. 05.55.
50 50 55.5.2.5
_ 54.0 5.5. 5.00 5m - 5.5 255,555. 05.55. i
0.5 5.5 5.5 .5 0.5 8.2535028 55555 5.52 5.50
_. 0 0 0 0 0 555-59
5.50 5.5. 5.00 5.50 5.50 5.55555 55.50.
0.5 0.5 5.5 5.5 5.5 .53-...555255. 55550. 5555 520
. i . 0 . . ace-55> J I
0.54 5.0 5.5 0.0 ufi 0.5 2525555350. i l
5.. 5.5 5.5 0.5 5.. 9.2-5.55.250. 5585.550. 5.00
_ 5 5 : 5 5 ate-55>
_ 5.5 5.5. 0.5. 5.5 .5. 2555550255.
T 5.. 5.5 0.. 0.0 0.0 9.2-59:255. 5555.55; 5.5
— 0. 5. 5. 5. 0. 5555-59
50.5% 00.5".0 50.5% 00.0"? 00.0".0
555855-55 555585.55 555855.55 555585.55 555555555 55555555550. 555525 5.5..
55.55.. 55555 55.55.. 5555.. 55.55.. .5 5255552
555... 55 5e... 55 5e... 5.0 555.0 550 5.5...- .5.

 

 

 

 

 

 

 

 

 

 

5:55.. .550. 5. 55552.0. 5.... .3553... 5. 55.55..
55.5 5...; 555585.0- 55 .5 5. 6.25 552.5 55550 \ 55:585-0- 55. 55555555... 5. 58.... 55...... 555.55
5.52 .5. 555552 55.58553... .5... 55.55552... .5 .38.. 5 55 8.555555% .5... 5.55...

 

 

149

Appendix D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5.55. 5.05. 3.2552555. 05.55.
5.00 5.555 2.2-55>. 05.50.
550 055 55.50-59
0.50 5.55 0.55 ..50 0.55 ..55 0.05 5.55 0.50. 5 0.55 0.5. 5.55 ..00 5.50 £5288l05fimfi|
0.. 0.0 5.5 5.. 0.5. 5.5 0.5 0.5 5... 5.5 0.5. 0.5 5.0 5.5 5.0 05.2-..o>;5.5n 6.500. 5.02 :-....
. 5. 5 0 .. 0 . .. 5 5 5. . 5 5 5 55.....-505
5.55. 5.55 5.50. 5.55. 5.55 5.50. 5.05. 0.55 5.0.0 0.0. 5.05 5.55 0.50. 0.05 0.5 £5>\oomv.05.5o _.
5.00 5.5 5.00 0.00 5.0 5.05 5.5. 5.50 ..0 0.5. ..5. 5.0 5.0. 0.5. ..0 05.2-..5>Y05.5Q 6.50m .....om :-0.
n5. 5 h: 5| 5. 5. 5 5. 0 5 k: 0 5&0. 0 llama-..-..g ‘
5.50 0.55 5.550 5.05 5.55 0.500 5.05 5.55 5.500 5.55 0.55 0.550 0.00 5.55 0.550 Quincmisfl
0.0 0.00 5.05 5.5 0.0. 5.55 5.5 5.0 0.55. 5.0 5.05 5.55 5.5. 5.55 5.50 05.55-553.335. 9:50.555. ..-.
5 50 5. 5 5. 5. 5 00 5. 0 50 .. 5 00 .. ark-55>
0.55 5.05 5.5 0.55 5.55 5.55 0.5 5.00 0.05 5.05 5.55 0.55. 0.05 5.55 0.00 25>08mv05.50
5.0 ..50 5.5 5.5 0.50 5.0 5.0 5.0. ..0 5.5 0.55 5.. 5.5 0.05 ... 052-53.05.5Q 9:80.555? ..-00_
W0 50 5 0 50 0 5. 00 0 5 05 . 5 .0 .10 55...r-;5> ili ‘
m .0. A m ... A m .0. A m .0. 4 MM— ... 1.
5c...om.mmm|-.w~. 5....oomummu..0...~. 5ccwwuommu.%0. 5....M.~00.MW......~. 55=wwofiwnnw§ 552.53.05.05 50.825 5......
52.55. 52.55. 8:55. 05.0.55. 05.0.55. .0 55.55552
5......- ccE 5...:- 555... 5......- EE... 5E:- ocooom 5......- .52.”.

 

h

 

5. own—E to..— G.

.0...» 5. 55:55.5 5.... 55.55.55... 5.5 55525.. 5.....0 5.0 5... .5 .55.. 555 ....0 5... .5 .55.. 5555.55

2. .5535 555.0 5......» 555855.50 ..5... 55555.55. 5. 55.5.... .- 55 0. 55...... .5555 5.5:.

... .5555: ..5..555.5.... .5... 55.....555... .5 ....53. 5 55 85553533355335 .2. 5.55...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5.50 5.5. 3.5252555. 05.50.
5.55 5.50 2.2-55>. 05.50.
50. 55. 5.50-55>
335% :35
n . H‘ 5 o . .
a 0- n. m .5. m N. o w m o. £05309 05.50
= .5.m w.m w. 0.. .50 2.2-:55. 05.50 9:50 5.02 cém P
m. c. o N. L .5. ark-:03 1
i." H LI I n H J"!
00.5 me 5.35 Qwv 0.0.0- 2505088 05.50
0.5 5.0 5.5. 5.5 5.5 052-553 05.50. 55550. .550. 5-0 _
m N. m u“ om . . 0 math-0.05
.58..- .58... .58..- .80... .58... 1 k
5.5 wd md Wm. m6 0.. ad 0... m6 w.m 25>?an 05.50
ad 0... 5.. 01.5 ... 5.0. n. 0.0 ... ... 05.2.53 05.50 .5550. .553 c-n
5 5 0. .0 5. 5. 5. 5. 5. .. 5.50-55>
M. 1. m 1. m 1. y. .0 m 1.
mmfi u .- Nmum u ... Sun u ... Omfi u ... H u 5500.53.58
55:85m-oo 59.88-06 59.88-95. 5c=8om10© mummowm- .....o ....m 0.0.88.0 5.51.
D
X
m 5.55.. .550. 5. ..5..55..0. 5.... 55.5.. 5...... 5......- 5..... 5. 55555.50.
m.- 6230 ...—m. 0.5 «gm ...—5.). .555 07. 39.5 89.0 :5; 0:0 5..—53%.: 0 mm .00-.5 5.5.). .50 59.0 63:56.5
A

.5 55555555. .5 555.52 5555555.... .5... 5555.555... .5 5555.. 5 55 55.555.550.333 "5.0. 5.55...

