1 arterial performance measurement mathew berkow, michael wolfe, christopher monsere and robert l....

1Arterial Performance MeasurementArterial Performance Measurement

Mathew Berkow, Michael Wolfe, Christopher Monsere and Mathew Berkow, Michael Wolfe, Christopher Monsere and Robert L. BertiniRobert L. BertiniIntelligent Transportation Systems LaboratoryIntelligent Transportation Systems LaboratoryMaseeh College of Engineering and Computer ScienceMaseeh College of Engineering and Computer SciencePortland State UniversityPortland State University

Arterial Performance Measurement Using Arterial Performance Measurement Using Transit AVL and Signal System DataTransit AVL and Signal System Data


IntroductionIntroduction

Develop an automated way to report speeds, travel times and performance measures using: Existing ITS signal infrastructure Automatic Vehicle Locator (AVL) data

Expand PORTAL to include arterial data

Develop an automated way to report speeds, travel times and performance measures using: Existing ITS signal infrastructure Automatic Vehicle Locator (AVL) data

Expand PORTAL to include arterial data


PORTAL -- Regions ADUSPORTAL -- Regions ADUSPORTAL -- Regions ADUSPORTAL -- Regions ADUS


What’s in the PORTAL Database?What’s in the PORTAL Database?

Loop Detector DataLoop Detector Data20 s count, lane occupancy, 20 s count, lane occupancy, speed from 500 detectors speed from 500 detectors

(1.2 mi spacing) (1.2 mi spacing)

Incident DataIncident Data140,000 since 1999140,000 since 1999

Weather DataWeather DataEvery day since 2004Every day since 2004

VMS DataVMS Data19 VMS since 199919 VMS since 1999

DaysDaysSince July 2004Since July 2004About 300 GBAbout 300 GB

4.2 Million Detector 4.2 Million Detector IntervalsIntervals

Bus DataBus Data1 year stop level data1 year stop level data

140,000,000 rows140,000,000 rows

001703

WIM DataWIM Data22 stations since 200522 stations since 2005

30,026,606 trucks30,026,606 trucks

Crash DataCrash DataAll state-reported crashes All state-reported crashes

since 1999 - ~580,000since 1999 - ~580,000


What’s Behind the Scenes?What’s Behind the Scenes?

Database ServerDatabase ServerPostgreSQL Relational Database PostgreSQL Relational Database Management System (RDBMS)Management System (RDBMS)

StorageStorage2 Terabyte Redundant Array 2 Terabyte Redundant Array of Independent Disks (RAID)of Independent Disks (RAID) Web InterfaceWeb Interface


Selected Arterial Performance MeasuresSelected Arterial Performance MeasuresSelected Arterial Performance MeasuresSelected Arterial Performance Measures

TABLE 1 SELECTED ARTERIAL PERFORMANCE MEASURES

Metric Measurement Interval Location Maximum Speed

per Vehicle per Person

per Distance per Time

(cycle, 15 min, hour, day)

per Lane per Lane Group per Approach per Segment per Facility

per Area

Average Speed Speed Indexa Density Running Time Travel Time Travel Time Variance Average Delay Maximum Delay Queue Length Platoon Ratio Number of Stops Signal Failure Duration of Congestion per Day Number of Incidents per Day/Peak Period Duration Incidents per Event Nonrecurring Delay aRatio of average speed to posted speed.


Inspiration – Signal System Data OnlyInspiration – Signal System Data OnlyInspiration – Signal System Data OnlyInspiration – Signal System Data Only


Inspiration – Bus AVL System Data OnlyInspiration – Bus AVL System Data OnlyInspiration – Bus AVL System Data OnlyInspiration – Bus AVL System Data Only

I-5

I-5

I-205

US 26

SR 217

Powell Blvd.


Powell Blvd. Corridor StudyPowell Blvd. Corridor StudyPowell Blvd. Corridor StudyPowell Blvd. Corridor Study

Study Corridor

Four-lane arterial 3 Miles

East-West 50,000 ADT

TriMet’s Bus Dispatch System (BDS)

On-BoardComputer

Radio

DoorsLift

APC (Automatic Passenger Counter)

Overhead SignsOdometer

Signal Priority Emitters

Memory Card

RadioSystem

Garage PC’s

Radio AntennaGPS Antenna

Navstar GPS Satellites

Control Head

BDS DataR

ou

te N

o.

Ser

vice

D

ate

Lea

ve

Tim

e

Sto

p T

ime

Arr

ive

Tim

eB

adg

e

Dir

ecti

on

Tri

p N

o.

