University of California Berkeley Phone: ( 510) 642- 4522
2105 Bancroft Way, Suite 300 Fax: ( 510) 642- 0910
Berkeley, CA 94720- 3830 http:// www. calccit. org
CALIFORNIA CENTER FOR INNOVATIVE TRANSPORTATION
CCIT Task Order 3
Corridor Management Plan Demonstration
Final Report – December 2006
Prepared By:
University of California at Berkeley’s
California Center for Innovative Transportation
in collaboration with
University of California at Irvine
and
System Metrics Group
For:
California Department of Transportation
Division of Traffic Operations
Division of Transportation Planning
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page i
Project Fact Sheet
Title: Corridor Management Plan Demonstration
Sponsor: Caltrans Division of Transportation Planning
Office of Policy Analysis and Research
Executing Organization: California Center for Innovative Transportation
2105 Bancroft Way
Berkeley, CA 94720
Phone: ( 510) 642- 4522 – Fax: ( 510) 642- 0910
Execution Period: 6/ 30/ 2003 – 6/ 30/ 2006
Contract Amount: $ 1,937,425
Principal Investigators: Dr. Samer Madanat, UC Berkeley
Dr. Hamed Benouar, UC Berkeley
Center Director: Dr. Hamed Benouar
Director, CCIT
Project Manager: Erik Alm, AICP
Senior Development Engineer, CCIT
Administrative Officer: Anne Crowe
Assistant Director, CCIT
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page ii
Acknowledgements
CCIT is pleased to recognize the hard work of the project team, as well as the agencies,
cities and counties that provided staff resources and expertise to this effort.
Project Team
CCIT
Dr. Hamed Benouar, Director
Erik Alm, AICP, Project Manager
Dr. Lianyu Chu, Research Engineer
Dr. Xuegang ( Jeff) Ban, Research Engineer
Dr. Koohong Chung, Postdoc Researcher
Angela Sugihara, Student Assistant
UC Irvine
Dr. R. Jayakrishnan
Hyunmyung Kim
Ji Young Park
Klayut Jintanakul
Pierre M. Auza
Tyler Bonstead
Jennifer Yoon
Chih- Lin Chung
System Metrics Group
Tarek Hatata
Tom Choe
Bill McCullough
Chris Williges
University of Minnesota/ Utah State Univ.
Dr. Henry Liu
Liang Ding
Caltrans
John Wolf, HQ Operations
Pat Weston, HQ Planning
Fred Dial, HQ Operations
Steve Hague, HQ Operations
Sarah Chesebro, HQ Operations
Al Arana, HQ Planning
Juliana Gum, D4 Operations
Cesar Casteneda, D6 Operations
Farid Nowshiravan, D12 Operations
Cambridge Systematics
Vassili Alexiadis
Krista Jeannotte
Braidwood Associates
Richard Braidwood
Wiltec
Moses Wilson
Agency Participants
I- 880 Corridor
• Metropolitan Transportation Commission
• AC Transit
• Alameda County Congestion Mgmt. Agency
• Bay Area Rapid Transit District
• Caltrans HQ and District 4
• City of Alameda
• City of Fremont
• City of Hayward
• City of Newark
• City of Oakland
• City of San Leandro
• City of Union City
• County of Alameda
• Port of Oakland
• Union City Transit
SR- 41 Corridor
• Caltrans HQ and District 6
• City of Fresno
• Fresno Council of Governments
I- 5 Corridor
• Caltrans HQ and District 12
• Orange County Transportation Authority
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page iii
Table of Contents
Executive Summary ............................................................................................................ v
List of Figures .................................................................................................................... ix
List of Tables ..................................................................................................................... xi
1. Introduction................................................................................................................... . 1
1.1 Corridor Management Plan and Demonstration ....................................................... 1
1.2 Current Progress of CMPD....................................................................................... 3
1.3 Organization of the Report ....................................................................................... 6
2. Stakeholder Needs and Corridor Selections ................................................................... 9
2.1 Stakeholder Needs and Involvement ........................................................................ 9
2.1.1 Stakeholder Needs ............................................................................................. 9
2.1.2 Stakeholder Involvement ................................................................................. 10
2.2 Corridor Selection and Arterial Coverage .............................................................. 12
2.2.1 Corridor Selections .......................................................................................... 12
2.2.2 Determination of Corridor Arterial Coverage.................................................. 15
2.3 Micro- simulation Based Modeling Technique ....................................................... 18
3. Data Needs and Collection............................................................................................ 21
3.1 Data Needs.............................................................................................................. 21
3.1.1 Corridor Description Data................................................................................ 22
3.1.2 Traffic Description Data .................................................................................. 27
3.2 Data Sharing and Field Collection.......................................................................... 31
3.2.1 Data Sharing..................................................................................................... 32
3.2.2 Field Data Collection ....................................................................................... 34
3.3 Performance Evaluation Data ................................................................................. 40
3.4 Simulation Model Data........................................................................................... 41
3.4.1 Network Coding Data Needs ........................................................................... 41
3.4.2 OD Matrices Estimation and Simulation Calibration ...................................... 43
3.4.3 Scenario Evaluation ......................................................................................... 45
3.5 Data Cleaning and Processing ................................................................................ 47
4. Corridor Performance Evaluation ................................................................................. 51
4.1 Performance Measures............................................................................................ 51
4.1.1 Mobility Measures ........................................................................................... 52
4.1.2 Safety Measures ............................................................................................... 57
4.1.3 Reliability Measure.......................................................................................... 59
4.1.4 Productivity Measure ....................................................................................... 60
4.2 Corridor Bottleneck Analysis and Verification ...................................................... 61
4.3 Current Performance............................................................................................... 64
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page iv
5. Baseline Simulation Model Development .................................................................... 68
5.1 Introduction............................................................................................................. 68
5.2 Microscopic Simulation Network Coding .............................................................. 68
5.2.1 Determine the Configuration of the Most Basic Inputs ................................... 69
5.2.2 Code the Skeleton Network ............................................................................. 77
5.2.3 Code Traffic Control........................................................................................ 83
5.2.4 I- 880 Network Coding ..................................................................................... 84
5.3 Calibration data preparation.................................................................................... 88
5.3.1 Existing Data.................................................................................................... 88
5.3.2 Preparation of Ramp Data................................................................................ 90
5.3.3 Preparation of Mainline Data........................................................................... 93
5.4 Model Calibration ................................................................................................... 94
5.4.1 Methodology .................................................................................................... 94
5.4.2 Calibration of Driving Behavior Models ......................................................... 96
5.4.3 Initial Calibration of Route Choice Model .................................................... 100
5.4.4 Dynamic OD demand estimation................................................................... 103
5.4.5 Network Performance Calibration and Validation ........................................ 119
5.5 Calibration results ................................................................................................. 124
5.5.1 Link flow GEH .............................................................................................. 124
5.5.2 Time- Dependent Section Travel Time........................................................... 127
5.5.3 Bottleneck Calibration Results ...................................................................... 128
5.6 Challenges of Micro- Simulation Model Development......................................... 133
6. Improvement Scenarios .............................................................................................. 135
6.1 Improvement Scenarios for SR- 41 ....................................................................... 135
7. Next Steps ................................................................................................................... 139
7.1 Lessons Learned ................................................................................................... 139
7.2 Recommendations for Future Study ..................................................................... 139
7.2.1 Stakeholder Involvement ............................................................................... 140
7.2.2 Finalizing the Three Simulation Studies........................................................ 140
7.2.3 Developing the Template for Corridor Management Plans........................... 140
7.2.4 Initial Implementation of Corridor Management Plans ................................. 142
Appendices..................................................................................................................... 143
Appendix A: System Metrics Group. Draft I- 880 Corridor Management Plan.
November 2006.....................................................................................................................
Appendix B: Ding, Liang. California SR- 41 Corridor Simulation Study – Calibration
Procedures and Results. 2005. .............................................................................................
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page v
Executive Summary
The Corridor Management Plan Demonstration ( CMPD) aims to develop a template for
corridor system management plans that can be used for both planning and operational
analysis. The primary objective of CMPD is to improve traditional corridor management
planning by incorporating detailed, multi- modal performance measurement and
evaluation, and innovative micro- simulation modeling techniques. The template will
help to address the problem of lost system productivity during congestion; it will also
help to create effective corridor management plans, thus improving statewide
transportation mobility, safety and productivity. CMPD represents the first attempt by
the California Department of Transportation ( Caltrans) to develop phased and integrated
corridor system management strategies by incorporating state- of- the- art operational
analysis into more traditional transportation planning processes.
This is the Final Report for CCIT Task Order 3 and is NOT intended to be a final report
of CMPD. Rather, it is a summary of the progress and status of CMPD to date ( June 30,
2006). The CMPD effort continues under CCIT Task Order 1015. Three corridors were
selected for the corridor management study during the course of CMPD: the SR- 41
corridor in Fresno, CA, the I- 880 corridor in the San Francisco Bay Area, and the I- 5
corridor in Orange County, CA. By the end of CCIT Task Order 3 ( June 30, 2006), the
project team had completed the simulation model development of SR- 41, finished part of
the I- 880 simulation model, and started to develop the simulation model for I- 5. The
project team had an ambitious research agenda under TO 3, and over the course of the
project encountered a number of challenges, both technical and institutional. As a result,
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page vi
the expectations of what could realistically be accomplished within the original schedule
and budget were adjusted. Stemming from the fact that the application of
microsimulation has not been attempted at this level of corridor planning analysis before,
the primary challenges encountered in Task Order 3 included:
• Data Collection issues: Unexpected gaps and network configuration issues
• Microsimulation model calibration issues: More adjustment iterations than expected
• No solid dynamic O/ D estimation method: Lack of a standard, accepted practice on
dynamic O/ D estimation required development of new method
Caltrans has approved a continuation research proposal with an updated schedule and
scope of work to accommodate the unanticipated gaps and new developments and to
enable the project team to complete the three studies ( CCIT Task Order 1015). More
importantly, this continuation research project under TO 1015 will enable the project
team to produce the template for corridor management plans by integrating all
experiences achieved from the three studies.
This report focuses on the methodologies developed and results obtained by the end of
Task Order 3 for CMPD, primarily covering the I- 880 and SR- 41 studies. The report
addresses the following topic areas:
Stakeholder Needs and Corridor Selections
Methodologies were developed for addressing stakeholder needs and selecting corridors
that were suitable for the study. Stakeholder needs were addressed by organizing
stakeholder meetings and corridor selections were conducted by balancing stakeholder
needs and resource limitations.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page vii
Data Needs and Collection
Data needs were defined for both performance evaluations and simulation model
development, and methodologies were developed to resolve data accuracy and
consistency issues. CMPD tasks needed corridor and traffic description data for corridor
level performance evaluations and simulation model development. To reduce the effort of
data collection, data sharing among stakeholders was critical. In addition, data pre-processing
( especially for traffic description data) was crucial to remove erroneous data
and maintain data consistency.
Corridor Performance Evaluation
Various mobility, safety, reliability and productivity measures have been defined and
used for corridor level performance evaluations. The evaluation results showed that the I-
880 corridor experiences heavy congestion and incidents during both AM and PM peak
hours, while the SR- 41 corridor only had light congestion during peak periods.
Baseline Simulation Model Development
Detailed methodologies and procedures were developed for baseline simulation model
development, including network coding, origin- destination demand estimation,
calibration, and model fine- tuning. The results showed a general match between observed
data and simulation data.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page viii
Evaluation Scenario Generation
Based on planned and programmed projects from Caltrans Headquarters and local
Districts ( i. e., Districts 6), improvement scenarios were generated for the SR- 41 corridor
( including short-, medium-, and long- term scenarios). Preliminary scenario development
began for the I- 880 corridor, but was not completed by the end of Task Order 3.
Finalizing and evaluation of these scenarios will be conducted in the continuation CCIT
Task Order 1015.