151

Appendix D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5.50 5.55 05.255585. 05.50.

5.5.. 5.55 5.2.53 05.50.

0.5 555 5555-555
5... <2 <2 «.0 <2 <2 5.5. <2 <2 ..0. <2 <2 5.5. <2 <2 05525580550.
... <2 <2 _ 5.5 <2 <2_ ... <2 <2 5.5 <2 <2 0.5 <2 <2 552-555 05.50. 55550. 5552 5-5.
5 5 <2 <2+m <2: <i 5 <2 <2 1m <Mu <2 5 <2 <2 5555-55>
5.0 <2 <2 5.0 <2 <2 5.5 <2 <2 5.0 <2 <2 5.5... <2 <2 29.555.05.50.
..5 <2 <2 75 <2 <2 _ 5.5 <2 <2 0.5 <2 <2 0.5 _ <2 <2 05.2553 05.50. 55550. .5555 5-5.
FEEVEEFEEFEEVEE 555.-.... II
5.05 .50 .55 ..5. 5.55 5.55 5.55 5.5. 5.5 5.50 5.55 5.55 5.55 5.55 5.55 053.555.05.50.
.... 5... 5.5 _ 5.5 m... ..5._ 5.5 5.0 5.5 0.5 5.5. 5.5 0.. 5.5. 0.5 05.25.53 05.50. 5555555.. 5-5
5 55 f—ouflmmh 5. 0n 0. 5 5 50 5 0 R 0 55.5%.;
55. 5.5. 5.55 ...N 5.55 .....N 5.0 ..0 5.5 5.55 5.55 5.55 5.5. 5.5 5.5. 255555.053
5.. 5.5 5.. 7.5 5.5 5.. _ ..5 0.. ..N 0.5 5... 0.. 5.5 5.5 5.5 5.2-555 05.50. 55555.55... 5-.
IPPIPEPIWLVIFIFINIRLNIVVFLI 5.5-50 |.
M F ... M .... A ~— h. 1.. ~— ..—. 1. M h. 1—

55mmwm¢o mumwwws 5......”me mummmwg 5.5mm“? MWMHWMMM. 5.85:0 5....1.

55.55.. 555.. .5... 5555.. 555.5 5555. 5555.. 555.5 5...... 55.55.. 555.5 555555 55.55.. 5.55:5...

 

 

 

6:55.. .550 .... 5.59.5 5.... 652.555.... 5. 55.50.
5E... 5...... 650.055 5. 6-2. 55.5 85.0 \ 5.5.553 3 .5... 55.55.20 5. 555.... :.:.... d 2.55. .310.
.5555 55.). .N 50.5.57. 5.55558... .5... 5.555550. .5 .553. 5 55 835.55% "0.0 5.0.5...

3: _ 9: _ Eozozoomv E8

.3: 3% 22.53 Eon
ms 8v 8.5;;

 

 

 

E: ”.3 gm 3: 5.2 a: 3 .3 0.3 EN a? ”.2 in can 08 EoZoomKEoo
2. 3m 3 3 3m 3 3 ed 3 2 an 2 3 2: 3 AEZéoZQBQ .255 552 :-2

oEMMEmoomonnmfio mach-;o>

 

 

 

 

$3 32 «.3 to: we: 3: od 3 $2 38 ..S was 38 was 32 20289339

 

3m 3m «.2 0.3 nan SN 3. ad 2: SN we SN 3a 2; SN €_2-zo>:£oa 655m Ezom :-:

 

w 2 a S 2 2 o o 2 w : 2 m S 2 maker—29

 

 

 

 

SM 23 3% 0.:. 32 tom“ «.8 qt o «8 2x. 2% we. $H.~$ £028833Q

 

152
a

3m 32 2 E». 32 2 “.3 od 2 n3 n3 3. 3m NE AEEéQZbWEQ vasomawm :-—

 

m a 2 n a 2 m on o b 3 S n 8 : math-no>

 

 

”.8 3% an o8 3:. «.3 wow n8 3 who 2: OS 2.2 2” «.8 EoZBmZSoQ
2 ma 2 3. 2a 3 3 v.8 ed 3 we. 3 as 98 E EST—BBQBQ gamma? :-mm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

v Q n v a m m a. o o R o p 8 ~ 35205
M P A M _ F A M _ F A M F 4 fl ..—.. \—
mvnmuh man... Sum"? amnmuh Emu?
mucoooméfl $8832 mucoooméfi mucoooméd 88832 32.03885 .5885 is
85m 85m Eton 85m 85¢ .5 $5302
25% 5E oEF 5.50m 25% BEE oEC. 258m 08:. HEE

 

 

 

 

 

.5» g 5:925 3m 63:58... a £55.. oEF Em 623m 253 83E
:3 .85 5.52 .25 m a. h 6.5 .85 2:0 2...? ..Etam 3:3 3.2.. ._. a m $3.: sea :32
.= .252 5:883... .5“. 5:302: .8 in». a a 85:53 "ad 2...?

Appendix D

153

Appendix D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

v.2 3: Eafizooe Eon
mam QR Ezfiz ER.
8 a 35-5%
SN 98. S: 3: 3:“ 2289 50m
hm E .1 N.N «a E25393 258 582 fix
a n o h w 3.5-59
ofi 3: oé 3A. oi 20289 58
we ad no N; No §2¢o>v§uo 23m .38 o.%
N m N m _ $5-;o>
q: M: S: 2: n: 20289 58
N.N m._ 3 5 NM E35333 2.59am. 3:
2 2 o 2 : .35-;9
«.2 9o NE m. a mam 30289 Son
3 Va m.m 3 cs 25-5338 250933 a-”
— 2 2 2 : 2 3.5;?
Sane SHE. Sang 03 menu?
wccoooméo massage “28%? mucsomé 8808-8 $05285 8825 is
veron— nOtom poten— cOtom cocoa ho 8:5on
as? 5m 25... .3 2:: Em 2:: EN 2:: E

 

 

 

 

 

 

 

 

 

 

€55: enaw m_ 5:925 9.: dais—3.5 m_ 62.8.—
oEF FEE. .mvguomfi E :.0 m_ 6.5 «3.5m $9.0 \bEam ageoomfi cor—Sm 3: 520 95.8
.085 £52 .a 53:52 5.83.5.5 .8“— :33535 .3 =36: a ma Gums—wag “ad «Bah.

