Institute of Transportation Studies and
Department of Civil & Environmental Engineering
University of California Irvine
Irvine, CA 92697-3600 USA

Funded by a grant from the National Science Foundation. [NSF Summary]

September 15, 2003 through March 15, 2005

The project consists of three overlapping research tasks. The first examines the design and performance of the distributed travel database and prediction system using simulation experiments. This task will develop and use a new modeling framework that integrates a microscopic traffic model and a multi-agent, activity-based travel demand model. The second task explores how such predictions can be applied to improve the design, management, and operation of the transportation system. For instance, local traffic controllers can move beyond turning fractions, and instead predict whether and where a platoon will disperse as it moves through the traffic network. Local intersection control schemes can incorporate the global consequences of their control measures. Path flow predictions made by the system can also serve as a direct input to existing dynamic traffic assignment formulations. The third task administers a small survey of traveler attitudes on privacy and utility issues involved in such a system. The results will help ascertain the feasibility of the system and balance the functional requirements with stakeholder and end-user interests in security and privacy.

The broader impacts of the research include the following:

1. In teaching, training and learning, the project lends itself to small, discrete projects ideal for completion as undergraduate or graduate class projects ranging from the study of wireless technology to full scale travel simulation.
2. To underrepresented groups, as all travelers could benefit from the optimal use of network capacity resulting from the use of the system.
3. In developing infrastructure for research and education, in that the simulation environment will become a novel resource available to any future student research.
4. From disseminating the results to enhance scientific and technological understanding, via publication of all findings in conference presentations and journal articles, and through a project web site.
5. To the society at large, where the impacts are many and varied. The system hinges upon traveler participation, which turns traffic control away from a command and control paradigm towards one of shared responsibility.

The proposed activity will be conducted in two phases. During the first phase, the simulation framework will be implemented including the basic features of the distributed travel demand database system. The second phase will involve experimental analysis of a set of applications that leverage the distributed database infrastructure.

Phase 1: Activity-based Traffic Operations Simulation System
During this phase of research, the prototype integrated activity-based traffic operations model will be extended and refined. Preliminary research tasks include:

1. Add a distributed database plug-in to the ParaDyn model: The ParaDyn simulation package is easily extended through a public API. For the purposes of this research, the distributed database function can be implemented simply as tags (cookies) written to a vehicle as it passes traffic controllers in the network. Read/write access to these tags will be made available for the implementation of traffic management applications that can leverage the data.
2. Link the agent-based activity simulation system to ParaDyn: Again, these models can be linked using the ParaDyn travel demand API, which allows external control over the vehicles generated in the system. The only modification required in the activity simulation is the creation of a travel activity implementation object that calls the ParaDyn trip generation API.
3. Conduct database-specific experiments: A series of experiments will be designed to evaluate the core hypotheses related to the distributed database system itself, independent of the traffic management applications that will use it.
4. Conduct a traveler attitude survey: Using UCI ITS's in house wireless enabled GPS data collection system, called Tracer, a survey will be designed which will strive to simulate a realistic, operating traffic cookie scheme for a small number of households (100). The preliminary design at this stage is to track each vehicle of all participating households for a week. At the end of each day, participants will be asked some simple questions about their comfort level with being tracked, the relative degree of sensitivity they have for different parts of their daily travel, and so on. It is expected that travelers will feel uncomfortable with publicizing data on their destinations, but will be less concerned about the en-route portions of their trip. An alternate survey methodology might be tested in which travelers are also provided with traffic information while using the device. While travelers may hesitate to surrender trip information for free, they may voluntarily release that data if they see it provides them with direct benefits.

Phase 2: Evaluation of Traffic Management Applications
During phase 2 of the proposed research, two traffic management applications will be evaluated. These research activities are generally independent and may be carried out in parallel. The first is developing a globally optimizing distributed traffic control system; the second is to enhancing DTA with high resolution data. The investigator was directly involved with FHWA DTA evaluation project, which compared the performance of DynaMIT and Dynasmart-X. If time and resources permit, an effort will also be made to link the travel demand database to one or both of these models.

It is anticipated that a post card survey of a targeted population will be used. Tentatively, UCI faculty, staff, and students who drive to and park on campus will be selected. A twp part post card will be distributed to drivers entering selected parking lots and structures (or left on windshields). One part will explain the purpose of the survey and ensure respondent confidentiality (there will not be any personal information whatsoever on the postcard. The second part will contain a small number of questions and will be returned by mail (or other means such as drop boxes).

When the final design is completed, the two part survey will be posted to this web site.

Related ITS Research Projects:

• ATMS Testbed
• ZEVNET -- A Shared-use Station Car Program
• Autonet -- Ad Hoc Peer-to-Peer Vehicle Communications
• NSF/USDOT -- Autonet

Project Working Papers and Reports