Next Steps
CCIT Task Order 1015 will complete the three studies, and more importantly, complete
the template for development of corridor management plans, as well as provide technical
assistance to Caltrans staff on these corridor management planning techniques.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page ix
List of Figures
Figure 1.1 Caltrans System Management Pyramid
Figure 1.2 Overview of Methodology for Corridor Management Plan
Figure 2.1 I- 880 Corridor in the Bay Area
Figure 2.2 SR- 41 Corridor in Fresno
Figure 2.3 Arterial Coverage for I- 880 Corridor
Figure 2.4 Arterial Coverage for SR- 41 Corridor
Figure 2.5 Output from Toolbox for I- 880 Study
Figure 3.1 Snapshot of Incident Report in the Bay Area ( CHP)
Figure 3.2 Field inspection Results for I- 880 vs. SR- 84 Interchange
Figure 3.3 SR- 41 Freeway Volume Data Collection Locations
Figure 3.4 Tach run Check Point Locations for SR- 41 Corridor
Figure 3.5 I- 880 Freeway Segments for Travel Time Collections
Figure 3.6 Planned and Programmed Improvement Strategies for SR- 41 Corridor
Figure 3.7 Excel Template for Data Consistency Checking
Figure 3.8 Method for Data Consistency
Figure 4.1 VMT for NB I- 880 on March 1, 2005
Figure 4.2 Truck VMT for NB I- 880 on March 1, 2005
Figure 4.3 Delay for NB I- 880 on March 1, 2005
Figure 4.4 Speed Variation for NB I- 880 at 4: 00 PM on March 1, 2005
Figure 4.5 Incidents on NB I- 880 for March of 2005
Figure 4.6 Average Daily Accidents for I- 880 ( Both Directions)
Figure 4.7 Travel Time Reliability for I- 880 NB
Figure 4.8 Lost Productivity for I- 880 NB on March 1, 2005
Figure 4.9 Bottleneck Plot for I- 880 NB
Figure 4.10 Average Speed Bottleneck Plot
Figure 4.12 Speed Contour for SR- 41 for PM Peak ( NB)
Figure 4.13 Speed Contour for SR- 41 for PM Peak ( SB)
Figure 5.1 The Study Network
Figure 5.2 Paramics Network
Figure 5.3 Mainline Data Adjustment
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page x
Figure 5.4 Model Calibration
Figure 5.5 Incorporating Traffic Simulations into Conventional Planning Procedure
Figure 5.6 Subarea Analysis on a Static Network by TRANSCAD
Figure 5.7 Network Data Conversion from Static Network to Dynamic Network
Figure 5.8 Framework for OD Demand Matrix Estimation
Figure 5.9 PARAMICS Seed OD Table Estimation
Figure 5.10 Bi- level Dynamic OD Demand Estimation Process
Figure 5.11 Speed Plot from Observed Data ( NB)
Figure 5.12 Speed Plot from Simulation Results ( NB)
Figure 5.13 Speed Plot from Observed Data ( SB)
Figure 5.14 Speed Plot from Simulation Results ( SB)
Figure 6.1 Short- Term Improvement Strategies for the SR- 41 Corridor
Figure 6.2 Medium- Term Improvement Strategies for the SR- 41 Corridor
Figure 6.3 Long- Term Improvement Strategies for the SR- 41 Corridor
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page xi
List of Tables
Table 2.1 Corridor Scope Overview
Table 3.1 Major Data Providers and Available Data
Table 3.2 Locations for Link Counts and Turning Volumes for SR- 41 Corridor
Table 3.3 Data Needs for Performance Evaluation
Table 3.4 Data Needs for Network Coding
Table 3.5 Data Needs for OD Estimation and Simulation Calibration
Table 3.6 Arterial Data Collection for Northbound of I- 880 Corridor
Table 3.7 Arterial Data Collection for Northbound of I- 880 Corridor
Table 4.1 SR- 41 NB Travel Times
Table 4.2 Delay of I- 880 Corridor
Table 5.1 Vehicle Group Definition and Percentage
Table 5.2 HOV Percentage Estimation
Table 5.3 Geometry Data for Network Coding
Table 5.4 General Description of Two Ramp Data Sources
Table 5.5 Wisconsin DOT Freeway Model Calibration Criteria
Table 5.6 GEH Statistic for NB
Table 5.7 GEH Statistic for SB
Table 5.8 Section Travel Times for NB
Table 5.9 Section Travel Times for SB
Table 5.10 Summary of AM Simulation Calibration Results
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 1
1. Introduction
This report presents the methodologies developed and results obtained so far ( by the end
of FY 05- 06) for Corridor Management Plan Demonstration ( CMPD), the first phase of
developing Corridor Management Plan ( CMP). CMP represents the first attempt by
California Department of Transportation ( Caltrans) to develop phased and integrated
corridor system management strategies by incorporating state- of- the- art operational
analysis into more traditional transportation planning processes. Although conducted in
California, the CMP research is the first of its kind in integrating operational analysis
( such as dynamic traffic network modeling and traffic simulation) into the traditional
planning process for urban and congested corridors.
1.1 Corridor Management Plan and Demonstration
Caltrans has been developing system management strategies with the aim of managing
the California’s transportation system more effectively and efficiently. As depicted in the
following system management pyramid in Figure 1.1, system monitoring and evaluation
are the basic foundation upon which other strategies are built. These strategies range from
maintenance and preservation to system expansion and completion. This strategy
pyramid represents comprehensive management strategies, including both planning and
operations, for a maturing transportation system in which infrastructure expansion,
although still important, is not the only strategy to address mobility and safety needs of
Californians. Operational improvements and strategies, on the other hand, can help to
make the most efficient use of existing transportation system. Therefore, Caltrans
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 2
recognizes the emerging needs of development systematic management strategies to get
the most out of current system and to evaluate potential facility expansion if necessary.
Figure 1.1 Caltrans System Management Pyramid
Guided by the system management concept in Figure 1.1, Caltrans investigated
developing a Corridor Management Plan ( CMP) which focuses on developing
management strategies on state- wide corridors. Upon success, CMP can be a backbone
for developing more comprehensive system management strategies for an entire region.
CMP can be defined as:
A Corridor Management Plan is a document that identifies the recommended system
management investment strategies for a State Highway facitility within the context of the
full multi- modal corridor. The strategies and their phased implementation are
recommended based on the comprehensive performance assessment of the State Highway
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 3
facility and its interactions with other modes, and should maximize the return on the
recommended investments.
To demonstrate CMP, Caltrans initialized the Corridor Management Plan Demonstration
( CMPD) as the first phase to develop appropriate management strategies on selected
corridors and demonstrate the feasibility of developing CMP. CMPD aims to develop a
template of corridor system management plans that can be used for both planning and
operations of Caltrans. The primary objective of CMPD is to improve traditional corridor
management planning by incorporating detailed, multi- modal performance measurement
and evaluation and operational analysis with state- of- the- art analysis and modeling
techniques ( e. g. microscopic traffic simulation tools). The template will help to address
the problem of lost system productivity during congestion; it will also help to create
effective corridor management plans, thus improving statewide transportation mobility,
safety and productivity. Led by California Center for Innovative Transportation ( CCIT)
under University of California, Berkeley ( UCB), partners involved in CMPD include
regional, local and congestion- management agencies. Researchers from the University of
California campuses ( Berkeley and Irvine) and outside the state ( Utah State University
and the University of Minnesota) are supported by System Metrics Group, Inc. ( SMG)
and other private transportation consultancies.
1.2 Current Progress of CMPD
Three corridors have been selected for the corridor management study during the course
of CMPD, reflecting a joint effort by Caltrans and multiple state and local level
stakeholders, supported by the project team. The first corridor is the SR- 41 network,
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 4
located in Fresno, CA. It is a mid- size network with many arterials, especially two
parallel streets. This corridor currently experiences only light congestion and the purpose
of the study is to provide a proof- of- concept of developing a CMP, particularly the
feasibility of incorporating detailed operational analysis ( e. g. dynamic traffic network
modeling and micro- simulation) into traditional planning in a corridor level, an effort that
has rarely been conducted before. Researchers from Utah State University and the
University of Minnesota have completed the base simulation model development.
Evaluating the benefits of selected short-, medium, and long- term operational
improvements is on- going.
The second corridor is I- 880, an inter- regional and multi- modal corridor located east of
San Francisco. It is a 34- mile urban freeway route with parallel arterials and cross roads,
including 143 metered lanes, 157 actuated traffic signals, and 25 fixed- time signals.
Transit and inter- modal facilities ( e. g., sea port and airport) serve as alternative modes of
the corridor. It currently experiences heavy congestion for both AM and PM peaks and
multiple traffic incidents for almost any given “ typical” day. Therefore, the I- 880 corridor
study is a “ full- scale” case for demonstrating system management concepts within the
CMP. While the I- 880 performance assessment was completed, the microsimulation
model to evaluate improvement scenarios was not completed before the expiration of
Task Order 3. In Task Order 1015 Braidwood Associates ( BA) will continue finalizing
the simulation study previously conducted by University of California, Irvine ( UCI).
The third is the I- 5 corridor in Orange County. It is of similar size to the I- 880 network,
and is somewhat less congested. The research team completed a highway- only
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 5
microsimulation for I- 5 and is working on incorporating arterials into the model as part of
Task Order 1015. The I- 5 corridor will serve as a test case for verifying methodologies
developed in I- 880 and SR- 41 studies, and will be covered in the report for Task Order
1015.
These three studies have demonstrated both the challenges and benefits of integrating
planning and operations into traditional corridor management. As a result, the concept of
corridor system management plans received significant attention from stakeholders and
decision makers at all levels, despite the three demonstration corridor studies not being
completed by the end of FY 05/ 06 ( June 30, 2006). The project team had an ambitious
research agenda; over the course of the project the team encountered a number of
challenges, both technical and institutional. As a result, the expectations of what could
realistically be accomplished within the original schedule and budget were adjusted.
Stemming from the fact that the application of microsimulation has not been attempted at
this level of corridor planning analysis before, the primary challenges encountered in
Task Order 3 included:
• Data Collection issues: Unexpected gaps and network configuration issues
• Microsimulation model calibration issues: More adjustment iterations than
expected
• No solid dynamic O/ D estimation method: Lack of a standard, accepted practice
on dynamic O/ D estimation required development of new method
Caltrans has approved a continuation research proposal with an updated schedule and
scope of work to accommodate the unanticipated gaps and new developments and to
enable the project team to complete the three studies ( CCIT Task Order 1015). More
importantly, this continuation research project under TO 1015 will enable the project
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 6
team to produce the template for corridor management plans by integrating all
experiences achieved from the three studies. Therefore, this report also sets up the basis
for the continuation efforts of CMP development.
1.3 Organization of the Report
This report is NOT intended to be a final report of CMPD. Rather, it is a summary of the
progress and status of CMPD to date ( i. e., June 30, 2006). In particular, it summarizes the
methodologies that have been developed for CMP and results obtained from SR- 41 and I-
880 corridor management studies. Due to the majority efforts put on the I- 880 corridor by
the project team, this report will focus primarily on the I- 880 study; the developed
methodologies, however, are suitable for both SR- 41 and I- 880 corridors in most cases.
The results of the I- 5 corridor will not be covered in this report; as previously noted these
results will be provided in a separate report as part of the findings of CCIT Task Order
1015.
This report contains a main body summarizing the overall methodologies and results, as
well as appendices of a report by SMG on developing the corridor management plan for
the I- 880 study ( Appendix A) and a master thesis documenting the development of the
simulation model for the SR- 41 corridor ( Appendix B).
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 7
Figure 1.2 provides a high- level overview of the methodologies ( procedures) to conduct
the CMPD. The main body of the report is organized as follows:
• Stakeholder Needs and Corridor Selections ( Section 2): Describes the
methodology for addressing stakeholder needs and selecting corridors that are
suitable for the study;
• Data Needs and Collection ( Section 3): Describes data needs and collection, as
well as data accuracy and consistency issues, for both performance evaluations
and simulation model development;
• Corridor Performance Evaluation ( Section 4): Discusses the methods and
results of performance evaluations of selected corridors;
• Baseline Simulation Model Development ( Section 5): Details the methodology
for developing the baseline simulation models;
• Evaluation Scenario Generation ( Section 6): Presents the corridor management
improvement scenario generation for the SR- 41 and I- 880 corridors; and
• Next Steps ( Section 7): Summarizes activities recommended be conducted in the
next steps ( i. e., during the extension of CMPD).
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 8
Figure 1.2 Overview of Methodology for Corridor Management Plan
Stakeholder Needs and Corridor
Selections
( Section 2)
Data Needs and Collection
( Section 3)
Corridor Performance Evaluation
( Section 4)
Baseline Simulation Model
Development
( Section 5)
Evaluation Scenario Generation
( Section 6)
Next Steps Study
( Section 7)
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 9
2. Stakeholder Needs and Corridor Selections
This section describes the methodologies for addressing stakeholder needs regarding
corridor management planning and how the studied corridors were selected.
