Abstract
This research involved an important extension of existing field-implemented and tested PATH research by the authors on individual vehicle reidentification, to develop methods for assessing freeway and arterial (and transit) system performance for the Caltrans PeMS (Performance Measurement System). PeMS has been adopted by Caltrans as the standard tool for assessing freeway system performance, but lacks capabilities for assessing arterial and transit system performance, and strategies that combine freeways, arterials and/or transit and commercial vehicle fleets. It was shown that the research methodology of this project could directly address these limitations in PeMS. A systematic investigation was conducted of anonymous vehicle tracking using existing inductive loop detectors on both freeway and arterial street facilities combined with new, low-cost high-speed scanning detector cards (that were utilized by the authors in PATH TO 4122) to meet the needs of PeMS. Both field data and microscopic simulation were utilized in a major travel corridor setting, using the Paramics simulation model and field sites that were part of the California ATMS (Advanced Transportation Management Systems) testbed network in Irvine, California. The experience and insights of the research team obtained from extensive previous and current PATH research on vehicle reidentification techniques for single roadway segments and signalized intersections was used to investigate and develop methods for tracking individual vehicles (including specified classes of vehicle such as buses and trucks) across multiple detector stations on a freeway and an arterial street network to obtain real-time performance measurements (including dynamic or time-varying origin-destination (OD) path flow information such as path travel time and volume). This study presented a framework for studying the feasibility of an anonymous vehicle tracking system for real-time freeway and arterial traffic surveillance and performance measurement. The potential feasibility of such an approach was demonstrated by simulation experiments for both a freeway and a signalized arterial operated by actuated traffic signal controls. Synthetic vehicle signatures were generated to evaluate the proposed tracking algorithm under the simulation environment. The PARAMICS microscopic simulation model was used to investigate the proposed vehicle tracking algorithm. The findings of this study can serve as a logical and necessary precursor to possible field implementation of the proposed system in freeway and arterial network. It is also believed that the proposed method for evaluating a traffic surveillance system using microscopic simulation in this study can offer a valuable tool to operating agencies interested in realtime congestion monitoring, traveler information, control, and system evaluation. Furthermore, the automatic vehicle classification system developed in this study showed very encouraging results.