 

 

154

Appendix D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.....N ...m. 3.02.2.3. Eon.
v.9. ...... 2.2-53 ER.
8 .... ......T..>
N... S; o... o... # .3 23.89.28
2.. N... m... o. ...m 2.2-5.5128 2.8.. 552 gm
2 a . c a 85.?
3... 3% o... 3. ed 2589 EX.
.... 3 o... ...o I. .=.2-...>.§8 28m 58m 3%
m v o u N N 85.;
= no ....N 2 - o... o... 20289.28
__ m. m... ...o N. N.N 25-6%....8 28...... E:
N. n o o N. ...E-..o>
...: N... EN ed ...... 258... E8
3. 6o ow ...o 3 2.2-5358 28983 a-..
o. o. ... ... m. 85-53
8%”. Sam. 3%». 9% NE”.
8:838 888m... .280m.% .28...-8 85%-... Eggs... 8525 .5.
8.5.. no.5; Stun. totem 8...»... .o 3.38.). . .
as: am 25 5.. 2:: 2m 2:... EN 2:: ...

 

 

 

 

 

 

 

 

 

 

 

.23.. .83 .. 8.88... a... .8353... m. 8......
2.5 9...... ......exm .. 8...... 6.2. ......m 8.5 . 8.58m ... .2 82...... m. :85 95.”... .85
:..... a .2252 8.68.3... .5. 5.2.58... ... .38.. a 8 8.323%; 5.... 2......

 

 

Appendix E

Case 4 Preemption Results

155

Appendix E

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

...om N.Nm ...... N...N ...... ...S N...N ..mN new .... ....o ...Nm_...8 5.. ...... N5 N.N .... 28285.5...
N.N m... N.N .... 9N N... N... 3N ...N o... .... m... _ mm ...m w... ...m ...m .... .=.2-..o>.._..oo 6.55.2 :-....
v N. m m a rm v N. v .. w m m N N .. IV. M. ......Nés
N8 N8 N.NN N... n... ...; main. ...... ...8 ....N ...NN 0.8. ...: ...... N8 ...8 ...NW ......N .5285.....oo
...N. ...: m... N.NN v... ...m N... ..NN 3N ...m n... N.NN ..m .... n2 .... w... 5.: ....2-..o>.....oa .28.... ..-N.
w m. a N. : v m m. a. m w 2 ... M. m. a ... ... 85.; .
...... N.N. .....m ...N... ...... 28 ...NN :8 S8 ...... ...... .... N... N... 22 EN ..8 ...... 2.285.828. .
...... ...mm «.2 N.N .....N «.5 ...N Na ...... N... QN o... N... N.N. qNN .... ...N .....N ....2-.....>.....8 9.8.... ..-.
o ... m. .V on N. N M. .N m 5N .. . NN ... N N. m. 8....-8>F
3... .....N. ...... ...... ...Nm. .... 3.: ...N... 2.... 2m ...: .... N... ...N n... ...... .....N ...N 2.288.284
N2 .... N... N2 .... o... ...m ...N. ...N ...m 2N o... nN .....m N... ..o. 5.8 v... .....2-...>.§R. 2.5....» :-mm
.. ..m . 5 NM .. N N m a on .. 5 8 N N ..m . 85.?
N. .N .. N. N .. N. .. 4 N. .. .. N. ... .. N. ... ..
tank m .5"... m . any :5"... 85“.? No.55
5.2.85-8. 59.8.5-8. 5.2.85-3. 5.2.85-8. 5.2.88-3. 5.2.85-8. mmo=o>..8..m 5.825 2....
.85.. 8.5.. 8...... 8.5.. 8.5.. 8...... ... 85.82
2...... 5.2m 2:... 5..". 2...... 5.50”. 2...... 2...... 2...... 2.02% 2...... .8...

 

 

 

 

655m .8...» 5. 5.8.2.5 55.. 50382.8. 5.5.3.. :3 ...... EM

2.. ... 0.:. 53559.50 ...... 5.2—E8...— ...o._..um ...—E855... 63.55 5...... ... an m 55-... «3.5m :....2

...... 2.8.25 .. 8...... .... .38.. .85 :..... 8:85-.. ... ...u m. 8...... m a ... 5-2. .85 880
... 8......z 5.888.... ...... 5.2.58... ... .38.. a 8 8.8.853 ...... 2......

 

 

156

Appendix E

 

II

 

NM.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

...o. 0w ON. C... 202009 .9200
9.. N.N 9N ...N 0.. 9.2-:03 >200 950.. 5.52 0......
.. 2 N. m. T N. 85-..; -
..N. .....N fl 0% ...mm ..N. .5289 .28..
3.. ON ...w 0.0 QN ASE-:03 20.00 ......om ..N.”. 0..-
.So... .20... A. .20... .80... .80... I
o6 v.0 m6 ...: ed ed o... m... M: M.» 202008 ......0Q
o. .... Wm w.m ..N o. N... a. N.N .mm 3.2-:03 .....0n. 2.23.303 c-m
a m. m. m. ... o . ... N. ..N 8......9
m -. N. 1. M. -. m x. y. 4
....mn... . u min"... . n u u mm0..0>..00....m.
22.88-00 mWhomom-wo .....ooom-oo mWhowom-Nwo muwwowm-m-o 5.80.5 0......
8...... 2...... 5m ......0. 2...... 5w 8.5.. 0......- Em 02.0.. 2...... ....N 8.5.. 2.2-.... ..o 8.300.).

 

 

 

.25.... .8... a. 528.... ...... .8258... .... 8.8.. 2...... 2...... ...». ...5
.85 .....2 ...... ..z 885 820 =5... .5 358%... . .... 88.0 :..... .5... .85 2...... ... 8....
82... ... 82.52 8:883... .5... 8.2.58... ... .38.. a 8 8.3.3353 MN... 2...?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

02 SK 20289 35
N.N 3 235% ES 28m 58m w:
w a 855$ -
ado 3m 3: 20289 35
_ 3 3: ma 3 no 22.506 38 2:8 552 gm
_ v w w u g 35.5%
. Zn 3; 5m v.2 in 258$ 35
m S: «.0 2: 3 2: £2.53 38 28m .5 3
a ..
S cm VM flw um H a m Mm 50> n I
ma go we no of 258$ 33
on S 3 § mm £2.53 55 225 a»? 2
3 ON om 2 § £5-53
Sane mane Sane Sane a any
883% $83-8 $803-8 $8830 888%?» accoéootm 8825 V25
totem v3.8a carom cote; woven no 85302 . . .
08F 5w 25L. Ev 08E. Em 08F Ucm 08F HE
E
.m
m .252. as”. a 8:925 8m 32.58.... a 8...».—
m 2.5 2.; 3:88.». 3 .5 g 3% sea 820 \ 3:3”... €=§m.m Bram 3: :85 95.8
A .85 :32 .n :..—5.2 £8023... Sm 5.952.. ... :33. a 8 85:33? "3 2.5