2.1 Stakeholder Needs and Involvement
Properly addressing stakeholder needs and their regular involvement are critical to the
success of any transportation related research project. This is particularly true for CMPD
since it is an inter- regional project involved with Caltrans Headquarters, local districts,
local transportation authorities ( such as Metropolitan Transportation Commissions ( MTC)
in the Bay Area), cities, and transit agencies.
2.1.1 Stakeholder Needs
To adequately address stakeholder needs, the project team started with preliminary
performance assessments for candidate corridors. After the initial assessment was
complete, the team organized multiple stakeholder meetings to discuss the stakeholder
needs and resolve possible conflicts of CMPD with previous and existing planning and
project development efforts. During these meetings, findings from performance
assessments were presented to stakeholders and their needs were discussed extensively
and summarized by the project team. Following is a summary of general stakeholder
needs in terms of corridor management plan from those stakeholder meetings:
( 1) Seeking short- term ( within five years) and/ or medium- term ( five to 15 years)
operational improvements to reduce traffic congestion in the corridor;
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 10
( 2) Seeking short- term ( within five years) and/ or medium- term ( five to 15 years)
operational improvements to improve corridor safety;
( 3) Seeking short- term ( within five years) and/ or medium- term ( five to 15 years)
operational improvements to resolve corridor- level environmental issues;
( 4) Seeking long- term ( 15 years and after) capital investment and/ or land use
strategies to better accommodate increasing traffic demands.
Discussions with stakeholders confirmed that different types of corridors require varied
corridor management planning strategies. A highly urban and congested corridor ( e. g.,
the I- 880 corridor in the Bay Area) requires more analysis and modeling to evaluate
short- term and/ or medium- term operational improvement strategies ( i. e., item ( 1) in the
above list). On the other hand, a less urban and less congested corridor ( e. g. SR- 41 in
Fresno) may require less modeling and more emphasis on safety and environmental
issues ( i. e., item ( 2) or ( 3) in the above list). The project team thus needs to develop
methodologies that can accommodate all these stakeholder needs.
2.1.2 Stakeholder Involvement
Appropriate stakeholder involvement is necessary to better address stakeholder needs. It
helps identify gaps of current corridor management plans and facilitates the data
collection efforts of the project team. In summary, the major issues that were discussed
through stakeholder involvement are listed as follows:
• Stakeholder needs - the major motivation of conducting CMPD;
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 11
• Related previous and ongoing projects - Provides the current state of the
corridor and necessary documentations for the project team to identify gaps in
terms of corridor management planning;
• Corridor selections and arterial coverage – Defines the scope of the study, and
• Data needs and sharing - Addresses data needs of the project team and
availability of data among stakeholders. A large amount of data was needed by the
project team ( see Section 3). The data was obtained from PeMS ( Performance
Measurement System) and other dedicated sources ( e. g., CHP and TASAS). Some
data ( e. g., signal timing and detailed corridor descriptions) was provided by
stakeholders. This can significantly reduce the project team’s efforts for data
collection and reduce data collection costs.
In CMPD, the stakeholder involvement was conducted by organizing regular stakeholder
meetings for each of the corridor studies. For example, for the I- 880 study alone, more
than 12 formal meetings and 20 presentations were conducted as part of the stakeholder
participation efforts to date. For each stakeholder meeting, research findings and
recommendations were provided by the means of presentations and handouts. At the
same time, meeting minutes were developed and feedback from stakeholders was colleted
and integrated into the project. In addition to these meetings, regular contacts with
stakeholders have been maintained by the project team, especially for data availability
and collection purposes.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 12
2.2 Corridor Selection and Arterial Coverage
2.2.1 Corridor Selections
The corridor selection for each involved District started with the list of candidate
corridors provided by the District during the stakeholder meetings. Two criteria were
initially used by the project team for corridor selections: potential corridor benefits and
detection availability. A rough assessment of potential benefits was conducted for each
candidate corridor. The assessment provided indicators based on existing and projected
future traffic conditions, preliminary simulation results, and the inventory of Traffic
Management System ( TMS) elements currently installed. In addition, one day’s detection
availability was collected from PeMS to serve as the second criterion since several
performance measures ( e. g., reliability, productivity, delay, etc.) need detailed detection
data. Additional criteria were then developed during stakeholder meetings. The identified
criteria for D4 are summarized as follows:
• Primary Criterion
o The selected corridor should not conflict with other ongoing studies.
• Secondary Criteria
o The selected corridor should serve significant inter- regional travel.
o The selected corridor should be multi- modal in nature.
o Congestion on the selected corridor should be high and projected to grow.
o The selected corridor should have a high potential for benefits.
• Final Criteria
o The selected corridor should have a good detection data.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 13
o The selected corridor should serve the goods movement industry.
By evaluating 24 candidate corridors in D4, the I- 880 corridor in the Bay Area ( from
Grand Ave in Oakland to the SR- 237 interchange in Milpitas) was selected. For District 6,
the SR- 41 corridor in Fresno ( from E Friant Rd to the SR- 99 interchange) was initially
selected by District staff with no extra selection effort by the project team. Figures 2.1
and 2.2 below show maps of these two corridors. In these two figures, the “ star” symbols
indicate the starting and ending points of the selected corridors.
Figure 2.1 I- 880 Corridor in the Bay Area
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 14
Figure 2.2 SR- 41 Corridor in Fresno
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 15
2.2.2 Determination of Corridor Arterial Coverage
The arterial coverage for each selected corridor was also discussed during the stakeholder
meetings. For the I- 880 corridor, a major arterial was selected to be included in the
modeling extent based on the following criteria:
o Whether it could be a major diversion possibility under congestion,
o Whether the arterial plays a significant role in distributing freeway traffic to
various ramps ( and vice versa to the city streets),
o If the dynamic flows on an arterial are more affected/ influenced by the freeway
conditions ( more than the larger network conditions), and
o Resource constraints on collecting intersection count data and calibrating properly
To some extent, determination of the arterial coverage reflects tradeoffs between
stakeholder needs and resource constraints of the project. On the one hand, many
stakeholders, especially the local transportation management agencies, may prefer to
have a wider arterial coverage so that more improvement strategies can be evaluated and
tested. On the other hand, because of the time and budget constraints, the project team
has limitations on including too broad a network of arterials into the corridor network. In
particular, micro- simulation tends to have restrictions on the size of the studied network
in order to run the simulation efficiently.
One lesson learned from conducting current CMPD studies is that stakeholder meetings,
especially those involving a large number of stakeholder groups are essential at the
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 16
beginning of the study to get early consensus on arterial coverage. The addition of new
arterials to the I- 880 corridor network after the simulation team had finished the network
coding caused delay to the original schedule.
The finalized arterial coverage for the two corridors represents a balance of these issues.
The arterial coverage is depicted in Figure 2.3 and Figure 2.4 respectively for the I- 880
and SR- 41 corridors, representing the simulation networks developed in Paramics for the
two corridors.
Figure 2.3 Arterial Coverage for I- 880 Corridor
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 17
Figure 2.4 Arterial Coverage for SR- 41 Corridor
In summary, Table 2.1 provides an overview of the two corridor studies.
Table 2.1 Corridor Scope Overview
I- 880 Corridor SR- 41 Corridor
Freeway scope The Grand Street interchange in
Oakland to the SR- 237 interchange in
Fremont
Abosolute Postmiles:
NB: 8.073 to 45.027
SB: 45.027 to 8.105
E Friant Rd to SR- 99
interchange
Absolute Postmiles:
NB: 121.573- 137.573
SB: 137.573- 121.573
Study period AM peak hour 6: 30 am – 9: 30 am *
PM peak hour 3: 30 pm – 6: 30 pm 4: 00 pm – 7: 00 pm
Number of zones 168 210
Number of ramp meters 56 ( 143 metered lanes) 15
Number of pre- timed traffic signals 25 0
Number of actuated traffic signals 157 99
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 18
Transit BART, Inter- city trains, AC Transit N/ A
Inter- modal Facility Oakland airport, Oakland seaport N/ A
*: Caltrans D6 staff is conducting the AM study based on the methodologies developed in CMPD
2.3 Micro- simulation Based Modeling Technique
To identify gaps of traditional corridor management planning practices, the project team
reviewed relevant documents, including guidelines and reports of related previous and
ongoing projects. By reviewing these documents, it is clear that traditional corridor
management planning in Caltrans District 4 and District 6 uses the four step modeling
process ( i. e., trip generation, trip distribution, modal split, and traffic assignment). The
applied modeling tool is mainly travel demand models such as TP+, EMME2, TransCAD,
etc. Although more advanced modeling techniques such as micro- simulation have been
applied in a project by project basis ( e. g., evaluation of intersection timing plan and ramp
metering control algorithm), they have not yet been widely used in the corridor level. The
potential advantages of micro- simulation models are summarized as follows:
• Dynamic, implying that the dynamic characteristics of traffic, which is critical
especially for evaluating short- term and medium- term operational improvements,
can be properly captured;
• Capable of evaluating multiple ITS- related strategies, including incident
management, traveler information, traffic signal control and coordination, and
ramp metering control; and
• Adding the perceived objectivity of modeling and analysis to the perceived
subjectivity of staff experience and judgment to identify problem areas of the
corridor.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 19
However, micro- simulations suffer from drawbacks such as intensive efforts for model
development and extensive data needs. Therefore, to identify the proper modeling
technique to better address stakeholder needs and the project objectives for CMPD, the
project team used a toolbox recently developed by the Federal Highway Administration
( FHWA) called Traffic Analysis Toolbox Volume II: Decision Support Methodology for
Selecting Traffic Analysis Tools1. The following figure depicts the results of using this
toolbox for selecting appropriate tools for the I- 880 study. For the SR- 41 study, micro-simulation
was pre- selected as the modeling tool by the District Staff. The simulation
network for SR- 41 had been developed by staff at District 6 in Paramics before the
corridor was selected in CMPD.
Figure 2.5 Output from Toolbox for I- 880 Study
1 http:// www. ops. fhwa. dot. gov/ trafficanalysistools/ tat_ vol2/ index. htm
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 20
The micro- simulation technique was recommended by the toolbox for the I- 880 study.
The project team adopted Paramics as the micro- simulation package recommended by
Caltrans Traffic Operations. The expectation is that by using micro- simulation modeling
the project team will be able to evaluate various ITS related operational improvement
strategies while considering detailed dynamics of traffic flow in the studied corridors.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 21
3. Data Needs and Collection
Compared with traditional corridor planning using travel demand models, CMPD
requires significantly much richer dataset for both performance evaluations and
simulation studies. This section addresses the data needs and collection issues in CMPD.
We start with descriptions of basic data categories and major data providers. We then
discuss in more detail the specific data needs and collection for performance evaluation
and simulation model development.
3.1 Data Needs
The data needed for conducting CMPD can be grouped into the following two categories:
• Corridor Description Data – Provides a general description of the corridor geometry,
traffic control and detection facilities, demand characteristics, land use, modal
services, and environmental factors.
• Traffic Description Data – Describes the basic traffic flow characteristics such as
origin- destination ( OD) matrices, vehicle type mix, volume, occupancy, speed, travel
time, traffic incidents, etc.
Each of the two categories is addressed in more detail as follows.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 22
3.1.1 Corridor Description Data
Corridor description data provides the geometry of the corridor ( including both freeways
and arterials), major transportation management systems and traffic detection devices,
neighboring land use, demand characteristics, and environmental scan. Although it may
vary for different corridors, in general a complete corridor description should include the
following elements:
Roadway Geometry
This element explains the coverage of the corridor, including the freeways, major cross-routes
and parallel facilities, as well as the physical dimensions of the facilities within it.
Generally, the roadway geometry should include the following components:
• The extent of the corridor ( major freeway), including total lane miles and centerline
miles as well as beginning and ending postmiles ( e. g., for the I- 880 corridor, it begins
at the SR- 237 interchange and ends at Grand Ave).
• Key geometric information, such as the number of lanes along the corridor in general
or major sections, the presence of high- occupancy vehicle ( HOV) facilities, typical
lane and shoulder width, and type of medians.
• Major structures, such as bridges and overpasses.
• Key interchanges and/ or intersections with other state highways and major arterials
( e. g., the I- 238 interchange of the I- 880 corridor).
