Freeway Accident Detection Model Based on Support Vector Machine

In this paper, the key technology of expressway accident detection is studied in order to set up an automatic accident detection system and promote the efficiency of accident rescue. On the bases of characteristic analysis of the existing models, the expressway accident detection model based on support vector machine (SVM) is put forward Furthermore, the performance influenced by many factors such as upstream input or upstream and downstream input, different kernel function and input aggregate that compose of different input features are experimented with database which established by EAD-simulations system. The study shows, that the excellent performances of the model are demonstrated by contrast with California 8# Model. Its' detect rata raised 179%, error detect rata dropped at 0.50% and average detect time cut down 81%. In addition, the input characteristic aggregate which made up of occupancy and velocity in upstream is received.