KLK712: Improved Simulation of Stop Bar Driver Behavior at
Signalized Intersections
Principal Investigators
Michael Kyte, Michael Dixon and Ahmed Abdel-Rahim
Project Objectives
The Federal Highway Administration has made a major shift in its
role in microsimulation modeling of traffic flow. While previously
supporting the development of the CORSIM model, FHWA is now focusing
its resources on the development of core driver behavior algorithms
and the supporting data sets representing driver behavior in a range
of traffic environments. The NGSIM program has thus far produced new
data sets for both freeway and arterial operations as well as new
algorithms that are being included in private sector developed
microscopic simulation models. The new algorithms include freeway
lane selection, cooperative/forced freeway merging, oversaturated
freeway flow, and arterial lane selection algorithm.
The NGSIM program has identified a set of priority algorithms for
arterial operations that are yet to be developed. These algorithms
include starting/stopping behavior and permitted left turns at
signalized intersections. Limited research budgets, however, mean
that FHWA is not able to invest in the development of these new
algorithms.
FHWA has reached out to the University Transportation Center
community to help support nationally identified research needs that
it now is not able to fund, including the NGSIM program. This
outreach has led the four UTCs in the Pacific Northwest to join
together to form a new mechanism for meeting national transportation
research needs. Three of these UTCs (including four universities,
Portland State University, Oregon State University, University of
Washington, and University of Idaho) have formed a regional
simulation collaboration to specifically address the simulation
modeling needs of FHWA as well as the individual states in the
region. These UTCs are ready to invest in both arterial and freeway
simulation research. This document describes the first such
collaborative project, “Improved Simulation of Stop Bar Driver
Behavior at Signalized Intersections.”
With respect to NIATT’s own strategic plan, this project directly
supports Goal 1 and Strategy 1.2 from its University Transportation
Center Prospectus.
- Goal 1. Reduce congestion and improve safety by developing
arterial traffic management tools that can be used by
practitioners and researchers
- Strategy 1.2: Develop improved driver behavior algorithms
for congested and uncongested conditions on arterials in support
of FHWA’s NGSIM program and develop improved modeling
capabilities for arterial operations for TRANSIMS program
This project also is consistent with two example activities
listed in the Prospectus.
- Collect new vehicle trajectory data and, together with the
NGSIM arterial data set, study basic arterial traffic flow
theory and phenomena
- Develop new algorithms and make them available to private
software vendors and researchers
Two objectives have been established for this project. Each
objective will directly support the NGSIM program:
- Develop characteristics for stop line behavior at both the
beginning and end of green.
- Develop algorithm(s) for driver behavior during queue
discharge and phase termination periods at the approach to a
signalized intersection (response to signal control state
changes) using field data collected in the NGSIM project.
Task Descriptions
Seven tasks have been identified that will result in products
that will meet the three objectives listed above. These tasks are
listed below.
Task 1. Document existing driver behavior algorithms during
signal phase change states in VISSIM and in the literature for
signal phase change periods. The documentation will include the
inputs, logic, outputs, and context of application for each
algorithm. The documentation will also include a description of
any testing (calibration, validation) that has been completed
for the algorithms. Product: Technical memorandum.
Task 2. Document the characteristics of the NGSIM arterial
data sets (Los Angeles and Atlanta), including vehicle
trajectory data, vehicle point (event location) data, and signal
status data. Synchronize and combine trajectory, point, and
signal status data sets as needed. Prepare a report on the scope
of the data sets including time resolution, location resolution,
parameters recorded, parameter ranges, and parameter
distributions. Product: Technical memorandum.
Task 3. Integrate data set components into master data base
management system and publish results on web site. Make data
base available online. Product: Data sets published on-line.