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 23
• Relevant stakeholder information, such as cities and counties, regional planning
agencies, and air basins included in the corridor. For example, the I- 880 corridor
covers five cities, Alameda and Santa Clara counties, one regional planning agency
jurisdiction and one air basin.
• Detailed arterial geometric description. This includes the major cross routes and
parallel arterials. The definition of “ major arterial” should be discussed with local
stakeholders ( see Section 2) to reach consensus. Generally, routes included in the
regional planning model are a good starting point for these discussions. Once major
arterials are identified, corridor descriptions should provide geometric information of
arterials regarding number of lanes, presence of turn lanes, signalization and
channelization, and major areas served.
Traffic Management System ( TMS) Elements
TMS elements provide a summary of existing and planned traffic management and
control devices within the coverage of the corridor. Major components include:
• Vehicle detection and data availability. This includes the type of detections, e. g., loop
detectors or radar or probe vehicles, and its configuration such as locations of the
detectors, data formats, etc. This also provides how data can be achieved, e. g.,
through PeMS and/ or Advanced Transportation Management System ( ATMS).
• Ramp metering. This includes locations of the meters and the extent of deployment
( e. g., every ramp, every other ramp, a portion of the corridor, etc.), as well as the type
of the ramp metering ( e. g., pre- timed or traffic responsive or adaptive) and control
design ( one or two vehicles per green, HOV bypasses). In order for the micro-
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 24
simulation model to work properly, the detailed ramp metering control logic is also
needed. The logic could be rather different for different corridors and should be
provided by corresponding Caltrans local districts.
• Number and location of Changeable Message Signs ( CMS) and Closed- Circuit
Television Cameras ( CCTV).
• Availability of Highway Advisory Radio ( HAR)
• Signal timing plans and coordination for arterial intersections, as well as coordination
of ramp metering and arterial traffic signals ( if any).
Origins and Destinations of Travel
This element describes the primary origins and destinations of traffic that uses a given
corridor. This includes Origin- Destination ( OD) pairs within the corridor, from within the
corridor to outside the corridor, and both outside with the corridor used as a major route.
The OD information can be provided by the regional planning agency and/ or Caltrans
local district offices. For instance, for the I- 880 corridor, the Metropolitan Transportation
Commissions ( MTC) developed a region- wide model that includes a base calibration year
and several projection years. For the SR- 41 corridor, Caltrans District 6 has the planning
model that defines the origins and destinations for the studied corridor.
Land Use
This element provides an overview of current land- use and major changes planned for the
future. The data collection for CMPD is focused on land- uses that influence travel
patterns such as major office developments, shopping locations, and residential lots. The
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 25
primary source of land- use information should be the regional planning agency. In the
case of I- 880, the information was collected from the Association of Monterey Bay Area
Governments ( AMBAG). For the SR- 41 corridor, the land use data was obtained directly
from Caltrans District 6 planning staff.
Transit and other Inter- modal Facilities
Transit information provides a summary of complementary transportation modes along
the corridor. All forms of transit should be considered, including local and express bus
service, light- rail, heavy rail/ subway as well as long- distance passenger rail service. For
the I- 880 corridor, this will include at a minimum, the Bay Area Rapid Transit ( BART)
and AC Transit. For each transit operator the data that collected should include
information on daily ridership, frequency of service, major origins and destinations
served, and major stops.
Other inter- modal facilities consist mainly of seaports, airports, and freight rail and
transshipment facilities. This type of facilities turns out to be more relevant for the I- 880
corridor than SR- 41 since the former is physically adjacent to both passenger facilities
( such as the Oakland Airport) and freight facilities ( such as the Port of Oakland and rail
transshipment sites). For the Port of Oakland, the information that is collected includes:
• The type of ships served ( container versus cargo)
• Annual tonnage
• Most typical commodities carried
• Major trading destinations
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 26
• Primary access modes and routes
For the Oakland airport, the collected data includes:
• The extent of passenger and freight service
• Annual enplanements
• Access modes, especially the extent to which the corridor and certain interchanges are
used.
Appendix A of this report provides more detailed transit and inter- modal facility data
related to the I- 880 corridor.
Environmental Scan
This element summarizes the potential environmentally sensitive issues along the
corridor, including locations of wetlands, sensitive habitats, hazardous waste sites, and
general air quality. The purpose is to provide a background about issues that may need to
be considered when developing system management strategies for addressing corridor
mobility problems. It is not, however, intended to identify previously unknown
environmental problems. The environmental scan data was collected from Caltrans GIS
layers. Refer to Appendix A for more detailed data on environmental scan data for the I-
880 corridor.
Planned and Programmed Improvement Strategies
The strategies provided are a list of planned and programmed improvement projects that
are related to the studied corridor. They are obtained from Caltrans and local and regional
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 27
transportation management authorities. This information is crucial to developing base
year and future year improvement scenarios that can be evaluated through micro-simulations.
3.1.2 Traffic Description Data
Traffic description data provides basic parameters for describing traffic conditions and
states. Traffic performance measures, e. g. delay, reliability, and productivity, can all be
derived from these basic traffic description data. Except for OD matrices and vehicle
types, traffic description data was collected from Tuesdays to Thursdays to capture the
traffic conditions during “ typical” weekdays. In particular, basic traffic description data
includes:
Origin- Destination Demand Matrix
The OD demand matrix ( the most critical input to the simulation model) represents the
number of trips from a given origin to a destination. This could be static, representing the
average trip pattern between the OD pair, or dynamic, capturing the detailed time-dependent
( e. g. in each 15- minute interval) demand for the given OD.
For micro- simulations, a dynamic OD matrix is normally needed. Without a reasonably
correct OD matrix, micro- simulations tend to produce erroneous results that can not be
resolved easily by later calibration steps. In practice, the OD matrix ( either static or
dynamic) is not directly observable. Usually, static OD is available from traditional
corridor planning using travel demand models. On the other hand, dynamic OD needs to
be estimated from static planning OD, traffic flow, counts, and turning volume data. This
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 28
requires data with much finer level of detail, especially in the temporal domain. For
example, 5 or 15- minute link volume data is usually needed for entire rush hours in order
to produce a reasonably correct dynamic OD matrix.
Vehicle Type Mix
Vehicle type mix provides the types of vehicles and their percentages in the studied
corridor. Particularly for micro- simulations, vehicle type mix is an important input since
different types of vehicles ( e. g. passenger cars and trucks) tend to have varied
characteristics and behaviors such as accelerations and decelerations and gap acceptance
for lane changing, etc. This will in turn greatly impact the overall traffic flow
characteristics. For a Paramics simulation, vehicle type and percentage need to be
integrated with OD demand matrix. In the I- 880 study, vehicle type mix data was derived
from data collected at Weight- In- Motion stations ( WIM), based on the Federal Highway
Administration ( FHWA) 2 definition of vehicle categories. Detailed data processing
procedure is provided in Section 5. For the SR- 41 corridor, the vehicle type data was
provided by District 6 staff.
Besides vehicle type mix data, micro- simulations also require vehicle performance data
such as maximum acceleration and deceleration. However, this kind of data is difficult to
collect in practice, therefore the default values provided in Paramics were used.
Traffic Volume, Occupancy, Speed
2 FHWA, Manual for Vehicle Classification
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 29
Traffic volume is the most basic parameter describing the state of freeway traffic. Traffic
volume data should be collected at identified mainline and ramp locations. The volume
data is critical for corridor performance evaluations and micro- simulations. However,
these two applications normally require volume data at different levels of granularity. For
performance evaluation, 30- minute data works reasonably well; while for simulation ( e. g.
dynamic OD estimation purpose) 5- or 15- minute volume data is more preferable.
Occupancy and speed data are normally collected at specific locations along the freeway
mainline. They can serve as input data for micro- simulation calibrations. For example,
speed data can be used as the primary data type for building speed contour maps in order
to identify possible bottlenecks along the corridor. Occupancy data can also be used for
simulation calibration and bottleneck analysis. Generally speaking, occupancy and speed
data can be collected in a 5- or 15- minute interval.
A major source for collecting statewide traffic volume, occupancy and speed data in
California is via the Performance Measurement Systems ( PeMS) 3 which archives loop
detector data from eight districts ( see Section 3.2). The data provided by PeMS includes
30- second raw data and 5- minute, 15- minute, and hourly aggregated data. For dual loops,
speed data is measured; for single loops, speeds are estimated using updated vehicle
length4.
3 http:// pems. eecs. berkeley. edu/ Public/
4 Kwon, J. Joint estimation of the traffic speed and mean vehicle length from single- loop detector data. In
Proceedings of the 82nd Annual Meeting of the Transportation Research Board ( CD- ROM), Washington,
DC.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 30
For the I- 880 corridor, on and off ramp volume data was obtained through two sources:
hourly census data from Caltrans District 4 and 30- second data from the Traffic
Management Center ( TMC). For SR- 41, all traffic volume, occupancy and speed data
were collected by District 6 staff at 11 count locations since PeMS was only archiving a
limited amount of detector data for the SR- 41 corridor during the study period.
Traffic Counts and Turing Volumes
Traffic counts and turning volumes are mainly used with arterial streets. Traffic counts
are collected for arterial roadway links while turning volumes are for intersections. This
data is usually collected in a 15- minute or at least 30- minute interval, which are
important inputs for simulation calibration and dynamic OD estimations. In case of the I-
880 study, traffic counts and turning volumes were obtained from local and regional
transportation agencies. For the SR- 41 corridor, historically collected data was available
and additional field data collections were also conducted to reflect the most up- to- date
traffic conditions along the corridor.
Travel Time
Travel time is one of the major criteria for assessing the mobility performance of the
studied corridor, for both freeways and arterials. It is also one of the most important
forms of traveler information that is currently provided to the driving public by various
means ( such as CMS). Besides performance evaluations, travel times can be used for
simulation calibration purposes in Paramics. The travel time data was collected through
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 31
Tach runs5 by the districts staff and made available to the project team. Other data
sources ( such as 511 travel time in the Bay Area) also provide link- based or route- based
travel times that can be used in CMPD.
Traffic Incidents
The number and types of traffic incidents are important measurements for evaluating
corridor safety performances. In California, traffic incident data is readily available
through the California Highway Patrol ( CHP) database. For traffic accidents, Traffic
Accident Surveillance and Analysis System ( TASAS) provides detailed records on
accidents that are State highway related.
3.2 Data Sharing and Field Collection
Given the massive data requirements for conducting CMPD as indicated in Section 3.1, it
would be too time and resource consuming if all data were collected from scratch.
Fortunately, most of the data needed by CMPD are available through various agencies
and other freely accessible and dedicated data sources. Under certain circumstances,
however, extra data was needed for specific purposes ( e. g. dynamic OD estimations). In
these cases, additional field data collection efforts were made.
5 Caltrans, District 4, Congestion Monitoring Procedures and Guidelines, September, 1996.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 32
3.2.1 Data Sharing
Much of the data needed for the CMPD is readily available from various public agencies,
including Caltrans ( the Headquarters and local Districts), regional transportation planning
agencies, cities and counties, and other transportation management authorities ( e. g., Port
of Oakland). Most of these public agencies are corridor stakeholders, therefore data needs
and sharing can be addressed via stakeholder meetings and other means of stakeholder
involvement ( see Section 2.1). Table 3.1 summarizes the major information providers and
their available data.
Table 3.1 Major Data Providers and Available Data
Data Provider Available Data
GIS Layers Endangered species, wetlands, vegetation, air quality
Political boundies
Photolog Detailed geometry and number of lanes of highway sections
Traffic Operations Number and location of TMS elements
Caltrans Metering strategy and detailed control algorithm, if any
Local Districts
Extra traffic description data, if any, including ramp volume,
arterial link counts and/ or turning volumes
Tach run travel time data
Regional planning agencies
Land use
Major origins and destinations and planning OD demand matrices
Travel desire lines
Mode shares
Cities and counties
Signal timing plans
Arterial geometry
Link counts and turning volumes, if any
Transit agencies Description of routes and services
Ridership
Airport and seaport authorities ( I- 880)
Description of facilities
Annual enplanements, tonnage and/ or use
Available access modes and mode split
Major access roads
Project reports and environmental
study documents
Description of existing Caltrans facility and state- of- practice
Environmental indicators
TMS baseline inventory Number of existing and planned TMS field elements by districts
Other data ( such as majority of the traffic description data) can be provided by some
dedicated data sources, including PeMS, CHP incident database, and Caltrans Accident
Surveillance and Analysis System ( TASAS) reports.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 33
PeMS archives loop detector data from eight districts of Caltrans, primarily volumes,
occupancies and speeds. CHP incident database contains information related to incidents
on State highways in California. The information includes time and location, as well as
type of the incident. Real- time CHP data can be obtained via the Internet6 and archived
CHP data can be downloaded directly from the PeMS website. Figure 3.1 shows the
incident report at CHP website for the Bay Area at a given day.
