Research On Fusion Event Detection Algorithm Based On D-S Evidence Theory

With the constantly growth of the urban scale and rapid increasing of population, more and more big cities are suffering from bad traffic circumstances, among all the problems traffic jam is most significant. There are many reasons for traffic congestion; the basic one traffic accident which has a huge impact on urban traffic, not only caused casualties and property losses but also lead to traffic congestion and even the collapse of the local or the entire road network. In order to detect traffic incidents effectively and reduce the impact of traffic incidents at utmost, advanced event management system and effective detective methods should be introduced to the road management and monitoring. However due to the influence of different acquisition environment and the various signal transmission circumstance or some other reasons, the single event detection algorithm detection effect is not ideal, prone to the phenomenon of false and missing detections. Based on the factors discussed above, this paper proposes a fusion event detection algorithm based on D-S evidence theory, therefore to improve the detection performance.Firstly, according to the change law of traffic flow and historical traffic data, the paper proposes a multi parameter California algorithm, section flow-velocity algorithm and neural network algorithm. Multi parameter California algorithm uses the traffic flow data share while the section flow-speed algorithm uses velocity and flow data. The reason we use different traffic flow data is that when we miss one of the traffic data we may use another algorithm to finish the detection. The neural network algorithm has its own advantage in processing of multidimensional nonlinear traffic flow data. Thus in this paper, we choose the neural network algorithm as a sub algorithm fusion algorithm. We also make sure of its higher detection rate and lower false alarm rate through the actual data.Secondly, this paper presents the fusion event detection algorithm based on D-S evidence theory and combines three single sub algorithms of detection results using D-S evidence theory fusion. Also the article soundly introduces the fusion rule of D-S evidence theory and decision rules.Finally, in order to verify the effectiveness of the algorithm, this paper uses the traffic flow data of Beijing Urban Expressway. From what has been discussed in the paper, we may safely find that the fusion event detection algorithm proposed in the article is better than the other three event detection algorithms, especially in reducing the false alarm rate which reflects the high detection performance.