Evaluation and Refinement of Incident Detection Strategies for Major Urban Arterial Streets
or Dr. David Shen, (305) 348-1869, firstname.lastname@example.org
Incident detection along limited access facilities in FDOT districts where ITS projects have been implemented is currently done via surveillance of CCTV cameras by operators stationed at Transportation Management Centers (TMCs). However, as ITS service coverage areas continue to expand, it will not be possible for operators to visually monitor all locations where CCTV cameras are installed. Consequently, for most locations it is necessary for operators to visually monitor only when a potential incident is detected. Since these locations are not visually and continuously monitored, incidents must be detected automatically. This is usually done through incident detection algorithms that make use of real-time traffic data feed, which generally includes vehicle volume, speed, and occupancy.
While much research has been done over the past decades on freeway incident detection, similar research on arterial streets remains largely at the initial stage of development and testing. Freeway incident detection techniques are not generally applicable to arterial streets, which feature a variety of traffic controls, turning movements, easier lane changing and rerouting of vehicles, thus, a much more complex environment for incident detection. Existing arterial incident detection strategies have ranged from the traditional rule-based techniques to more advanced techniques that that make use of image processing technologies, artificial neural networks, data fusion, etc.
This project will review and evaluate existing strategies for incident detection on urban arterial streets and identify one or more that can be adopted for refinement and implementation by FDOT. The research will include system evaluation using one or more existing traffic simulation software, calibration of parameters, field tests, and refinement of methods, as appropriate.