1
Bus Signal Priority Based on GPS and Wireless Communications
Chen-Fu LiaoMinnesota Traffic Observatory
Department of Civil EngineeringUniversity of Minnesota
Gary DavisDepartment of Civil Engineering
University of Minnesota
ITS Advanced Transportation Technologies Seminar Series, September 25, 2007
Outlines
uBackground
uPhase I - Simulation Study
uPhase II – Prototype system development
2
What is Transit Signal Priority (TSP)?
“A tool to help make transit service more reliable, faster, and more cost effective. It has little impact on general traffic and is an inexpensive way to make transit more competitive with the automobile.”
- ITS America TSP planning and implementation handbook
“ Transit signal priority systems use sensors to detect approaching transit vehicles and modify signal timings to improve transit performance. ”
University of Minnesota
Priority vs. Preemption
uSignal priority modifiesnormal signal operation process to accommodate transit vehicles
uPreemption interrupts the normal signal operation process for special events (railroad crossing, emergency vehicles)
3
Transit Signal PriorityDetection Technologies
u Advantage: § Compatible with commonly used loop detectors§ Relatively reliable§ Does not require line-of-sight or visibility
u Disadvantage: § Require in-pavement loop detectors§ Prone to failure due to pavement flexing
u Example: v IDC LoopComm
u Implementation: v LA, Chicago
Inductive loop-based detection
Transit Signal PriorityDetection Technologies
u Advantage: § Widely used in U.S. for EVP, § Well tested for many years
u Disadvantage: § Require Line-of-sight clearance between emitter and
detectoru Example:
v 3M Opticom, Optronix/Tomar Strobecom, Novax Bus Plus
u Implementation: v Oakland, Tacoma, Portland, Vancouver
Light-based (Infrared) detection
4
Transit Signal PriorityDetection Technologies
u Advantage: § Emergency vehicle does not need additional equipment§ Facilitate inter-jurisdictional emergency response§ Does not require line-of-sight or visibility
u Disadvantage: § Not practical for transit signal priority (need additional
audible siren)§ False activation from building/car alarms
u Example: v Sonic Sonem 2000, EPS II
Sound-based detection
Transit Signal PriorityDetection Technologies
u Advantage: § Does not depend on line-of-sight or visibility
u Disadvantage: § Require RD tag installation at upstream curbside§ None directional vehicle information (with no RF tag
installation)u Example:
v TOTE (RF Tag), Econolite (EMTRAC)u Implementation:
v King County Seattle
Radio-based detection
5
Transit Signal PriorityDetection Technologies
Satellite (GPS)-based detectionu Advantage: § Does not depend on line-of-sight or visibility§ Can easily notify controller when vehicle has cleared
u Disadvantage: § Slow AVL polling rate§ Low/no GPS reception in urban canyon
u Example: v 3M GPS Opticom
u Implementation: v City of Edinburg, TX?
TSP Experiences in USAn Overview of Transit Signal Priority, ITSA
6
Transit Signal Priority (TSP)Background
u Bus signal priority has been implemented in several US cities (Seattle, Portland, LA, Chicago, St. Cloud, etc.)*
u Metro Transit performed tests on Lake Street using 3M Opticom technology (problem w/ nearside bus stops & trigger timing)
u Metro Transit contracted SEH Inc. to conduct TSP conceptual design along Northwest (Bottineau) corridor
u Most TSP deployments use sensors to detect buses at a fixed or preset distance.
* An Overview of Transit Signal Priority, ITS America 2002
Adaptive Bus Signal Priority Objectives
u Provide efficient and reliable bus transit service to traveling public.
u Reduce transit operation cost.u Use already installed GPS/AVL on bus & a
wireless communication based adaptive signal priority system with minimum impact on other traffic.
u Conduct traffic modeling and simulation to analyze and evaluate the possible impact.
7
u Bus transmits signal priority request to traffic controller based on its readiness not presence.
Bus Signal PriorityOur Approach
I am ready…
Bus Signal PriorityOur Approach (continue)
u Adaptive bus signal priority strategy using GPS/AVL and wireless communication technology.
u Provide conditional signal priority based on bus’s schedule adherence, speed, location and estimated dwell time at bus stop.
u Transmit priority request wirelessly from bus to intersection signal controller.
u Evaluate benefits and impacts of signal priority from simulation model.