 

158

Appendix E

 

 

 

 

 

 

   

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. S: 3: 2388 Son
. no no 2:53:30 28m 58m 93
_ N SES> I!
. 3N 3N 238$ 55 F
we mo 3. S S Ezéfifig 28m 582 EN
m N m m m 85-59
”.2 .2 3 Na 3 258$ 38
an E 3 3 S. 23-5338 258 sum 3
ON om 3 a! we amwfi>3 I1
3: 0.8 v: ”m 4.2 EoZMmIamo II
3 2w 2 1: S: E25338 28mm»? 3
a S Ma N.N mm £5-59
_ mmumup Sane Sane oNnmuF Sane
mU:OQOM-O© mficooomuoo mUCOoomAvc mfiCOQomAVO mUCOUOm-O© mmoco>muootm cocoobfl x54
totem totem cocoa cocoa notom mo 8.3282 . . .
08F 5m 08C. 5v 08:. Em 085. EN 2.5. “fl

 

 

...—=5: .83 m_ .839th as: ...czamcogficu
e2 .8358... g SE.— 85 9...: 62.3% g 6.5 .85 880 \ 3.55 x :85 $9.6
39.5 532 .m 53:52 2:83:55 .8,.._ 5:9:er .3 :33— a ma 833553.353 and «Ba...

 

 

159

Appendix E

 

 

o.v~

3:.

j
l|ll

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SN ed 3 538% 58
M3 mo We No to 52-53 58 258 552 2-8
N m N N m 3.5.5>
odm ofl odm odm o.v~ 20288 58
no No 3 no to 2:253 58 58m 58m 2-;
_ _ N 5 _ any-55> *
S w.m 3. 3.. SW 5389 58 i.
o; no wd 9o 3 52-53 58 55855 SNN
a w : 2 N 85-5>
M2: 5.: Ni 3 NE 5389 58
3 mm N.N o.N N.N 52-53 58 “.5853 2-0
2 ”N 2 E 2 8.5-55
oNnnnh. 3m"..- .Nnmnp NNumne NNHE
.5... 5...... ...... ...... .5...
25 5m 25 5.. 2::- E as: EN 2:: 5

 

 

 

 

 

 

 

5.58 58 5 .5525 ...: .5355...— 5 8:8
«:.:. 22:. 5:88..” B .5 5 6-3 525 520 \ 558m...” 5 8.5.5 5 596 95.8 .85
.52 .2 52.52 558.5... S... 5558... .... 58¢ a 8 855533 a... 5.5-N

 

 

160

Appendix E

 

 

 

 

 

_.NN «$123 N.No Maafla nod ofl EL 3 a? N.NN. w? a QNN Eozoomisoo
9N 3: E N.N i 3 3 $3 tn 3 ed N.N ad 2: N.N i 3V 3 23-5358 .2582 :-2
N : m N 2 m N NN N c o v o E o n N n 8th.; __
32 Nam N3 EN 22 3.2 N.NNN 38 can 3 od N.No 32 N3 3N N.N2 9% N.Nn £38958
9:. 3 N.N: New N.NN 3 N.Nv 9mm N.N_ ad ad Ni ”.2 SN N.N 2: q: 2: E23658 658% :-N_
N_ n Nfi 2 3 N 2 Elna c o] 2 e a N N N_ : 8th.; _
N.N2 3.: N.NNN 3: Ni: N.Nom o... S. 2% N.N_ 3a .3 N? o? ENN on N.Nn N.NNN Euzoomisoo
N2 N.NN 3» q: ..8 SN ad 3 no: NB «.3 3 no S: 33 3 EN N.NN. 93:91.28 958 .m :-_
N 3“ 9 N X 2 o o «N N WN c _ 8 2 N NN 2 watt; 4
N.N2 3.3 N.NNN N.N: Ni: N.Nom Qo 3 SS 33 32 ed 38 N.N2 N.NN: N.NNN 3: n; 2028858

:. N6: Wm Wow wdo 96 Cd $6 0.2 eém 0;: od wd— Wm: w._ w.NN Wax fig A52é0>v zflofl ccsom.>» :-mm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a: - o
N mmlfl m w Nm o o" o o 2 no a w Nm _ cl wN N uh ..N:>
m ,r A m ... q a .r A m p. \— M p. A m H 4
mm” mug. Nm N ”mu? om” mu? hm” mug. mm” mu? mm” mug.
mucoooméfi mccoooméfi mucoouméfi mocoooméfi mucoooméfi mucoooméfi 3053085 .5325 x51.
venom carom totem notom venom notom .0 8.33.2
08¢ 55m 2:? NEE oEF 550m 08:. .55. 08F 25on oEF NEE

 

 

 

 

 

 

 

 

...—50m .83 m_ 5:925 25 50382.8.— .meotem Se 25 ...—m 2: E 95 533:2..N80 EB gun—53.5
«.323— BuuSm 8a.... h Q ~— Aafié Em ENE 23. 4.2—aim 95 83.5 :3 95.8 385 :32 % ~— a. h
8-2V .085 880 .= .352 ..363.35 ...E =o_:_e§.. 3 =33. N a azésmgauag 6d 2.5

161

Appendix E

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we 3: N.N_ 3w 2: £338 528
N.N N.N 3 f N.N 32-6% N28 258 552 EN
2 2 o N_ 3 853$
53 me mi” 03 hum £0288 NAN—on
N.N N.N 3: N.N 3 32-5% 38 25am sum ON
2 m ON : N. 3529
" W ” |"1
y Bop EoN EoN 38. 238‘
_.N N? no N.N_ S 3 to of ON no 2288 N28
no no N.N 3 3 no N.N N.N o; 2 EESa 38 958 so? 3
o N 3 ON 0 m 2 E w 2 acN-;u>
m 4 m 1. m a m a m 4
mmum u .N New u ,N 35 u N can u ,N aNum n N 382805 858.: s
$808.8 $588.8 $508.8 mucoooméo 8:80” on . .Q x .4
woven. 08F 5m 85m oEF 5v coron— oEE. Em vote; 08:. cam notom 08:. “2 .No 35232

 

 

 

 

 

...—:5: Ram— m_ 5:025 «an 63.5.— 2.5. ...—Eh. 2F m_ coup—Sum vote.—
.c2 :3 team :32 1:.“ :2 39.5 820 EB... :.0 3.88%.: \ mm .085 :32 .5..— 586 8533‘ 3
3503.3 6 53:52 538235 .5...— =£EE¢P£ 3:581 “ ma 8E=Sm§ "Nam «Ea...