Figure 3.1 Snapshot of Incident Report in the Bay Area ( CHP)
The TASAS accident database contains specific data for accidents that are State highway
related. Each accident record is a location specific to a ramp, intersection or highway
postmile address. The master file contains records for 10 years plus the current year. The
individual records in the TASAS accident database contain general accident information
including:
o Location
o Date and time
6 http:// cad. chp. ca. gov/ default. asp
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 34
o Severity
o Primary collision factor
o Environmental items
o Roadway conditions
o Type of collision
o Number of vehicles involved
3.2.2 Field Data Collection
Besides data obtained from public agencies and other dedicated data sources, extra data is
sometimes needed ( e. g., ramp volumes, arterial link counts and turning volumes). Often
field inspections are desirable as well to observe the real- world traffic states and verify
bottleneck analysis results. All these require additional field data collection efforts.
Generally, extra field data collections will be made if any of the following four
conditions is met:
• Data Coverage is not sufficient ( e. g., for the I- 880 study, ramp data was not available
from PeMS);
• Collected data is out- of- date ( e. g., for the SR- 41 study, data collection was conducted
in 1998);
• Level of detail is not sufficient ( e. g., only hourly data is available while finer dataset
is more desirable);
• Field inspections for bottleneck evaluations. ( e. g., this was conducted for the I- 880
study to verify results of bottleneck analysis).
Field Data Collection for I- 880 Corridor
The most challenging field data collection issue for the I- 880 study was related to ramp
data. Since PeMS did not archive ramp data for the I- 880 corridor, 30- second raw
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 35
detector data was obtained from Caltrans District 4. However, due to detector
configuration problems ( i. e., how to interpret the 30- second data) and data processing
( e. g., aggregating 30- second data to 15- minute), the complete set of ramp data for the I-
880 study was not available until early 2006. This resulted in study delays as ramp data is
critical to the dynamic OD estimation and simulation model development.
Other field data collection efforts for I- 880 mainly focused on field inspections. The
major purpose was for calibrating micro- simulations or to verify bottleneck analysis
results. They also helped to identify causes of bottlenecks and develop potential
improvement strategies. Field inspections are usually qualitative, using visual
assessments. In certain cases, field data collection is conducted ( e. g. ramp volume counts).
Figure 3.2 depicts the field inspection results for the I- 880/ SR- 84 interchange. In this
effort, visual assessment was conducted to estimate the queue lengths of on and off ramps
for the interchange. The estimated queue length is used by the simulation team to verify
the calibration results.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 36
Figure 3.2 Field inspection Results for I- 880 vs. SR- 84 Interchange
Field Data Collection for SR- 41 Corridor
Due to the fact that the available traffic description data for SR- 41 corridor was limited
and largely obsolete, intensive field data collection efforts were conducted for this study.
The traffic data of SR41 network was collected in 1998, which does not reflect current
traffic conditions. In addition, the demand information ( counts) and traffic performance
data had not been collected on the same day, posing difficulties for data processing and
analysis. The newly collected data for SR- 41 includes:
Freeway loop detectors count data
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 37
Freeway traffic volume data, including those of mainline, on/ off ramps and freeway
interchange, were collected by the loop detector from 11 count locations. Figure 3.3
demonstrates the count stations along SR- 41 corridor. Data for three consecutive days
was collected from Tuesday to Thursday ( November 16 – 18, 2004) to present the typical
traffic conditions. The collecting period was five minutes. Since the congestion in this
corridor was mainly in the freeway system, the freeway volume data is the most
important to match in the calibration.
Figure 3.3 SR- 41 Freeway Volume Data Collection Locations
Arterials traffic data
Previously there were some historical arterials turning count data for most of the
signalized intersections. However, most of them were outdated and were only collected
by hour. As a result, they cannot fit the calibration requirement. New count data
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 38
collection was conducted from Tuesday to Thursday which can reflect the typical week
day traffic pattern. Nine locations were selected for collecting traffic counts and
intersection turning volumes, as shown in Table 3.2.
Table 3.2 Locations for Link Counts and Turning Volumes for SR- 41 Corridor
Intersections turning Cordon link
Blackston & Herndon Nees west of Blackston street
Fresno & Herndon Nees east of Fresno street
Blackston & Shaw Bullard west of Blackston street
Fresno & Shaw Bullard east of Fresno street
Blackston & Shields Ashlan west of Blackston street
Fresno & Shields Ashlan east of Fresno street
Blackston & Mckinley Belmont west of Blackston street
First & Mckinley Belmont east of First street
Locations
First & Tulare Tulare between T and U street
Tach run data
Vehicles equipped with GPS devices were sent out to traverse the study area ( SR41
freeway, and two parallel arterials) to collect travel time data. The vehicle location was
recorded every second. At the end of the trip, all records were output to a text file and
processed by specialized software. Travel time and speed data between sections were
then generated. They are called the “ Tach run” data7. With archives of multiple days and
multiple vehicle runs, the travel time data can be used in calibration. In this study, we
conducted Tach runs on SR41, Blackstone Ave, and Fresno Ave. Figure 3.5 depicts the
check point locations for the Tach runs conducted for SR- 41 corridor.
7 Caltrans, District 4, Congestion Monitoring Procedures and Guidelines, September 1996.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 39
Figure 3.4 Tach run Check Point Locations for SR- 41 Corridor
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 40
3.3 Performance Evaluation Data
System performance evaluation is the basis upon which corridor management plan
strategies are developed. In general, system performance evaluation includes base
performance measurements related to current or recent conditions, and future
performance measurements related to projections derived from the models. This report
focuses on performance evaluations on the State Highway System. For arterials,
performance evaluation is much harder to conduct due to the complexity of computing
performance measurements for arterials, e. g., delays. This report focuses on the travel
times of arterials only.
In CMPD, various measurements have been used to evaluate the system performance,
including delay, travel time, productivity, safety, and reliability ( see Section 4). Generally,
these performance measurements indicate, at a relatively high- level, how the system
performs ( base case) or will perform ( future case). Therefore, data required for
performance evaluation is mainly to compute these measurements in an aggregated
manner. In most cases, less detailed data could suffice compared with data needs for
micro- simulation. For example, 15- minute or 30- minute or even hourly is often adequate
to compute performance measures, while much finer data is usually required for micro-simulation
calibrations and dynamic OD estimations. Performance evaluations basically
need traffic description data. Table 3.3 summarizes the data needs, possible sources, and
level of details for computing different performance measurements.
Table 3.3 Data Needs for Performance Evaluation
Data Needs Level of Details Measurements Sources
Volume, Occupancy, 15- or 30- minute VMT, Delay, speed, PeMS
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 41
Speeds productivity, bottleneck
analysis
Travel time 15 or 30- minute
interval
Travel time, reliability PeMS, Tach run, 511
Truck Volume 15 or 30- minute
interval
Truck VMT PeMS
Incident/ Accident reports / Safety TASAS, CHP, PeMS
3.4 Simulation Model Data
The general steps and data requirements for conducting micro- simulations can be found
in the California Department of Transportation’s Guidelines for Applying Traffic
Microsimulation Modeling Software8 and FHWA’s Traffic Analysis Toolbox, Volume III:
Guidelines for Applying Traffic Microsimulation Modeling Software9.
For CMPD, the data requirements for micro- simulation model development are divided
into three categories: network coding, OD matrix estimations and simulation calibrations,
and improvement scenario evaluations. The following three sub- sections will discuss
each in more detail.
3.4.1 Network Coding Data Needs
Network coding is the first step to develop a micro- simulation model. The data needed is
primarily the corridor description data, including freeway and arterial geometry, ramp
8 Dowling Associates. Guidelines for Applying Traffic Microsimulation Modeling Software, Caltrans
Report, September 2002.
9 Federal Highway Administration, Traffic Analysis Toolbox, Volume III: Guidelines for Applying Traffic
Microsimulation Modeling Software, Publication No. FHWA- HRT- 04- 040, June 2004.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 42
metering locations and control plans, signal timing plans and coordination, etc. Table 3.4
summarizes data needs for network coding.
Table 3.4 Data Needs for Network Coding
Data Needs Descriptions Sources
Photolog Number of lanes, locations of on/ off ramps,
signs, and lane drops
Caltrans
Aerial photo Curbs, number of lanes Teraserver. com
Geometry map for each
arterial intersection
Intersection geometry, number of lanes, lane
assignment
Cities and counties;
Traffic signal Type of signal, signal timing planning and
coordination
Cities and counties
Ramp metering Type of ramp meter, metering control plans Caltrans local districts
Due to the inter- regional nature of many corridors, collecting intersection signal timing
plans imposes the most challenging part of the overall data collection effort. For the I- 880
study, the project team has developed the following sequential data requesting and
collection steps that proven to be effective to reduce the overall efforts required:
• Develop a master list of all signalized intersections to be included in the simulation
model.
• Send the master list to the Caltrans District 4 Division of Operations as the starting
point.
• Once the team receives the data from Caltrans, divide the remaining intersections by
cities and send city- specific requests.
• For those intersections that are not covered by Caltrans or the cities, contact the
Counties for data.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 43
3.4.2 OD Matrices Estimation and Simulation Calibration
OD estimation and simulation calibration require primarily traffic description data ( e. g.,
freeway mainline and ramp volumes, arterial link counts, and intersection turning
volumes). For calibration purposes, 15- minute data may be adequate; while for dynamic
OD estimations 5- minute data is sometimes needed. For OD estimations, definitions of
OD pairs and the planning OD demand matrix are also important input data, which can be
obtained from regional or county travel demand models. Table 3.5 summarizes the data
needs for OD estimations and simulation calibrations.
Table 3.5 Data Needs for OD Estimation and Simulation Calibration
Locations Data Needs Time Resolution Source
On/ off ramps volumes 5- or 15- minute PeMS; Caltrans ( Census
hourly); TMC ( 30- second);
Mainline loop stations Volume,
occupancy, speed
5- or 15- minute PeMS; Caltrans ATMS
Freeway
Selected segments Travel times 15- or 30- minute PeMS; Tach run; 511
Major links Link counts 15- minute
Major intersections Turning volume 15- minute
Caltrans; cities; field data
collection
Arterials
Selected routes Travel times 15- or 30- minute Tach run
The actual locations and coverage of freeway, arterial and travel time data collections for
SR- 41 corridor are shown in Figure 3.3, Table 3.2 and Figure 3.4, respectively. For I- 880
corridor, 65 NB loop detector locations and 73 SB loop locations, as defined at PeMS,
were chosen to collect mainline data. And 44 locations were selected for collecting on/ off
ramp data. Table 3.6 and 3.7 further list the selected arterial links and intersections for
northbound and southbound I- 880 corridor respectively. Figure 3.5 shows the selected
freeway segments for computing and comparing travel time performances.
Table 3.6 Arterial Data Collection for Northbound of I- 880 Corridor
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 44
Table 3.7 Arterial Data Collection for Northbound of I- 880 Corridor
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 45
Figure 3.5 I- 880 Freeway Segments for Travel Time Collections
3.4.3 Scenario Evaluation
To evaluate base year or future year improvement strategies, the planned or programmed
corridor improvements from Caltrans are the most critical dataset. Figure 3.6 depicted
the improvement projects at D6 for the SR- 41 corridor. This report only covers the
generation of the improvement scenarios for the SR- 41 corridor. Detailed scenario
Grand St Interchange
29th St
98th St
I- 238
SR- 92
SR- 84
Automall Pkwy
SR- 237
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 46
evaluations are not yet available for any of the study corridors, and will be provided in a
separate report as part of CCIT TO 1015.
Figure 3.6 Planned and Programmed Improvement Strategies for SR- 41 Corridor
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 47
3.5 Data Cleaning and Processing
Some of the collected data, especially the traffic count/ volume data, may have
discrepancies with each other since they may have been collected during different days.