Lo

cati

on

ID

Dw

ell

Do

or

Lif

t

On

s

Off

s

Est

. L

oad

Max

Sp

eed

Pat

tern

D

ista

nce

X C

oo

r.

Y C

oo

r.

9 01NOV2001 8:53:32 8:49:15 8:53:28 285 0 1120 4964 0 0 0 0 0 21 41 10558.58 7644468 676005

9 01NOV2001 8:55:00 8:51:41 8:54:46 285 0 1120 4701 4 0 0 0 1 20 50 15215.05 7649112 676328

9 01NOV2001 8:56:22 8:52:00 8:55:08 285 0 1120 4537 36 3 0 6 0 26 34 15792.35 7649674 676220

Route Number Vehicle Number Service Date Actual Leave Time Scheduled Stop Time Actual Arrive Time Operator ID Direction Trip Number Bus Stop Location

Dwell Time Door Opened Lift Usage Ons & Offs (APCs) Passenger Load Maximum Speed on

Previous Link Distance Longitude Latitude

Probe Vehicle Location Over Time

39T

H

DIVISION

26T

H

33R

D

1ST

21S

T

FRANCIS

11T

H

LINCOLN

HARRISON

20T

H

MILW

AU

KIE

14T

H

13TH

3-D Speed Contour

Time

Distance

3-D Speed Contour

Conceptual Bus

Hypothetical Bus – run without stops or passenger activities (no dwell)

Pseudo Bus – travel at maximum achieved speed record within link

Modified Pseudo Bus – travel at maximum achieved speed with additional time spent at stops

0

0.5

1

1.5

2

2.5

3

7:07:00 AM 7:12:00 AM 7:17:00 AMTime

Dis

tan

ce

(m

ile

s)

Bus

Hypothetical Bus

Pseudo Bus

Modified Pseudo Bus

Ross IslandBridge

Thurs. Nov. 1, 2001

BU

S

Hyp

o

Pse

ud

o Mod

ified

Stop circle

0

0.5

1

1.5

2

2.5

3

1:04:00 PM 1:09:00 PM 1:14:00 PMTime

Dis

tan

ce (

mil

es)

Ross Island

Bridge

Test VehicleBusHypothetical BusPseudo BusModified Pseudo Bus B

us

Hyp

o

Pse

ud

oM

od

ified

Pse

ud

o

Veh

icl

e

Test Vehicles vs. Buses

5:12

3:52

5:38

3:40

4:40

7:38

10:24

6:54

5:07

9:10

0:27

0:34 0:34

0:28 0:28

0:33

0:48 0:48

0:44 0:44

3:00

4:00

5:00

6:00

7:00

8:00

9:00

10:00

11:00

Test Vehicle Bus Hypo Pseudo Modify

Tra

ve

l T

ime

(m

in)

0:10

0:15

0:20

0:25

0:30

0:35

0:40

0:45

0:50

Bri

dg

e T

rav

el

Tim

e (

min

)

Entire Corridor

On the Bridge

Min

Max

3rd Quartile

1st Quartile

Median

Test Vehicle vs. Pseudo

Test Vehicle Pseudo Test Vehicle Pseudo

Weekday Morning Peak 28.8 37.2 0.77 5:34 4:17 1.30

Weekend Midday Off-Peak 32.0 35.8 0.89 4:36 4:23 1.05

b ttMean Travel Time

EastboundMean Speed b speed




b tt

Mean Travel Time

WestboundMean Speed b speed

Ross Island Bridge

Four-lane traffic (2 lane on both side)

No shoulders Approaches are

bottleneck No interruptions No bus stops

Test Vehicle vs. Pseudo




b ttBridge: Mean Travel Time

EastboundBridge: Mean Speed b speed




b tt

Bridge: Mean Travel Time

WestboundBridge: Mean Speed b speed


This Project: Combine Signal and Bus AVLThis Project: Combine Signal and Bus AVLThis Project: Combine Signal and Bus AVLThis Project: Combine Signal and Bus AVL


Past and Ongoing EffortsPast and Ongoing EffortsPast and Ongoing EffortsPast and Ongoing Efforts

Using stop bar detectors to generate Using stop bar detectors to generate an arterial performance mapan arterial performance map Hallenbeck, Ishimaru, Davis, KangHallenbeck, Ishimaru, Davis, Kang LiangLiang

Performance based on summation of Performance based on summation of delay componentsdelay components Liu, MaLiu, Ma Skabardonis, GeroliminisSkabardonis, Geroliminis

Intersection/Signal PerformanceIntersection/Signal Performance Sharma, Bullock, BonnesonSharma, Bullock, Bonneson Smaglik, Bullock, SharmaSmaglik, Bullock, Sharma