8
Key Features of Our Approach
u Non-proprietary wireless communication (802.11x protocol)
u Use existing GPS/AVL system on busu Incorporate passenger count and bus lateness
to provide conditional priorityu Consider bus stop dwell time for signal priority
request when it is ready (Intersection arrival time forecast)
u Include controller phasing and timing status in priority request, forecast and decision making
Study Site – Franklin Ave. Minneapolis
27th Ave Dupont Ave. 3 miles, 22 signalizedintersections
Hiawatha LRTI-35W
To I-94& I-394 I-94
9
Simulation Model Development
u Intersection Capacity Analysis (Synchro)uData CollectionvSignal timing plan (from Minneapolis)vVolume and turning movements (Use Jamar)vTravel time (Vehicle probe)vBus dwell time & delay (20%) at intersectionvBus stop location (GIS from Metro Transit)
uUse AIMSUN Micro-Simulator
Simulation Model Development (con’t)
uCapacity Analysis
10
Simulation Model Development (con’t)
uNetwork Modeling
Simulation Model Development (con’t)
uNetwork Model Calibration08:00 ~ 08:15AM EB
Reaction Time = 0.75s, Reaction Time at Stop = 0.8s
0
50
100
150
200
250
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
Detector Station (42 intersections)
15-m
in V
olum
e
EB SIM EB BAL
Franklin Avenue Eastbound
11
Signal Priority Strategy
Intersection i
Intersection i-1 Intersection i+1
Bus Stop j Bus Stop j+1Eastbound
Westbound
Busde, j
de, i
Bus Stop kBus Stop k+1dw, k
dw, i
Bus Stop k-1Bus
Bus Stop j-1
Far side bus stop
Nearside bus stop
Signal Priority Request
delaybrb
jeaj TT
vd
T ++×
=467.1
,
bv : is bus speed, in MPH.
jed , : is the distance from the current bus location to bus stop j, in feet.
brT : is bus braking/stopping time.
delayT : is the traffic delay on bus route.
uNearside Bus StopAverage bus travel time to stop j:
12
Signal Priority Request
bv : is bus speed, in MPH.
: is the distance from the current bus location to intersection i, in feet.
delayT : is the traffic delay on bus route.
uFar side Bus Stop
Average bus travel time to intersection i:
delayb
iwai T
vd
T +×
=467.1
,
iwd ,
Poisson Distribution(Lambda=3)
0.000
0.050
0.100
0.150
0.200
0.250
1 3 5 7 9 11 13 15
Number of Passenger
Prob
abili
ty
Bus Dwell Time Calculation
uDwell Time Forecast at Bus Stop
[ ] boardingkkjB
dj TjtjttT ×−×= − )()()( 1λ
13
Signal Priority Acknowledgement
uRequest Time, TF (Time factor)vFirst come first serve
uBus Schedule Adherence, LF (Lateness Factor)vLF=0 if bus is ahead of schedule
uNumber of Passenger, PF (Passenger Factor)
Signal Priority Treatment: Green Extension
Cycle iPlan 1 2
Priority Phase
3 4 1
Cycle iPlan 2 2
Priority Phase
3 4 1
Priority Request
Green Extension
14
Signal Priority Treatment: Red Truncation
Cycle iPlan 1 2
Priority Phase
3 41
Cycle iPlan 2
2
Priority Phase
3 41
Priority Request
Early Green Extension
Signal Priority Treatment: Phase Insertion
Cycle iPlan 1 2
Priority Phase3 4 1
Cycle i+1 2Priority Phase
3 4 1
Priority Request
Cycle iPlan 2 2
Priority Phase3 4 1
Extension
2
Resynchronization
15
Signal Priority Strategy Modeling
GETRAM EXTAPI Interface
AIMSUNTraffic Simulator
Bus arrival/travel timeestimation model
Passenger arrival model at each stop
Intersection signalphasing/timing model
Signal priority, recovery, resynchronization model
Traffic Controller Model
Bus Signal PriorityController
Bus dwell time model at each stop
Bus SignalPriority Strategy
Read intersectionand bus stop data
Acquire bus(k) location,compare its distance to
next stop(j) and intersection(i)
dist2NextStop>dist2NextIntsc ?
Far side bus stop Nearside bus stop
Estimate time when buswill pass next signalized
intersection
t_estimate - sim_time<= system response time ?
No
Estimate bus dwell, arrivaltime at bus stop and travel
time to next intersection
dist2BusStop <=system response distance ?
Submit priority requestusing green extensionor red truncation for
intersection (i)
No
NoYes
Yes Yes
Exit extension phase and return back to phasep+1, where p is the phase before extension
Resynchronize signaltiming at next cycle
GETRAM APIInterface
TrafficSimulator
Is intersection (i)priority enabled ?
