Intelligent Traffic Incidents Detection Method in Freeway Corridors

Traffic incidents detection has become an important issue in the past decades, which would result in a large amount financial cost and considerable reduction in traffic efficiency. Most previous studies mainly pay their attention on automatic incident detection in freeway corridors based on the traffic data collected by inductive loops or video monitors. In this paper, we have presented a support vector machines (SVM)-based method to confirm whether a traffic incident occurs or incident-free in freeway corridors with the utilization of the real-time traffic data collected and transmitted by wireless sensors. We also have extracted the most crucial traffic variables as features and utilized them into the SVM model for data process. Several experiments have been conducted to evaluate our method's performance compared with a representative relevant work, in terms of detection rate, mean time-to-detect and false alarm rate. The experimental results depict that our method could obtain better performance in most case.