162

Appendix E

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

30289 38
E2€>128 28m 58m N-N
grins H I.
258935 L In
3 3 N.N N.N N.N EESEEE 28m 552 SN
w w o N N 35.5%
MEN N.NN Sm N.: n2 20288 35
v.2 :N N.: N.: v.2 EESaEB 2:85am 3
8 mm E N_ u" 9N i 85-;
ed N.N: of v.2 03 20288 55
N.N 3 Ne Ne 3 32-50335 28983 N-N
oN mN x cm ; grins
omnmue Nomi NmnnuN ommuN mmunuN
mUCOoomAvc mficongomAvo mfiCOoomuoo mficoooméo mUCOoomAVo wmocozuootm :OSUOBD x51—
votom totem cotan— votom voted No 85322 . . .
08F Em 08:. :3» 08F Em 06:. EN 08F H2

 

 

 

 

 

 

6:25 6am m_ 5399.5 9m 45358.5 fl totem
oEE. PE; .258on 3 :.0 m_ Amiv 39.5 89.0 \ Siam meccaomé 325m 3: 520 A34:
«025 :32 .m 53:52 53896.:— Sh 55:58...— .3 .38: a ma fiumzfimdfig and... «BE.

 

 

163

Appendix E

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ o.NN 98' of no. 98 39.88.58
m... N.N n. N.N N.N 9.2-5358 28m 582 N-NN
o. o o N N act;
W o.~. o6. ow. QVN o6 Eoioomv-QBQ
to N... no N.. ..o 3.2-5% N28 .525 5:8 N-NN
N N N N . 83.;
m... 3 o... N.m N.: 38289 .28
fi ..N ... QM Nd ..m 3.2-cot N280 vasomaam 0.0.
: o. o. w m. mace-=9
__ N.N N.N. N.N N.NN 8.8 £588.25
_ v.0 N.N N.N qm N.N 3.2-=9. N28. 252. .83 a
T m. N 3 N N 8.528
N3"..- Emu-.- ovmnN New mama-N
8:888 8:88.8 8888-8 8288-8 8:883 888.85 828.5 .5.
85m totem cocoa. cocoa. Bron. .o 8.3.3.2 . .
2::- sm 8.: 5.- 8.:- EN 2:: EN 8.: a.

 

 

 

 

 

 

 

 

 

.28.. as. a. 5.89.... a... .8858... m. 8.8..
«E... ......F “Begum-m E ...u m. 6.2. .85 8.80 . 8:5. museum-m Bram 8.. ..85 $3..
.85 5..: .N 2:52 5.888.... .5. 8.2.58... ... .38.. a 8 822.38% a... «EN-N

 

 

164

Appendix E

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N... ..N.. ..N.. ....N a... .....N_...N2 ...... 3..— o... o... ...—N.NN o... ...N _ N... .... o... 29.8838.
N. ...: .... a... .... N. _ ... ..3. N... _ a... o... N. _ ... .... N.N _ ... Nd. .... 2.2-5%....8 .2892 ..-...
m m. N N ..N.. . rum 8“. . o o N N . . _ m N. . 2.5.;
..N.. .... .... ...... ...... N... ...NN ...... N... o... o... ..N.. «SN ..N. NS. ...NoN ...NN ..N. £828.12...
.... ...... a... o... N... ..N .... N... .... o... o... N... N.N ..N. .....N .... .... N.N. .....2-..o>..a.oo .28.... ..-N.
N N. .. E NN . N. .. .. o o ... w : N. a N. .. math-30>
...... ....N. 23 ...NN. ...: :..... c... o... SN. .... o... o... N.N ...... ...ov N.N. .... N.N... 2.2885...
.... ..N. Ne: ...... .... ...... o... c... ...... .... 3N o... ... ...... a... .... .... .... 2.2-53.3.5 2.8m... :-.
v 88. ... N m. .. o o .N F .. o . 8 c. ..N.. NN N. mace-sow
...... 58 o... 3.8 ...... N.N o... o... ...N. ...N. N.: o... ......NN ..3. ..N.. ...... ..N.. N... €288.28.
... ..NN. o... ..N. ...S. .... o... o... N... .... ..N... o... ....N ...... N... .....N ...... .... 3.2-5358 28...? :-..
. .. o . m. m o o m m .N c N ..N . a ..N . math-50>
w. ... \. m. ... A y. .- .. w. ... A N. ... x. y. ... 1.
NguN .....u... .....uN .....uN ...".N 8.".
mucoooméN— mucoooméN— mucooomém. mocoooméN. mucooomém— mucooomém— mwoco>..oo....m c2825 3:3
cocoa vouch. woven. cocoa. cocoa. coton. .o .2...on
2.....- Sim 2...... 5..... 2...... 5:5". 2.....- E...... 2.....- .Eooom 2...... ......

 

 

 

 

 

 

 

 

 

.25.... .83 m. 8.8.... ...... ......888. «8.8.. 5.. ...... ....

2.. ... 2:. :c..a.=2.EcU ....a 5.2.58... 5..-...... museum-mm 63.8% 82... ... a m $9.... 82.5 5...).

.22 62...... .. 8...... a... .3... .8... :..: £589.-.. ... :6 .. 8...... a a N ...z. .8... 820
... .852 5.88.3... ...... 5.2.58... ... .58.. a ... 8.2.3.3 5..... 8......

Appendix F

Volume Sensitivity Results

Appendix F 165

Table El: Cumulative Network Statistics; With and Without Preemption.

Veh-Trips

 

 

U2:1 No Preemption

Delay
Veh-Hours

Delay
Min/Veh-Trips

1.15

 

 

 

Preemp., No Comp.

1.50

 

Preemp. W/ Comp.
No Preemption

 

    

1.34

 

 

Preemp., No Comp.

   

 

 
 

 
 

Preemp. W/ Comp.

 

No Preemption

 

     

Preemp., No Comp.

 

Preemp. W/ Comp.
No Preemption

 
  

 
 

 

 

Preemp., No Comp.

   

 

Preemp. W/ Comp.

   

 

 

No Preemption

   

 

Preemp., No Comp.

   

 

Preemp. W/ Comp.

—-—————————_.

 
  

 
  

 

 

   

 

 

   

 

 

   

 

   

 

   
  

 
  

——_-—_—_——I

No Preemption

No Preemption I 3324 16.83 0.30
Preemp., No Comp. | 3324 17.32 0.31
Preemp. W/ Comp. | 3324 17.46 0.32

No Preemption l 2770 E 14.96 L 0.32
Preemp., No Comp. l 2769 15.88 0.34
Preemp. W/ Comp. | 2770 15.53 0.34

 

Preemp., No Comp.

   

 

Preemp. W/ Comp.

_:

   
   

 

 

No Preemption

 

 

 

Preemp., No Comp.