Yes
No Wait for next signalizedintersection with enabled
signal priority strategy
Signal Priority Control Flowchart
University of Minnesota
16
Bus Signal Priority Phase I Simulation Study Results
u Bus travel time reduction v AM-Peak: 12-15% v PM-Peak: 4-11%
u Bus delay time reductionv AM-Peak: 16-20% v PM-Peak: 5-14%
Overall Network Measures
Network Flow and Speed
19.818.0
13.497
18.128
0
5
10
15
20
25
Flow (1000 veh/h) Speed MPH
Ave
rage
Sta
tistic
s
AM Peak PM Peak
Network Travel and Delay Time
0:01:35
0:01:00
0:01:55
0:01:20
0:00:00
0:00:17
0:00:35
0:00:52
0:01:09
0:01:26
0:01:44
0:02:01
Avg. Travel Avg. Delay
Ave
rage
Tim
e
AM Peak PM Peak
17
Bus Travel Time and Speed (AM Peak)
AM Peak Bus Speed
9.1
10.4
9.2
10.7
8
8.5
9
9.5
10
10.5
11
No Priority With Priority
Ave
rage
Spe
ed M
PH
EB WB
AM Peak Bus Travel Time
0:16:27
0:19:53
0:17:30
0:19:02
0:14:24
0:15:07
0:15:50
0:16:34
0:17:17
0:18:00
0:18:43
0:19:26
0:20:10
No Priority With Priority
Trav
el T
ime
EB WB
Bus Travel Time and Speed (PM Peak)
PM Peak Bus Speed
9.2
7.7
8.38.3
6.5
7
7.5
8
8.5
9
9.5
No Priority With Priority
Ave
rage
Spe
ed M
PH
EB WB
PM Peak Bus Travel Time
0:21:580:22:41
0:19:39
0:21:03
0:14:24
0:15:50
0:17:17
0:18:43
0:20:10
0:21:36
0:23:02
0:24:29
No Priority With Priority
Trav
el T
ime
EB WB
18
MOE Analysis of Major Intersections
uHennepin – No significant changeuLyndale – 10% decrease in delay & stopsuNicollet – about 30%increase in travel
time and 47% increase in delayuChicago – about 5% decease in travel
time & delay timeuCedar - about 24% increase in travel time
and 30% increase in delay
Video Clip
With Priority
Without Priority
University of Minnesota
19
Phase II Study Overview
uDevelop wireless communication prototype using commercial off-the-shelf (COTS) product
u Implement & validate GPS/AVL and wireless communication based signal priority strategy
uField testing and validation
Phase II Prototype Systems
AntennaModemTx/Rx
AntennaModemTx/Rx
PriorityRequestServer(PRS)
PriorityRequest
Generator(PRG)
GPS/AVL
APC
I/O
TrafficController
OBURSU
VID, # Passenger, Stop Location,Schedule, Door Status, etc.
Ref: NTCIP 1211 V01.37 Protocol
Ex. Door Open/Close
GPS 18
20
Signal Priority System Hardware
RSUOBU
GPS Receiver
Controller Cabinet
NEMA Controller
AMD GX500256 MB RAM
Signal Priority Embedded System
Embedded System
Radio Modem
PC
Power Converter
21
Wireless Communications
uMinneapolis Wi-Fi Network
u5.9 GHz WAVE Radio
Minneapolis Wi-FiImplementation StartNovember, 2006
Phase OneAnticipated Completion................................................June, 2007
Phase TwoAnticipated Completion......................................September, 2007
Phase ThreeAnticipated Completion..........................................October, 2007
Phase FourAnticipated Completion..........................................October, 2007
Phase FiveAnticipated Completion......................................November, 2007
Phase SixAnticipated Completion......................................December, 2007
http://www.usiwireless.com/service/minneapolis/schedule.htm
30-40 nodes/sq mile
22
5.9 GHz Radio
Denso WAVE (Wireless Access in Vehicular Environment) Radio Module Prototype
u Bandwidth: 75 MHz (5.850 ~ 5.925 GHz) u 5.89 GHz (IEEE 178, control channel)u Channels: 10 MHz per channel (20 MHz optional)
Frequency (GHz)
5.85
0
5.85
5
5.86
0
5.86
5
5.87
0
5.87
5
5.88
0
5.88
5
5.89
0
5.89
5
5.90
0
5.90
5
5.91
0
5.91
5
5.92
0
5.92
5
5.82
5
5.83
0
5.83
5
5.84
0
5.84
5
US Spread Spectrum Allocation
Uplink
Downlink
IEEE DSRC Standard, http://www.ieee802.org
PublicSafety/Private
Public SafetyIntersectionsControl
ChannelPublicSafety/Private
PublicSafety/Private
US and Potential Mexican DSRC Allocation
IntersectionsControl Veh-Veh
Dedicated Public Safety
Short Rng ServiceMed Rng Service
Shared Public Safety/Private
PublicSafety/Private
PublicSafety
Veh-Veh
5.9 GHz DSRC Band Plan
23
Bus Stop & Intersection Geo-Database
Bus route #2Travel Pattern #1
Test Plan and Data Collection
uEquip a passenger vehicle for initial testing
uWireless communication reliability and latency data collection and analysis
uSignal priority algorithm verification and validation
uFinal testing and validation (one bus + one intersection)
24
Acknowledgement
§ ITS Institute, University of Minnesota.§ Don Sabania and Scott Tacheny at City of
Minneapolis.§ Gary Nyberg, John Levin at Metro Transit.§ Hun-Wen Tao & Wu-Ping Xin, CE students.§ Priya Iyer, EE student.§ Ted Morris & Pi-Ming Cheng at ITS Institute.§ Tony Juettner at Brown Traffic Products, Inc.
Thank You Very Much!
I am ready…