As a result, more data cleaning and processing are necessary. Three methods have been
applied in CMPD for this purpose. The first approach is to use a tool built in the Paramics
Estimator, called “ Validate Data”, which can detect count discrepancies automatically.
A second approach is to use an Excel template developed by Caltrans engineers to locate
the discrepancy data. One example is provided in Figure 3.7 which is to check the data
consistency for a given arterial intersection for the SR- 41 study.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 48
Figure 3.7 Excel Template for Data Consistency Checking
A third method, used in the I- 880 study, is to simultaneously process the ramp and
mainline traffic volume data to make them consistent with each other. The method
assumes the ramp data is perfectly accurate, and the mainline volume is adjusted
accordingly. The process starts from the upstream end of the highway, where its data is
regarded to be more accurate, and proceeds until the downstream end of highway is
reached. However, if this method is used directly, the data set will be very sensitive to
errors. As one example, if there is an error at the starting point of count adjustment, all
counts points on the downstream will be impacted by the error. To prevent this kind of
error propagation, the adopted method should have multiple reference points ( i. e., one
point every 2- 3 miles). A reference point corresponds to a mainline loop detector station
494
369
35
0
0.156
0.688
0.156
1.000
0.000
0.000
1.000
1.000
77
340
77
494
0
0
35
35
76 0.131 0 0.000
504 401 0.693 579 579 1.000
420 102 0.176 382 0 0.000
579 1.000 579 1.000
0.283 119 0.000 0
0.629 264 579 1.000 382 579
0.088 37 382 0.000 0 428
1.000 420 1.000 382
26
174
41
241
0
0
11
11
0.108
0.722
0.170
1.000
0.000
0.000
1.000
1.000
479
241
0
11
479
241
ZONE138
ZONE139
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 49
whose data is considered to be 100% accurate. To select a reliable reference point,
original data was compared with count data provided by the PEMS website. If count data
for a point has a big difference among three surveyed days or it has poor consistency with
other reference points, it cannot be used as a reference point. Moreover, if the observation
rate of the point is less than 100%, it cannot be used as a reference point either.
Usually, count adjustments are started from the point to the next reference point. For a
mainline detector station between two adjacent reference points, although PeMS may
report data for this location, its data is actually estimated based on data on reference
points and ramps.
Figure 3.8 provides an example of using the third method. In this figure, rows between
two reference detectors are separated, so errors in a detector cannot influence beyond the
isolated segment. In consistency checking, one- hour counts were used because it is
assumed that only ramp census counts are reliable. If an upstream count and a
downstream count have inconsistency, it is distributed to all detected points between the
two reference points. More details on this data processing method are provided in Section
5 of this report.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 50
Isolated segment
1 hr count
Reference detector
Figure 3.8 Method for Data Consistency
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 51
4. Corridor Performance Evaluation
As discussed in Section 3.3 and also shown in Figure 1.1, system performance evaluation
is the basis for developing corridor management planning strategies. It can help both
traffic engineers and decision makers identify problems within the study area. This is
normally the starting point for developing new management strategies. Generally,
performance evaluation can be conducted for freeway systems, arterials, and transit. This
report only focuses on performance evaluations on the freeway systems. The evaluation
includes base performance measurements related to current and recent conditions, and
future performance measurements related to projections derived from the models. This
report only covers current performance measurements and separate reports will be
developed for future performance measurements. In particular, this section mainly
focuses on the definitions of the performance measures used in CMPD; detailed corridor
performance assessment results can be found in the appendices of the report.
4.1 Performance Measures
Performance measures define quantitatively how the corridor system performs at current
or projected future situations. In CMPD, several popularly used measures have been
adopted, including various mobility, safety, reliability, and productivity measures. These
measures can be computed using collected data ( refer to Section 3 for more detail on data
collection), although some are readily available from data providers such as PeMS.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 52
4.1.1 Mobility Measures
As the primary objective of transportation system is to transport people and goods,
mobility measures are crucial for corridor system performance evaluations. In CMPD, the
mobility measures used include vehicle miles traveled ( VMT), truck VMT, delay, speed,
and travel time.
Vehicle Miles Traveled ( VMT)
For a given unit of time and a given section of the freeway, VMT is defined as the sum of
freeway miles driven by each vehicle. For a section of fixed length ( denoted as L), the
number of freeway miles driven is simply the flow for a period of time multiplied by the
length L10. VMT can be readily obtained from PeMS. Figure 4.1 below is an example of
the VMT plot computed by PeMS for NB I- 880 on a particular day ( March 1, 2005). It
shows the expected variation of VMT with more VMT in the daytime compared with
night time. Also the two peaks appear during AM and PM rush hours.
10 http:// pems. eecs. berkeley. edu/ Help/ index. phtml? content= calculations
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 53
Figure 4.1 VMT for NB I- 880 on March 1, 2005
Truck VMT
Truck VMT is the VMT measured and computed particularly for truck flows. This can
also be obtained directly from PeMS. Figure 4.2 below is the truck VMT for NB I- 880
for March 1, 2005. Note the different trend of truck VMT compared with that for entire
traffic flow for the same day in Figure 4.1.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 54
Figure 4.2 Truck VMT for NB I- 880 on March 1, 2005
Delay
Delay is defined as the amount of additional time spent by the vehicles on a section of
road due to congestion. This is the difference between the travel time at a non- congestion
speed and the current speed. Various congestion, or threshold, speeds can be used. In
CMPD, 60MPH was selected for computing the total delay and 35MPH for severe delay.
Delay can be defined mathematically as D = F * ( TT - TT_ t) = F * ( L/ V - L/ V_ t). In these
formulas, TT is the travel time at current speed, TT_ t is the travel time at the threshold
speed, F is the flow, and L is the length of the segment, and V and V_ t are the current
speed and threshold speed ( i. e. 60 MPH), respectively. By this definition, delay can never
be negative. Again, delay can be readily obtained from PeMS. Figure 4.3 depicts the total
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 55
delay computed by PeMS for NB I- 880 on March 1, 2005. The curve contains the typical
AM and PM peaks.
Figure 4.3 Delay for NB I- 880 on March 1, 2005
Travel Time
Travel time is the most direct measure for mobility, measuring the time needed for
traveling from a given origin to a destination. In CMPD, travel time is collected from
Tach runs and some dedicated data sources such as 511. Table 4.1 lists the time-dependent
( in 15 minutes) travel times collected for NB SR- 41 from O St to Friant Ave.
Table 4.1 SR- 41 NB Travel Times
Route Periods Average Tach run
Travel Time ( sec)
Northbound 16: 00- 16: 15 559
SR- 41 from 16: 15- 16: 30 553
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 56
16: 30- 16: 45 549
16: 45- 17: 00 562
17: 00- 17: 15 609
17: 15- 17: 30 723
O St to
Friant Ave.
17: 30- 17: 45 574
Speed
Speed is another important mobility measure of the corridor system, which is available
from PeMS. Note that PeMS archives measured speeds for dual loop detectors, and
estimated speeds will be computed and used if detectors are single loops. Figure 4.4
below depicts the variation of lane by lane speed for NB I- 880 at 4: 00 PM on March 1,
2005. Speeds can also be used to generate speed contour maps that can be used for
simulation calibrations and bottleneck analysis, as discussed in detail in Section 4.2.
Figure 4.4 Speed Variation for NB I- 880 at 4: 00 PM on March 1, 2005
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 57
4.1.2 Safety Measures
Mobility and safety are two major objectives of transportation systems. In particular,
safety has been an increasing concern by both transportation management agencies and
the public. In CMPD, two safety measures have been adopted: incidents and accidents.
They are available from two data sources, CHP and TASAS.
Incidents
Traffic incidents can be defined as “ an unplanned randomly occurring traffic event that
adversely effects normal traffic operations.” 11 This definition of incidents is very broad
and may contain accidents ( car crashes), debris on roadway, stalled vehicles, etc. Figure
4.5 plots the incidents occurred along NB I- 880 averaged for March 1 – 10 of 2005.
11 Traffic Management Data Dictionary ( TMDD) and Message Sets for External Traffic Management
Center Communications ( MS/ ETMCC) Website. Institute of Transportation Engineers.
( http:// www. ite. org/ tmdd)
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 58
Figure 4.5 Incidents on NB I- 880 for March of 2005
Accidents
Accidents are specifically defined as vehicle crashes on the roadway. Therefore,
accidents can be treated as one type of traffic incident. Figure 4.6 depicts the average
daily number of accidents occurred for both directions of I- 880. The data was obtained
from the TASAS database and averaged for each month from 1999 to the end of 2004.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 59
0
2
4
6
8
10
12
14
Jan- 99
Mar- 99
May- 99
Jul- 99
Sep- 99
Nov- 99
Jan- 00
Mar- 00
May- 00
Jul- 00
Sep- 00
Nov- 00
Jan- 01
Mar- 01
May- 01
Jul- 01
Sep- 01
Nov- 01
Jan- 02
Mar- 02
May- 02
Jul- 02
Sep- 02
Nov- 02
Jan- 03
Mar- 03
May- 03
Jul- 03
Sep- 03
Nov- 03
Jan- 04
Mar- 04
May- 04
Jul- 04
Sep- 04
Nov- 04
Figure 4.6 Average Daily Accidents for I- 880 ( Both Directions)
4.1.3 Reliability Measure
The reliability measure adopted for CMPD is “ travel time reliability”, which describes
the variation of travel times for the studied corridor. One example is provided in Figure
4.7 which depicts the mean and variation of travel times for I- 880 NB.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 60
Figure 4.7 Travel Time Reliability for I- 880 NB
From Figure 4.7, it is clear that the larger the range between + 1 standard deviation and - 1
standard deviation, the worse the travel time reliability is. The worst cases normally
occur during the AM or PM peak hours.
4.1.4 Productivity Measure
Productivity is the number of lane- mile- hours on the freeway lost due to congested
conditions instead of under free- flow conditions. Thus it is also referred to as “ Lost
Productivity” or “ Lost Lane- Miles”. When the freeway is in congestion ( i. e., the speed is
below 60MPH in CMPD studies), we find the ratio between the measured flow and the
capacity for this location. This drop in capacity is due to the fact that the freeway is
operating in congested conditions instead of free- flow. Multiplying one minus this ratio
by the length of the segment can produce the value of lost productivity. Figure 4.8 shows
one example of the lost productivity for I- 880 NB on June 15, 2006. The plot was
-
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
6 8 10 12 14 16 18 20
Hour
Travel Time ( Minutes)
+ 1 Std Dev Travel Time
Average Travel Time
- 1 Std Dev Travel Time
Best Travel Time
AM Midday PM
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 61
generated by PeMS, and presents two obvious peaks during the AM and PM rush hours.
Also we can observe the obvious similarity between this lost productivity curve and the
delay curve in Figure 4.3.
Figure 4.8 Lost Productivity for I- 880 NB on March 1, 2005
4.2 Corridor Bottleneck Analysis and Verification
Bottlenecks are sections of the freeway that either have capacities less than or demand
greater than other sections. These are the locations that will probably be the first to
experience congested conditions as traffic grows. Due to the dynamic nature of traffic
flow, bottlenecks are not only location specific, but also time- dependent. Therefore,
identifying the exact time durations and spatial extents of bottlenecks along the studied
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 62
freeway is of great importance for system performance evaluation, simulation calibration,
and development of improvement strategies. Because of the temporal- spatial
characteristics of bottlenecks, in CMPD we use two- dimensional ( time and space) speed
contour maps to identify bottlenecks. Sometimes, two- dimensional volume and
occupancy plots can also serve for the bottleneck analysis. As one example, Figure 4.9
depicts the bottleneck plot for I- 880 NB for AM peak hours ( here we use 6: 00 am – 10:
am) for a single day ( March 1, 2005).
From the speed map, bottlenecks can be visually identified. For example, from the plot in
Figure 4.9 we can observe that for that particular day, three major bottlenecks occurred.
The one on the most south was located from Whipple Rd and Tennyson Rd with its
duration roughly from 7: 45 am to 9: 10 am.
For a highly congested corridor ( e. g. I- 880), using one- day speed data is questionable
since there is almost no “ typical” day for the corridor. In this case, speeds from multiple
days need to be analyzed to make the bottleneck analysis results more reliable.