   

 

Preemp. W/ Comp.

 

 

 

 

Appendix F 166

Table F.2: Cumulative Network-Wide Bus Statistics; With and Without Preemp-
tion. Route 1.

    
   
   
 

   
  

  
   
  

Person T—
'Iime
(Min)

T—Travel-
Time
(Bus-Min)

Mean '1‘-
Time
(Sec/Bus)

   
   
 
   

   
  

   
  
  

Bus

Person
Trips '

Trips

    
 

  
 
 

   
  
     
  
   

  
  
 

  
   
 

O

  
  
 

355.8
361.3

122.0
123.8

14.2
14.4

Preemp., No Comp.

   
 

Preemp. W/ Comp.

     
  
     

      

   
   
 

O

  
  
  

362.2
374.2

123.8
127.9

14.5
15.0

   
 

Preemp., N 0 Comp.
Preemp. W/ Comp.

  

   
  
       
   
 

O

   
  
 

      

351.3
530.4

120.4
121.9

   

14.1
14.0

Preemp., No Comp.

       

    

   

Preemp. W/ Comp.

    
  
        
 

   
  
  

   

O

  
   
 

360.8
362.5

123.9
124.5

14.4
14.5

   
   

Preemp., No Comp.
Preemp. W/ Comp.
0
Preemp., No Comp.

   
  
  

   
   
 

  
  
   

     
 

357.5
360.0

122.7
123.8

    
    

14.3

Preemp. W/ Comp. 14.4

   
    

   

O

   
 

  

350.0
352.1

       
 

 

120.4
120.9

14.0
14.1

Preemp., No Comp.

      

   

Preemp. W/ Comp.

         
   
 

     
  
  

    
   
      
  
 

   
   
  

  
 

o .
353.3
354.6

121.3
121.7

14.1
14.2

Preemp., No Comp.

   

Preemp. W/ Comp.

   
  
        
   
  

   
  

O

  
     
 

349.6
350.8

120.0
120.6

14.0
14.0

Preemp., No Comp.

   

  

Preemp. W/ Comp.

   
  
      

   
   
  

   

o
Preemp., No Comp.
Preemp. W/ Comp.

  
  
 

345.0
345.0

118.3
118.3

13.8
13.8

   
  

Appendix F

167

Table E3: Cumulative Network-Wide Bus Statistics. With and Without Preemp-

tion. Route 2

 

 

      

  

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T-Travel- Mean T- Person T-
Trips Time. Time Trips “We
(Bus-Mm) (Sec/Bus) (Mm)
U2:1 No Preemption 10" 123.1 132.6 250 578.3
Preemp., No Comp. 10 21.3 122.7 250 533.3
Preemp. W/ Comp. 10 21.4 123.1 250 535.4
U3:1 No Preem-p-tion 10 i 23.8 142.?— 250 594.2
Preemp., No Comp. 10 23.6 134.8 250 590.4
Preemp. W/ Comp. 10 25.2 143.5 250 629.6
U5:1 No Preemption 1 2274— 128. 250 558.8
1 Preemp., No Comp. 10 21.2 121.7 250 529.6
Preemp. W/ Comp. 10 21.2 121.9 250 530.4
M2:1 No Preemption ho 21.6 129.4 925F74O8
Preemp., No Comp. 10 20.5 123.0 250 513.3
Preemp. W/ Comp. #10 20.5 123.0 250 513.3
W No Preemption 10 21.3 127.1 250 531.3
Preemp., No Comp. 10 20.3 121.2 250 506.3
Preemp. W/ Comp. 10 20.3 121.2 250 506.3
‘ M5:1 No Preemption 10 21.2 126.9 250 530.0
Preemp., No Comp. 10 20.0 120.0 250 500.4
Preemp. W/ Comp. 10 20.0 120.0 250 500.4
1:234? Nopgémpnon fi'=—1=‘——m 7n 11331-75? 527.1
Preemp., No Comp. 10 20.1 120.6 250 502.1
Preemp. W/ Comp. 10 20.1 120.6 250 502.1
L3: fl No Preemption ‘1 10 21.9 125.8 250 548.3
Preemp., No Comp. 10 21.1 121.3 250 528.3
Preemp. W/ Comp. 10 21.1 121.3 250 528.3
L5:1 No Preemption 10‘ 217.8 125.2 250 = 545.4
Preemp., No Comp. 10 20.6 118.8 250 516.3
Preemp. W/ Comp. 10 20.6 118.8 250 516.3

 

 

 

 

¥

 

Appendix F 168

Table FA: Cumulative NETSIM Person Measures of Effectiveness For; Before and
After Preemption.

 

 

 

 

 

Person Travel Time Delay Avge. Delay
Trips (Person-Min) (Person-Min) Sec / Person
U2:1 No Preemption 6543 8803 6665 61.1
Preemp., No Comp. 6448 10735 8628 80.3
Preemp. W/ Comp. 6520 9917 7786 71.7
U3:1 No Preemption 6698 7677 5488 49.2
Preemp., No Comp. 6690 8912 6726 60.3
Preemp. W/ Comp. 6521 9309 7168 66.0
U5:1 No Preemption 6148 4191 2182 21.3
Preemp., No Comp. 6141 4235 2282 21.8
Preemp. W/ Comp. 6121 4435 2435 23.9
M2:1 No Preemption 5773 4337 2450 25.5
Preemp., No Comp. 5755 5714 3833 40.0
Preemp. W/ Comp. 5782 5001 3112 32.3
M3:1 No Preemption 5718 3141 1431 16.6
Preemp., No Comp. 5181 3214 1521 17.6
Preemp. W/ Comp. 5181 3174 1481 17.1
M5:1 No Preemption 4706 2560 1022 13.0
Preemp., No Comp. 4706 2598 1059 13.5
Preemp. W/ Comp. 4706 2610 1072 13.7
L2:1 No Preemption 3985 2241 939 14.1
Preemp., No Comp. 3994 2314 1009 15.2
Preemp. W/ Comp. 3994 2291 984 14.8
1.3:] No Preemption 3613 1905 724 12.0
Preemp., No Comp. 3613 1922 741 12.3
Preemp. W/ Comp. 3615 1936 754 12.5
L5:1 No Preemption 3297 1606 528 9.6
Preemp., No Comp. 3297 1616 538 9.8
Preemp. W/ Comp. 1624 560 10.2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Appendix F

169

Table F.5: Cumulative NETSIM Bus Statistics; With and Without Preemption.