Task 4. Analyze the data sets to determine driver
behavior/reactions to signal state changes (beginning of
phase/end of phase) in the context of existing driver behavior
algorithms. Consideration will be given to driver responses to
the beginning of the green interval, the beginning of the yellow
clearance interval, the all-red change interval (including red
light running). Consideration will also be given to driver
responses to other vehicles in the queue ahead of them, and to
the effect of vehicle type on these responses (e.g., heavy and
large vehicles interacting with each other and with
smaller/standard vehicles). Outputs from this task will include
statistical summaries and distributions of observed driver
response algorithm inputs and outputs for various signal state
change scenarios. Product: Technical memorandum.
Task 5. Conduct conceptual validation of the potential
algorithms, including specification of a baseline algorithm,
generation of variables of interest from the trajectory data,
development and testing of the likelihood estimation functions,
and completion of transferability tests. Product: Technical
memorandum.
Task 6. Implement and test the model within the framework of
the VISSIM and other models. Prepare a summary of these tests.
Product: Technical memorandum.
Task 7. Prepare final report summarizing the work completed
in each of the tasks, including recommendations on the
algorithms that were tested during the project. Prepare
technical papers summarizing the results of this work.
Milestones
The seven tasks listed above each include a specific product or
outcome. Each of these products or outcomes is an important
milestone for the successful completion of this project.
FHWA has agreed to set up a project oversight panel through its
NGSIM program. Meetings with the project oversight panel to review
the project status will be held during the following times:
- Month 5 (projected as January 2008) to review products
from Tasks 1 and 2.
- Month 9 (projected as May 2008) to review products from
Tasks 3 and 4.
- Month 17 (projected as January 2009) to review products
from Tasks 5 and 6 (including the draft final report).
Budget Information
UTC funds committed to this project: $85,099.
Student Involvement
One graduate student, months 1-16.
Technology Transfer Activities
Four technology transfer activities will be generated by this
project:
- The final report summarizing all work completed during
the project; the report will be published on the NIATT web
site, distributed to the NGSIM community, and presented to
the technical oversight committee.
- A database, published on line that includes the data
sets that are generated from this project and that will be
available to other researchers.
- A new driver behavior algorithm that will be made
available to researchers and private sector model
developers.
- Technical papers submitted to academic journals covering
the following topics: driver behavior characteristics at
signalized intersections, driver behavior algorithms, and
data set fidelity requirements for microscopic model
simulation.
Potential Benefits of the Project:
NIATT has led, during the past two years, the development of a
new northwest regional consortium whose purpose is to increase
collaboration between university researchers and state departments
of transportation. One of the outcomes of this effort is a
simulation collaborative, including the University of Idaho,
Portland State University, Oregon State University, and the
University of Washington. This current project comes directly from
discussions between the collaborative members and the FHWA. FHWA has
now recognized the benefits of this collaboration in the following
email excerpt, describing FHWA’s desire to reach out to the
university community to develop new research partnerships.
“As an example, the Region X UTC Consortium will be partnering
with the NGSIM program this summer to develop driver behavior
algorithms for starting and stopping behavior at signalized
intersections. The University of Idaho will lead this development
project for the Region X UTC Consortium. The UTCs will provide
funding for their research, while the NGSIM program will assist in
data collection activities, provide a framework and forum for peer
review of the research, and assist in outreach and communications
with commercial developers and the overall traffic simulation
community. The final driver behavior algorithms will be joint
UTC-NGSIM products that will be openly available to commercial
developers and the overall traffic simulation community with the end
goal of improving the quality, trust, and use of traffic simulation
models to enable better transportation decision-making.”
The results of this project will directly support FHWA’s
objective of developing new algorithms for microscopic simulation in
key areas, including the simulation of arterial traffic flow. All
parts of the NGSIM community, including practitioners, model
developers, and researchers, will benefit from the analysis and
algorithms developed from this project. Practitioners will have
access to revised models that are more reflective of driver
behavior, developers will have access to important new algorithms,
and researchers will have access both to new models and new data
sets of arterial operations.
Project status
Active Final Report
|