Depending on the purposes, average speeds or certain percentile plots may be used for
bottleneck analysis. Figure 4.10 below is the average speed plot for 20 days ( Jan. 24 –
Mar. 9 of 2005, for only Tuesdays, Wednesdays and Thursdays). Apparently, two
bottlenecks are significant in this average speed plot.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 63
Figure 4.9 Bottleneck Plot for I- 880 NB
Whipple Rd
Tennyson Rd
Washington Ave
98th Ave
66th Ave
Kennedy St
23rd St
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 64
Figure 4.10 Average Speed Bottleneck Plot
4.3 Current Performance
This section briefly discusses current performance of I- 880 and SR- 41 corridors. Detailed
performance assessment results can be found in Appendices A and B, respectively, for
the I- 880 and SR- 41 corridors.
Whipple Rd
Tennyson Rd
Hegenberger
Rd
23rd St
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 65
Generally speaking, the I- 880 corridor is experiencing heavy congestion during AM and
PM peak periods for both northbound and southbound directions. Table 4.2 lists the delay
information, in vehicle- hours, based on the time- of- the- day for both directions from 2003
to 2005. The heavy delays shown in Table 4.2 can also be confirmed by the speed
contour plots in Figure 4.9 and Figure 4.10.
Table 4.2 Delay of I- 880 Corridor
Year AM Peak Mid Day
Evening and Early
AM PM Peak Total Daily
2003 1,499 1,237 552 2,547 5,835
2004 1,124 1,067 360 2,317 4,867
2005 1,331 1,434 285 2,351 5,402
Year AM Peak Mid Day
Evening and Early
AM PM Peak Total Daily
2003 1,924 1,397 276 2,249 5,846
2004 1,728 1,796 291 2,677 6,491
2005 1,678 2,196 232 2,885 6,991
Northbound Direction
Southbound Direction
The I- 880 corridor is also incident- prone. As depicted in Figure 4.6, there have been at
least 8 collisions daily for the corridor ( both directions) from Jan. 1999 to Nov. 2004.
Therefore, we can conclude that the I- 880 corridor is a heavily congested and incident-prone
urban corridor for which short- and medium- term operational improvements are
needed to solve corridor level traffic congestion and safety problems.
For the SR- 41 corridor, only light congestion has been experienced. Figure 4.12 and
Figure 4.13 show, respectively, the speed contour maps for PM peak periods for the two
directions based on Tach run data. In these figures, each cell of the time- space diagram
represents the average speed over multiple Tach runs for a specific mainline freeway
subsection. Thus speeds in the figure are shown over time ( row summary) and over space
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 66
( column summary). Three levels of speeds are represented by different levels of shading:
below 49 mph, between 50 and 59 mph, and over 60 mph.
We can see from the figures that PM congestion mainly appears on NB of the freeway
from 4: 00 PM to 5: 30 PM. The most congested areas are the Mckinley Ave., Shields
Ave., and Friant Ave. However, for SB, there is almost no congestion. Therefore, the
focus of the SR- 41 study is long- term improvements with the purpose of accommodating
regionally forecast travel demand in the future ( 25 to 30 years timeframe).
>= 60 mph 50- 59 mph <= 50 mph
NB O ST DIVISADERO OLIVE MCKINLESHIELDSASHLAN SHAW BULLARD HERNDONFRIANT AUDOBO
22.949 23.763 24.749 25.266 26.461 27.471 28.463 29.463 30.447 31.683 32.165
4: 00: 00 PM 65 61 61 63 61 65 61 60 60 62 65
4: 05: 00 PM 63 66 62 25 40 59 63 63 66 64 66
4: 10: 00 PM 58 68 61 60 28 57 65 66 63 66 62
4: 15: 00 PM 61 66 62 50 37 54 60 64 64 63
4: 20: 00 PM 64 64 63 40 47 51 55 61 65 60
4: 25: 00 PM 61 57 60 57 54 49 57 63 64 61
4: 30: 00 PM 61 62 62 64 63 64 62 61 62 63 66
4: 35: 00 PM 64 62 63 60 49 59 60 62 62 63 65
4: 40: 00 PM 61 61 61 57 46 56 51 64 61 62 64
4: 45: 00 PM 63 62 61 51 50 49 54 63 62 61
4: 50: 00 PM 64 62 61 46 54 42 57 62 62 60
4: 55: 00 PM 61 60 49 41 47 59 38 57 60 64
5: 00: 00 PM 61 62 62 64 63 64 62 61 62 63 66
5: 05: 00 PM 65 62 65 60 40 51 58 61 62 55 59
5: 10: 00 PM 59 60 58 22 36 39 49 59 60 56 64
5: 15: 00 PM 62 60 24 27 34 34 44 54 61 59
5: 20: 00 PM 63 59 37 29 38 58 56 55 56 62
5: 25: 00 PM 59 63 33 17 47 36 41 59 61 56
5: 30: 00 PM 68 60 54 33 33 54 63 62 61 61 70
5: 35: 00 PM 64 66 66 64 60 57 53 60 59 45 65
SB FRIANT HERNDOBULLARD SHAW ASHLAN SHIELDS MCKINLEYOLIVE DIVISADERO ST RTE 41/ 99 JENSEN
31.683 30.447 29.463 28.463 27.471 26.461 25.266 24.749 23.763 22.949 21.798 21.113
4: 00: 00 PM 64 65 65 65 65 65 63 65 65 62 64 63
4: 05: 00 PM 60 63 63 59 62 61 61 64 66 63 62 64
4: 10: 00 PM 65 64 63 62 62 62 60 63 61 60 60 62
4: 15: 00 PM 66 65 63 62 62 62 61 63 64 62 62 64
4: 20: 00 PM 67 67 64 61 62 63 62 63 68 64 64 66
4: 25: 00 PM 59 60 65 65 64 61 61 58 64 63 63 65
4: 30: 00 PM 61 62 66 65 65 60 60 61 65 63 63 62
4: 35: 00 PM 63 65 66 66 66 59 58 64 66 64 63 59
4: 40: 00 PM 67 66 67 63 61 60 63 65 63 60 57 62
4: 45: 00 PM 69 64 63 65 65 65 66 65 68 65 66 68
4: 50: 00 PM 64 64 64 63 64 61 62 65 68 63 50 67
4: 55: 00 PM 69 68 68 67 67 63 67 67 69 67 68 69
5: 00: 00 PM 62 59 62 62 54 59 61 61 65 63 25
5: 05: 00 PM 69 66 64 62 59 58 59 65 66 51 53 61
5: 10: 00 PM 66 64 62 62 60 61 61 64 64 62 62 62
5: 15: 00 PM 63 63 63 61 58 57 60 62 62 63 58 63
5: 20: 00 PM 64 64 66 65 62 60 60 62 65 63 39 61
5: 25: 00 PM 68 68 68 64 64 67 64 65 70 65 62 69
5: 30: 00 PM 67 65 71 55 43 36 61 67 46 44 61
5: 35: 00 PM 64 64 52 50 58 60 60 63 63 63 63 65
Figure 4.12 Speed Contour for SR- 41 for PM Peak ( NB)
Figure 4.13 Speed Contour for SR- 41 for PM Peak ( SB)
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 67
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 68
5. Baseline Simulation Model Development
This section provides a detailed procedure for developing the simulation model for the I-
880 study using the Paramics microsimulation tool. For the SR- 41 simulation model
development, please refer to Appendix B.
5.1 Introduction
Before the development of a microscopic simulation model, project scoping is required,
which has been explained in Section 2 of the report. Based on project scoping, the
microscopic simulation model can be further developed, which involves network coding,
data collection and preparation, origin- destination ( OD) demand estimation, and model
calibration. Usually, OD demand estimation is treated as one step of the model
calibration process.
5.2 Microscopic Simulation Network Coding
A microscopic simulation network needs to be built based on a wide range of input data,
including data of network geometry, driver behavior, vehicle characteristics,
transportation analysis zones, travel demands, and traffic control systems and traffic
detection systems ( see Section 3). The procedure that may need to be followed is
suggested as follows:
( 1) Determine the configuration of the most basic inputs to the model;
( 2) Code the skeleton network;
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 69
( 3) Code traffic control;
( 4) Add zones and demands;
5.2.1 Determine the Configuration of the Most Basic Inputs
The following basic configurations of the model need to be determined first.
5.2.1.1 Link definition
As a basic input to a Paramics simulation model, the “ categories” file includes the
definition of all available link types within the study network. Each link category
includes information that is required for simulation modeling. An example of a freeway
link category is as follows:
category 1 lanes 1 speed 65 mph width 12.0 ft colour 0x000000ff type highway
major
median width 0.0 ft
headway factor 1.000 curve speed factor 0.0 toll 0.000 cost factor 0.800
signpost 820.2 ft, 3.3 ft
The most important thing to consider at this step is to have a certain network road
hierarchy. The purpose to define a suitable road hierarchy is to maintain stability of route
choice between OD pairs in a Paramics model. This is achieved by defining a set of base
categories with suitable category costs factors. A hierarchy suggested by Quadstone12 is
defined as follows:
( 1) Major Primary Link Cost Factor 0.8
( 2) Major Secondary Link Cost Factor 1.0
12 Quadstone Limited, Quadstone Paramics V5.1 Modeler Reference Manual, Aug. 16, 2004
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 70
( 3) Major Primary Link Cost Factor 0.8
( 4) Minor Secondary Link Cost Factor 1.0
According to Quadstone, “ the hierarchy of a network can be obtained from many sources;
ordnance survey maps that have major and minor routes defined, assessment of traffic
flow data and of course local knowledge, the latter being invaluable during the calibration
process. The creation of a suitable categories file incorporating your chosen cost
hierarchy is the primary task in the network building process; this should be undertaken
before any network coding ( links/ nodes etc.) is undertaken. ”
In addition, it is better for the same type of links to use the same color for the link
auditing purpose. For our network, all freeway links have the same color; arterials with
the same traveling speed have the same color; ramp links have the same color. Using the
link auditing tool, the above- mentioned link color settings will help find link coding
conflicts easily.
5.2.1.2 Vehicle definition
Paramics regards each vehicle in the simulation as a Driver Vehicle Unit ( DVU). The
“ vehicles” file is another basic input to a Paramics simulation model. It includes the
attribute data and demand proportions of all available vehicle types and their associated
drivers’ behavior data. The attribute data include vehicles’ physical size, color, weight,
characteristics, and etc. The drivers’ behavior data include settings for drivers’
perturbation and familiarity. An example of a vehicle type is as follows:
type 1 car length 15.42 ft width 6.23 ft height 4.59 ft
acc 11.81 fpss dec - 12.80 fpss crawl speed 5.00 mph horsepower 50.00
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 71
colour 0x00ffffff name " car"
matrix 1 proportion 19.500 perturbation 2.0 familiarity 5.00
Each vehicle model could have its own vehicle type in Paramics. For detailed modeling,
each year and even each color of the same vehicle model can have its own vehicle
definition. To decide how many vehicle types are required for a simulation model, the
following factors need to be considered:
( 1) How many demand matrices will the simulation model have? For example, there
may be one for HOV, one for SOV, and one for trucks.
( 2) What level of detail will the simulation model provide?
A practical simulation model usually simplifies the “ vehicles” file by defining limited
numbers of vehicle types. The basic vehicle types can include:
( 1) four- door sedan;
( 2) sports car;
( 3) SUV;
( 4) pickup;
( 5) min- Van;
( 6) bus;
( 7) loaded class 5- 8 truck;
( 8) empty class 5- 8 truck;
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 72
( 9) loaded class 9- 14 truck; and
( 10) empty class 9- 14 truck.
The proportion of each vehicle type in its associated demand matrix needs to be
calibrated based on data of the target network. Here, we introduce a two- step method to
establish the “ vehicles” file for a simulation model. The first step is to determine vehicle
groups based on the vehicle classifications by FHWA. The groups are also dependent
upon the actual network, especially the number of demand matrices that are used by the
simulation. Each of the vehicle groups can be further divided to represent multiple
vehicle types in the second step. The vehicle type file can then be created accordingly.
( 1) Vehicle Group Determination
Based on the analysis of vehicles in the I- 880 network, two demand matrices are used for
the simulation, defined as matrix1 and matrix2 respectively. Vehicles that circulate
within the network are represented by matrix1, and matrix2 is for those passing through
the network. By further investigating the vehicle definition of FHWA, it is determined
that vehicle classes 1- 8 correspond to matrix1 and classes 9- 13 and 15 correspond to
matrix2. Classes 1- 8 are further categorized to distinguish different vehicle groups. The
vehicle groups can be finally defined in Table 5.1.
