Paper title: Incident detection on freeways: a Bayesian network approach

words): This paper presents a novel approach to incident detection on freeway. The proposed incident detection algorithm is capable of detecting lane-blocking incident promptly as well as reporting incident location and duration. The algorithm consists of two major components: (1) data processing and (2) incident detection. Data processing is designed to deal with site specific traffic measurements. One standard traffic case, which contains states of selected traffic parameters, is generated using smoothed lane volume, occupancy and speed at each detection interval. Incident detection is performed by a dynamic Bayesian network through two-way reasoning using traffic cases. One step congestion and incident detection are fulfilled in the Bayesian network. The proposed algorithm is tested using simulated incidents. The results are very encouraging in terms of detection rate, false alarm rate, and mean time to detect. The Bayesian network based approach is considered promising.