U2:l

U5:1

M2:1

M3:1

M5:1

L2:1

L3:1

L5:1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Avg.
Links Bus- Delay
Trips (Min)
No Preemption I 18
Preemp., No Comp. 18 18.4 I 11.5 I 38.3
Preemp. W/ Comp. 18 18.6 11.8 39.3
.____________. 17
Preemp., No Comp. 18 19.6 13.7 I 45.7 I
Preemp. W/ Comp. 18 21.6 15.6 52.0
No Preemption 18 18.2 12.3 41.0
Preemp., No Comp. 18 17.2 11.3 I 37.7 I
Preemp. W/ Comp. 18 17.2 11.3 37.7
No Preemption 17 18.2 12.6 44.5
Il Preemp., No Comp. 17 16.6 I 11.0 I 38.8
Preemp. W/ Comp. 17 16.7 11.1 39.2
No Preemption 17 17.8 12.2 43.1
Preemp., No Comp. 17 16.3 10.7 I 37.8
Preemp. W/ Comp. 17 16.4 10.8 38.1
No Preemption 17 16.9 11.4 40.2
Preemp., No Comp. 17 16.1 10.7 I 37.1
Preemp. W/ Comp. 17 18.2 10.6 37.4
No Preemption 17 17.8 12.2 43.1
Preemp., No Comp. 17 16.2 10.7 I 37.8
Preemp. W/ Comp. 17 16.3 10.8 38.1
No Preemption 18 18.0 12.1 40.3
Preemp., No Comp. 18 17.0 11.1 I 37.0
Preemp. W/ Comp. 18 17.0 11.1 37.0
No Preemption 18 17.6 11.7 39.0
Preemp., No Comp. II 18 16.6 10.7 I 35.7
Preemp. W/ Comp. 18 16.6 10.7 35.7

 

 

Appendix F 170

Table F.6: Cumulative Network Statistics; Four Periods Before and Four Periods
After For Each of the Four Selected Preemptions. (No Compensation).

Delay Delay
Veh-Hours Min/Veh-Trips

Preemption Before /
# After

Veh-Trips
After 348 323.5 55.78
After 342 454.7 79.77

After 329 568.6 103.7

After 1.3 115.13

 

Appendix F 171

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

”“32”” 32132? Veh-Tfips v.38... Min/32%....
M: 1 Be ore WT
1 After 272 85.1 18.77
2 JBefore 263 85.0 1939—
2 After 269 76.8 17.13
3 _BLefore 274 77.9 17717—1
3 After 265 94.8 21.46
4 Before 264 66. 8 15. 18
4 After 264 105.1 24.00
C 016 . .
1 After 243 58.8 14.52
2 Before 246 66.9 16.32
I 2 After 242 53.6 13.29
3 1'3efore 240 53.6 13.40
3 After 243 60.8 15.01
4 '1'3efore 244 51.3 12.61
4 After 240 59.5 14.88
L2:1 1 Before 202 51.6 15.33
1 After 206 50.1 14.59 I
2 1‘3efore 206 80.1 23.33 I
2 After 200 50.3 15.09
3 ‘Before 201 46.4 13.85
I 3 After 202 62.0 18.42
4 Taefore 201 77.4 14.15 I
4 After 204 63.5 18.68
L: 1 Be ore 1 . .5
1 After 181 32.5 10.77
I 2 More 182 52.5 17.31
I 2 iAfter 118 34.0 11.46
3 Before 178 36.6 12.34
3 After 180 46.9 15.63
4 Before 179 37.6 12.60
4 After 183 40.1 13.15

 

 

 

 

 

Appendix F 172

 

 

 

 

 

 

 

”“11”“ 888:...
5:1 ° 1
I 1 After 164 29.3 10.72 I
2 Before 163 36.0 13.25 I
2 After 159 29.4 11.09
1| 3 Before 157 25.2 9-63—
3 After 163 25.9 9.53
I, 4 Before 160 22.2 8.32
4 After 160 27.9 10.46

 

 

 

 

 

 

 

 

Appendix G

Results of Carpools

Appendix G 173

Table 6.]: Cumulative Network Statistics; With and Without Preemption*.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Veh-Trips Vefiffllgurs Min/giligl‘tips
W

Preemption 3043 121 . 15 2. 39
5.0% Carpool 3048 114.76 2.26
Preemption 41: 3035 1 15.06 2.27
10.0% Carpool 3020 106.97 2.13
Preemption 2880 l 16.23 2.42

 

 

 

* Results are based on a 20-minute simulation period.

Table 6.2: Cumulative NETSIM Person MOEs; Before and After Preemption.

Base Volume
With Preemption

Travel Time Delay
(Person-Min) (Person-Min)

Avge. Delay
Sec / Person

 

 

 

40.8
40.6

 

   

 

5.0% Carpool I 38.1
With Preemption I 38.0

 

 

 

 

 

10% Carpool 35.9
With Preemption 39-8

Appendix G

174

Table 0.3: Cumulative NETSIM Bus Statistics; With and Without Preemption.

 

 

 

 

 

 

 

 

 

 

 

 

 

LinggtallBus- Travelil'ime DelayTlme Avge. Delay
Trips (Mm) (Mm) Sec / B-Tr1p
Base Case Volume I 28 55.9 38.3 82.1
With Preemption I 28 . 55.7 38.1 81.6 "
5.0% Carme I 29 56.5 38.1 78.8
With Preemption I 28 54.7 36.8 78.9
70% Car—pool I 29 56.2
With Preemption I 29 55-4

 

 

 

 

 

 

 

Table 6.4: Cumulative Network-Wide Bus Statistic; With and Without Preemption.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Meank k Person

Route Bus Travel- Travel- Person Travel-
Tr1ps Time Time Tnps T1me

(Bus-Min) (Sec/Bus) (Min)

Base Case 1 1 25
Volume 2 1 26.5 1208.4 25 662.9
With 1 1 29.1 1304.3 25 728.8
Preemption 2 1 26.4 1200.9 25 660.0
5.0% I 1 1 31.1 1329.7 25 778.8
Carpool I 2 1 25.3 1136.0 25 631.3
Mth 1 1 29.7 1329.1 25 743.3
Preemption 2 1 25.0 11 13.0 25 625.4
_—_' 1 ‘T;=—3T.—t‘_

2 1 26.8 1110.7 25 668.8
1 1 30.3 1275.8 25 757.9
2 1 25.0 1114.0 25 625.0

 

 

 

 

 

 

175

List of References

1. Alice, Maria, Prudencio Jacques, and Sam Yagar. Demonstration of the Characteristics
of the TRANSYT-7F Model as Modified to Passenger Near-Side Transit Stops. Trans-
portation Research Board Conference Preprints, Paper No 930715, Washington, DC,
U.S.A, January 1993.