Table 5.1 Vehicle Group Definition and Percentage
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 73
matrix1 matrix2
FHWA class 1- 2
( motorcycles or
cars)
FHWA class 3
( minivans or
pickups
FHWA class
4
( buses)
FHWA class 5- 8
( trucks with 3- 5
axles)
FHWA class >= 9
( trucks with >= 6
axles)
Total 73% 22.88% 0.12% 4% 100%
HOV
( 20%) 58.4% 18.3%
NA NA NA
Regular
( 80%) 14.60% 4.58%
NA NA NA
The percentage of each category was estimated as follows.
Total percentage of each class of vehicles was derived from the Weigh- In- Motion
( WIM) data.
The HOV percentage for class 1- 3 was estimated from two sources. The first source is
the 1- hour aggregated volume data from PeMS which provides lane by lane volume ( the
left most lane is the HOV lane). The second source is the HOV report for D4 for 2004.
The detailed estimation results are depicted in Table 5.2. From this table, it is evident that
20% is a good estimate for both directions and both morning and afternoon peak periods.
Table 5.2 HOV Percentage Estimation
PeMS 1 - hour data D4 HOV report for 2004
AM PM AM PM
I880 NB 19% 20% 25% 20%
I880 SB 21% 20% 21% 21%
( 2) Vehicle Type Determination
Each of the cells in Table 1 can be further classified into different vehicle types in line
with the requirement of Paramics. Currently, the following vehicle types are applied:
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 74
Classes 1- 2 include “ Mustang”, " Crown Victoria", and " Focus Sedan".
Classes 3 include " F- 150 Pick up", " Windstar Mini- Van", and " Ford Explorer".
Classes 4 do not include multiple vehicle types.
Classes 4- 8 include both “ empty” and “ loaded” types.
Classes >= 9 include both “ empty” and “ loaded” types.
Note that according to the specific network, vehicle types in each vehicle group may be
varied; but the summation of their percentage must agree with the total percentage for
that particular group. In addition, due to the lack of actual data, the percentage for each
vehicle type within a group is estimated by roughly splitting the total percentage evenly.
3. Final vehicles file
Finally, the vehicle type file for I- 880 can be listed as follows.
vehicle types
type 1 car length 15.42 ft width 6.23 ft height 4.59 ft
acc 11.81 fpss dec - 12.80 fpss colour 0x00ffffff draw style pmx model
name " Mustang"
matrix 1 proportion 19.5 perturbation 5.0 familiarity 95.00
type 2 car length 17.72 ft width 6.56 ft height 4.59 ft
acc 7.87 fpss dec - 12.47 fpss colour 0x00ffffff draw style pmx model
name " Crown Victoria"
matrix 1 proportion 19.5 perturbation 5.0 familiarity 95.00
type 3 car length 14.44 ft width 5.58 ft height 4.59 ft
acc 7.55 fpss dec - 13.45 fpss colour 0x00ffffff draw style pmx model
name " Focus Sedan"
matrix 1 proportion 19.4 perturbation 5.0 familiarity 95.00
type 4 car length 17.39 ft width 6.56 ft height 5.91 ft
acc 7.55 fpss dec - 12.14 fpss colour 0x00ffffff draw style pmx model
name " F- 150 Pick up"
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 75
matrix 1 proportion 6.3 perturbation 5.0 familiarity 95.00
type 5 car length 16.73 ft width 6.23 ft height 5.58 ft
acc 7.55 fpss dec - 10.17 fpss colour 0x00ffffff draw style pmx model
name " Windstar Mini- Van"
matrix 1 proportion 6 perturbation 5.0 familiarity 95.00
type 6 car length 15.75 ft width 5.91 ft height 5.91 ft
acc 7.87 fpss dec - 11.15 fpss colour 0x00ffffff draw style pmx model
name " Ford Explorer"
matrix 1 proportion 6 perturbation 5.0 familiarity 95.00
type 7 car length 15.42 ft width 6.23 ft height 4.59 ft
acc 11.81 fpss dec - 12.80 fpss colour 0x000000ff
name " Mustang HOV"
matrix 1 proportion 5.0 perturbation 5.0 familiarity 95.00
type 8 car length 17.72 ft width 6.56 ft height 4.59 ft
acc 7.87 fpss dec - 12.47 fpss colour 0x000000ff
name " Crown Victoria HOV"
matrix 1 proportion 5.0 perturbation 5.0 familiarity 95.00
type 9 car length 14.44 ft width 5.58 ft height 4.59 ft
acc 7.55 fpss dec - 13.45 fpss colour 0x000000ff
name " Focus Sedan HOV"
matrix 1 proportion 4.6 perturbation 5.0 familiarity 95.00
type 10 car length 17.39 ft width 6.56 ft height 5.91 ft
acc 7.55 fpss dec - 12.14 fpss colour 0x000000ff
name " F- 150 Pick up HOV"
matrix 1 proportion 1.5 perturbation 5.0 familiarity 95.00
type 11 car length 16.73 ft width 6.23 ft height 5.58 ft
acc 7.55 fpss dec - 10.17 fpss colour 0x000000ff
name " Windstar Mini- Van HOV"
matrix 1 proportion 1.500 perturbation 5.0 familiarity 95.00
type 12 car length 15.75 ft width 5.91 ft height 5.91 ft
acc 7.87 fpss dec - 11.15 fpss colour 0x000000ff
name " Ford Explorer HOV"
matrix 1 proportion 1.58 perturbation 5.0 familiarity 95.00
type 13 car length 14.44 ft width 5.58 ft height 4.59 ft weight 0.79 ton
top speed 100.00 mph acc 7.55 fpss dec - 13.45 fpss crawl speed 50.00 mph
horsepower 100.00 colour 0x00ffffff
name " Buses"
matrix 1 proportion 0.12 perturbation 0.0 familiarity 85.00
type 14 OGV1 length 40.03 ft width 8.53 ft height 13.45 ft weight 3.88 ton
acc 5.58 fpss dec - 12.14 fpss crawl speed 55.00 mph horsepower 200.00
colour 0x00ff0000 draw style pmx model group 4
name " Truck - Class 5- 8, empty"
matrix 1 proportion 2 perturbation 0.0 familiarity 85.00
type 15 OGV1 length 40.03 ft width 8.53 ft height 13.45 ft weight 7.09 ton
acc 5.58 fpss dec - 12.14 fpss crawl speed 42.00 mph horsepower 200.00
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 76
colour 0x00ff0000 draw style pmx model group 4
name " Truck - Class 5- 8, loaded"
matrix 1 proportion 2 perturbation 0.0 familiarity 85.00
type 16 OGV2 length 64.96 ft width 8.53 ft height 13.45 ft weight 12.50 ton
top speed 65.00 mph dec - 11.48 fpss crawl speed 42.00 mph horsepower
267.00 acc profile 4 dec profile 8 colour 0x0000758b shape 2
name " Truck - Class 9- 14, empty"
trailer count 1
(
trailer 1 length 52.99 ft
colour 0x001a2d8b
model type 0
)
matrix 2 proportion 50 perturbation 0.0 familiarity 85.00
type 17 OGV2 length 64.96 ft width 8.53 ft height 13.45 ft weight 23.00 ton
top speed 65.00 mph dec - 11.48 fpss crawl speed 28.00 mph horsepower
267.00 acc profile 4 dec profile 8 colour 0x00008b00 shape 2
name " Truck - Class 9- 14, loaded"
trailer count 1
(
trailer 1 length 52.99 ft
colour 0x00142c8b
model type 0
)
matrix 2 proportion 50 perturbation 0.0 familiarity 85.00
5.2.1.3 Demand structures
Choosing the demands structure relates to the initial assignment method, the composition
of the vehicles types file, and the number of OD matrices to be used during assignment.
The determination of demands structure needs to be based on features of the target
network.
For the I- 880 network, there are many trucks due to the proximity to the Port of Oakland
port. As a result, it was determined to have a separate demand matrix ( i. e., matrix2) for
big trucks with FHWA class 9- 14. For those trucks with FHWA class 5- 8, they are
regarded as part of the demand matrix for other vehicles ( i. e., matrix1). In addition, the
study network has HOV lanes, which are operated during peak periods ( i. e. 5- 9 am and 3-
7 pm). Although the ACCMA ( Alameda County Congestion Management Agency)
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 77
planning model has specific HOV demand models, it will be costly to estimate the HOV
demand matrix since HOV flow data is hard to observe and collect. As a result, the I- 880
network has two matrices. One is for FHWA class 9- 14 trucks and the other is for all
other vehicles, including SOV, HOV and FHWA class 5- 8 trucks.
5.2.2 Code the Skeleton Network
5.2.2.1 Background images
A simulation network is generally coded based on background images. For the network
coding purpose, it is desirable for background images to have a resolution equal or
smaller than 1 meter per pixel. Background images can be CAD drawing files ( dxf format)
that can be obtained from Caltrans and/ or cities or aerial photos ( in the JPG or BMP
format) that can be obtained from the following sources:
( 1) Caltrans Digital Highway Inventory Photography Program ( DHIPP):
http:// svhqdhipp: 8080/ dhipp/ view. html, which only has aerial photos for freeways
and areas close to freeways with known resolution information.
( 2) http:// www. terraserver- usa. com
( 3) http:// mapper. acme. com
( 4) http:// maps. google. com/
( 5) http:// local. live. com/
( 6) http:// www. terraserver. com
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 78
Among them, http:// www. terraserver- usa. com is one of the best sites for obtaining aerial
photos since it is free and also provides users with resolution information. Another
excellent site is http:// mapper. acme. com, which is a Google Map application. Google
map provides high- resolution of aerial photos. The first zooming level’s resolution is 0.25
meter per pixel resolution. The second level’s resolution is 0.5 meter per pixel and the
third level’s resolution is 1 meter per pixel13.
In addition, Google Earth could be used as a source, but how to use it needs to be further
investigated.
5.2.2.2 Other road geometry data
The data that may be needed for road geometry coding are summarized in the table below.
Table 5.3 Geometry Data for Network Coding
Data Type Data Sources
As- built map Number of lanes, locations of
detectors, on- ramps, off- ramps
and lane drops
Caltrans
Photolog number of lanes, locations of
on- ramps, off- ramps, signs,
and lane drops
Caltrans
http:// video. dot. ca. gov/ photolog/
Freeway
Data
Aerial photos Curbs, number of lanes See 5.2.2.1
As- built map Number of lanes, , lane
assignment, locations of
detectors
Arterial Cities
Data
Aerial photos Curbs, number of lanes See 5.2.2.1
Background images are the most common data source for details of road network. But, if
there is anything unclear on the background images, other geometric data sources shown
from the above table are required. If the available sources can not provide details of the
13 Due to the terms of Google Map, http:// mapper. acme. com does not allow users to download map but
users can obtain it by printing screen.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 79
network, researchers need to go to the field and write down the geometric details on hand
drawing maps.
5.2.2.3 Node naming convention
Before network coding can be started, there may be a need to have a node naming
convention. Paramics provides a node with a name automatically. However, a good
naming convention may provide additional information to modelers. For example, the
node naming convention for the I- 880 network is as follows:
( 1) Names of all nodes along the Northbound I- 880 start with “ 8”.
( 2) Names of all nodes along the Southbound I- 880 start with “ 6”.
( 3) Node names have at least four numbers, e, g, 8230, 8230a. For node “ 8230”,
“ 230” has its own meaning as well, which tells modelers that the node is located
at postmile 23.0.
In order to have the above node naming convention, three steps were performed:
( 1) Modelers added nodes with their default names provided by Paramics first;
( 2) A plugin developed by the research team was used to calculate the postmile
information for all nodes along the freeway based on a reference point. As a result,
a node name lookup table was prepared.
( 3) Node names along freeways were changed to their new names based on the
lookup table through Paramics GUI.
CCIT Task Order 3 - Corridor Management Plan Demonstration – Final Report
California Center for Innovative Transportation ( CCIT) Page 80
5.2.2.4 Geometry coding
The process of road geometry coding involves adding nodes and links, modifying
stoplines and Curbs, and allocating lanes to various vehicle movements at intersections
based on background images and other geometric data.
1. Nodes
Nodes are usually placed at locations where there is a physical change in road geometry.