DDDD

Signal System Data:Signal System Data:Portland’s Portland’s Detection Detection InfrastructureInfrastructure

Signal System Data:Signal System Data:Portland’s Portland’s Detection Detection InfrastructureInfrastructure

Data AggregationData AggregationCount Station Count Station

5 min5 minOther Detector Other Detector

15 min15 min7 Day Sample7 Day Sample


Case Study: Barbur Blvd. Speed MapCase Study: Barbur Blvd. Speed MapCase Study: Barbur Blvd. Speed MapCase Study: Barbur Blvd. Speed Map

SheridanHooker

Hamilton

3rdTerwilligerBertha

19th

I-5 Off-ramp

30th

Park & Ride

N


Will It Work? To The Video….Will It Work? To The Video….Will It Work? To The Video….Will It Work? To The Video….


Detectors at Barbur and BerthaDetectors at Barbur and BerthaDetectors at Barbur and BerthaDetectors at Barbur and Bertha


Density vs. OccupancyDensity vs. Occupancy

Density = number of vehicles per distance

Occupancy = percent of time with a vehicle on the sensor

Density = Occupancy X 1/(vehicle length + sensor length)

Density = number of vehicles per distance

Occupancy = percent of time with a vehicle on the sensor

Density = Occupancy X 1/(vehicle length + sensor length)

Density = 2 vehicles / 45 feet =.044

45’12’

6’

Density = .80 * 1 / (12 + 6) = .044

Occupancy = 80%


Flow vs. Occupancy: 5 Minute DataFlow vs. Occupancy: 5 Minute DataFlow vs. Occupancy: 5 Minute DataFlow vs. Occupancy: 5 Minute Data

Barbur & Hamilton, Northbound 2-12-07 to 2-20-07

0

500

1000

1500

2000

2500

3000

3500

4000

0 10 20 30 40 50 60 70 80

Occupancy (%)

Flo

w (

v/h

)


5 Minute Speed and Occupancy (at Hamilton)5 Minute Speed and Occupancy (at Hamilton)5 Minute Speed and Occupancy (at Hamilton)5 Minute Speed and Occupancy (at Hamilton)

0

10

20

30

40

50

60

70

80

90

0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00

Time

Sp

eed

(m

ph

)

0

10

20

30

40

50

60

70

80

Occ

up

ancy

(%

)

Speed

Occupancy


Generate Occupancy Map From Detector DataGenerate Occupancy Map From Detector DataGenerate Occupancy Map From Detector DataGenerate Occupancy Map From Detector Data


AM Peak Speed Map From Detector DataAM Peak Speed Map From Detector DataAM Peak Speed Map From Detector DataAM Peak Speed Map From Detector Data


Bus Geolocation DataBus Geolocation DataBus Geolocation DataBus Geolocation Data


ObjectiveObjective

Develop an automated way to report SpeedsTravel timesPerformance measures

Using Existing ITS signal infrastructureAutomatic Vehicle Locator (AVL) data

Develop an automated way to report SpeedsTravel timesPerformance measures

Using Existing ITS signal infrastructureAutomatic Vehicle Locator (AVL) data


1 Year’s Worth of TriMet Data1 Year’s Worth of TriMet Data

92 Bus Line 7,600 Stop Locations 273,469 Runs 140,751,486 Stops 50 Gigabytes of Data Extensive Network Coverage Opportunity to Evaluate Multiple Routes on

Same Arterial

92 Bus Line 7,600 Stop Locations 273,469 Runs 140,751,486 Stops 50 Gigabytes of Data Extensive Network Coverage Opportunity to Evaluate Multiple Routes on

Same Arterial


Time-Space DiagramTime-Space Diagram

Dis

tan

ce

Sig

nalize

d

Inte

rsecti

on

s


Detector Limited in SpaceDetector Limited in Space

Dis

tan

ce

Sig

nalize

d

Inte

rsecti

on

s


Bus Limited in TimeBus Limited in Time


Building on Powell Blvd. StudyBuilding on Powell Blvd. StudyBuilding on Powell Blvd. StudyBuilding on Powell Blvd. Study

Begin with limited signal system data.Begin with limited signal system data.

Gather Gather archivedarchived TriMet AVL data. TriMet AVL data.

Merge two data sources to examine synergies due to Merge two data sources to examine synergies due to data fusion.data fusion.

Use AVL data to calibrate influence areas from loops.Use AVL data to calibrate influence areas from loops.


