Abstract
Early and precise incident detection contributes early provision of information and incident handling to mitigate negative impacts of traffic incidents. Automatic incident detection algorithm has been introduced as freeway incident management system to detect incidents more early and precisely. Several conventional studies developed incident detection algorithms using observation data from roadside traffic detectors. However, it is difficult to install traffic detectors in the entire road networks because of their costs of installation and maintenance. Probe vehicle data enable us to observe traffic flow state. It is expected that the incident detection can be utilized without traffic detectors. This study developed incident detection algorithm based on shockwave theory. By using the probe vehicle data, the algorithm detects tail of congestions and estimates capacity reduction at the bottleneck. The detection rate and false rate of the models are examined using a traffic flow simulator, and the performance measures of the proposed methods are compared with those of previous methods.