2. Ann Arbor Transport Plan. BRW, Inc., Minneapolis, Minnesota. November 1990.

3. Benevelli, David A., A. Essam Radwan, and Jamie W. Hurley, Jr. Evaluation of a Bus
Preemption Strategy by Use of Computer Simulation. TRB, Transportation Research
Record 906, 1983, pp. 60-67.

4. Bishop, C. M., D. B. Richardson, G. W. Carr, and D. S. McMillan. Transit Signal Prior-
ity -- Another Look. ITE Compendium of Technical Papers, Washington, D. C., pp.
228-232, 1988.

5. Casey, R. P., L. N. Labell, S. P. Prensky, and C. L. Schweiger. Advanced Public Trans-
portation Systems: The State of The Art. Final Report. U. S. Department of Transpor-
tation. Urban Mass Transportation Administration, Washington, D. G, April 1991.

6. Chang, Gang-Len, Meenakshy Vasudevan, and Chih-Chian g Su. Bus-Preemption under
Adaptive Signal Control Environments. Transportation Research Board Conference
Preprints, Paper No 950155, Washington, DC, U.S.A, January 1995.

7. Davis, R, C. Hill, N. Emmott, and J. Siviter. Assessment of Advanced Technologies for
Transit and Rideshare Applications. Final Report, NCT RP Project 60-1A. Prepared for
National Cooperative Transit Research and Development, Transportation Research
Board National Research Council. Washington DC, July 1991.

8. Elias, Mlbur J. The Greenback Experiment--Signal Pre-emption for Express Buses: A
Demonstration Project. California Department of Transportation, Sacramento, Report
DMT-014, 1976.

9. Evans, H. K. and G. W. Skiles. Improving Public Transit Through Bus Preemption of
Traffic Signals. Traffic Quarterly, Volume 24, No.4, October 1970.

10. Hubschneider, H. Bus Priority Using a Bus Guidance and Control System - A Simula-
tion Study. Proceeding of the International Conference of Road Traffic Signalling,
London, United Kingdom, pp.33-37, The Institute of Electrical Engineers, London,
1982.

11. Ingalls, Larry, Kern Jacobson, and Ethan H. Melone. Alternatives for Providing Prior-
ity to High Occupancy Vehicles in the Suburban Arterial Environment. Transportation

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

176

Research Board Conference Preprints, Paper No 930583, Washington, DC, U.S.A,
January 1993.

Khasnabis, Snehamay, Rajasekhar R. Kamati, and William Cisco. Ann Arbor Signal
Preemption Study - Interim Report. Wayne State University, October 1994.

Khasnabis, Snehamay, Gangula Reddy, and Bharat Chaudry. Signal Preemption as a
Priority Treatment Tool for Transit Demand Management. Vehicle Navigation &
Information Systems Conference Proceedings, Dearbom, Michigan, U.S.A. October
1991, pp. 1093-1104.

LaBelle, Sarah and Darwin Stuart. Diverting Auto Users to Transit: Early Lessons
from CI‘A’s Orange Line. Transportation Research Board Conference Preprints, Paper
No 950134, Washington, DC, U.S.A, January 1995.

Liberman, E. B., A. Muzyka, and D. Schneider. Bus Priority Signal Control: Simula-
tion Analysis of Two Strategies. TRB, Transportation Research Record 663, 1978, pp.
26-28.

Levinson, H. S., C. L. Adams, and W. F. Hoey. Bus Use of Highways: Planning and
Design Guidelines. NCHRP, Report 155, 1975.

Ludwick, John S. Bus Priority System: Simulation and Analysis. Final Report pre-
pared by the MITRE Corporation, for the US. Department of Transportation, Report
UTMA-VA-06-0026-1, February 1976.

Radwan, Essam and Jamie W. Hurley, Jr. Macroscopic Traffic Delay Model of Bus
Signal Preemption. TRB, Transportation Research Record 881, 1982, pp. 59-65.

Richardson, A. J. and K. W. Odgen. Evaluation of Active Bus-Priority Signals. TRB
Transportation Research Record 718, 1979, pp. 5-12.

Roark, John J. Synthesis of Transit Practice -- Enforcement of Priority Treatment for
Buses on Urban Streets. National Cooperative Transit Research & Development Pro-
gram. TRB, National Research Council, Washington, DC. May 1982.

Rothenberg, Morris J. and Donald R. Samdahl; JHK & Associates. Evaluation of Pri-
ority Treatments for High Occupancy Vehicles. Federal Highway Administration, Jan-
uary 1981

Salter, R. J. and J. Shahi. Prediction of Effects of BusoPriority Schemes by Using
Computer Simulation Techniques. TRB, Transportation Research Record 718, 1979,

pp. 1-5.

Salter, R. J. and A. A. Memon. Simulation of a Bus-Priority Lane. TRB, Transporta-
tion Research Record 626, 1976, pp. 29-32.

Smith, Mark. Evaluation of Control Strategies For Bus Preemption of Traffic Si gnals--

177

Final Report. New Jersey Department of Transportation. In C00peration With Federal
Highway Administration, US. Department of Transportation. March 1985.

25. Sunkari, Srinivasa R., Phillip S. Beasley, Thomas Urbanik H, and Daniel B. Fambro.
Model to Evaluate the Impacts of Bus Priority on Signalized Intersections. Transporta-
tion Research Board Conference Preprints, Paper No 950490, Washington, DC,
U.S.A, January 1995.

26. The Transit/Highways Task Force. Bus Priority: Traffic Signal Preemption. Chicago
Area Transportation Study. December 1989.

27. TRAF User Reference Guide. FHWA-RD-92-060. US. Department of Transportation,
Federal Highway Administration. IVHS Research Division, McLean, Virginia, 1992.

28. Vincent, R. A., B. R. C00per, and K. Wood. Bus Actuated Signal Control at Isolated
Intersections: Simulation Studies of Bus Priority. Transport and Road Research Labo-
ratory, Crowthome, Berkshire, England, Report 814, 1978.

29. Wattleworth, Joseph A., Kenneth G. Courage, and Charles E. Wallace. Evaluation of
Bus- Priority Strategies on Northwest Seventh Avenue in Miami. TRB, Transporta-
tion Research Record 626, 1976, pp. 32-35.

30. Wood, K. Bus-Actuated Signal control at Isolated intersections--A Simulation Model.
Transport and Road Research Laboratory, Crowthome, Berkshire, England, Supple-
mentary Report 813, 1978.

31. Yamamoto, T. An Overview of the Current Japanese ATMS & ATIS. A lecture at Uni-
versity of Michigan, Ann Arbor. IVHS Lecture Series, 16 January, 1992.

"I I IIIIIIIIIII IIIII