Buses Inform Detector Readings – 2/12/07Buses Inform Detector Readings – 2/12/07Buses Inform Detector Readings – 2/12/07Buses Inform Detector Readings – 2/12/07


Congestion AlgorithmCongestion AlgorithmCongestion AlgorithmCongestion Algorithm

1. Calculate the slope of each bus segment (less dwell time) which is equal to the average bus speed

2. Identify a threshold speed that will be considered congestion

3. Find consecutive segments that experience travel speeds below the threshold ‘congested’ speed

1. Calculate the slope of each bus segment (less dwell time) which is equal to the average bus speed

2. Identify a threshold speed that will be considered congestion

3. Find consecutive segments that experience travel speeds below the threshold ‘congested’ speed

.

.

Location (x+1)

Location (x)

Location (x-1)

Segm

ent 1

Segment 2

..

.

Location (x+1)

Location (x)

Location (x-1)

Segm

ent 1

Segment 2

.


Midpoint Method Using 5-Minute DataMidpoint Method Using 5-Minute DataMidpoint Method Using 5-Minute DataMidpoint Method Using 5-Minute Data

Sig

nalize

d

Inte

rsecti

on

s


Adjust Influence Areas ManuallyAdjust Influence Areas ManuallyAdjust Influence Areas ManuallyAdjust Influence Areas Manually

Sig

nalize

d

Inte

rsecti

on

s


Bus Data Confirms AdjustmentBus Data Confirms AdjustmentBus Data Confirms AdjustmentBus Data Confirms Adjustment

Sig

nalize

d

Inte

rsecti

on

s


Reveals Gaps in DetectionReveals Gaps in DetectionReveals Gaps in DetectionReveals Gaps in Detection

Sig

nalize

d

Inte

rsecti

on

s


New Occupancy Map From Combined SourcesNew Occupancy Map From Combined SourcesNew Occupancy Map From Combined SourcesNew Occupancy Map From Combined Sources

Sig

nalize

d

Inte

rsecti

on

s


An Improvement Over Mid-Point MethodAn Improvement Over Mid-Point MethodAn Improvement Over Mid-Point MethodAn Improvement Over Mid-Point Method

Sig

nalize

d

Inte

rsecti

on

s


0

2

4

6

8

10

12

14

16

18

Actual Bus Signal Max Speed X1.25

Signal + Bus

Tra

vel

Tim

e (M

inu

tes)

Morning Peak Midday Off-Peak Evening Peak

Average Link Travel TimesAverage Link Travel TimesAverage Link Travel TimesAverage Link Travel Times

Ground truth


ObstaclesObstacles

System Signal Detector Very Limited Aggregation Access to Real Time Data Limited Detection & Spacing

Bus Access to Real Time Data

System Signal Detector Very Limited Aggregation Access to Real Time Data Limited Detection & Spacing

Bus Access to Real Time Data


Next StepNext Step

System Signal Detector Cycle level data (Gresham, OR – SCATS)

Bus TriMet Buses Can Be Probes Extensive Network Coverage Opportunity to Evaluate Multiple Routes on

Same Arterial

System Signal Detector Cycle level data (Gresham, OR – SCATS)

Bus TriMet Buses Can Be Probes Extensive Network Coverage Opportunity to Evaluate Multiple Routes on

Same Arterial


Bus FrequencyBus FrequencyBus FrequencyBus Frequency


CTRAN – 10 Second ReportingCTRAN – 10 Second ReportingCTRAN – 10 Second ReportingCTRAN – 10 Second Reporting


Next Steps: Gresham Burnside CorridorNext Steps: Gresham Burnside Corridor

Good detector density

Highly granular data (cycle-by-cycle)

Probe data available

No direct measurement of occupancy

16’ detectors No bus data

Good detector density

Highly granular data (cycle-by-cycle)

Probe data available

No direct measurement of occupancy

16’ detectors No bus data


AcknowledgementsAcknowledgements

TransPort Members FHWA: Nathaniel Price ODOT: Galen McGill PSU & OTREC (Local Matching

Funds) City of Portland: Bill Kloos, Willie

Rotich TriMet: David Crout, Steve Callas JPACT and Oregon Congressional

Delegation ITS Lab: John Chee, Rafael

Fernandez

TransPort Members FHWA: Nathaniel Price ODOT: Galen McGill PSU & OTREC (Local Matching

Funds) City of Portland: Bill Kloos, Willie

Rotich TriMet: David Crout, Steve Callas JPACT and Oregon Congressional

Delegation ITS Lab: John Chee, Rafael

Fernandez


Next Steps: Vancouver Mill Plain Blvd.Next Steps: Vancouver Mill Plain Blvd.

Detailed Data Available Do Not Have Sample Data Yet Possible Future Project

Detailed Data Available Do Not Have Sample Data Yet Possible Future Project

1 arterial performance measurement mathew berkow, michael wolfe, christopher monsere and robert l....

Documents