Road Safety Evaluation from Traffic Information Based on ANFIS

With the rapid growth of the number of various vehicles, the ratio of the traffic accidents to vehicle number is increasing greatly. To improve road safety, it is necessary to monitor and evaluate the traffic safety degree first of all. The factors that influence the road traffic safety include the basic traffic parameters such as velocity and density, the peccancy behaviors such as overspeed, overweight and retrogradation and the weather conditions such as the plane visibility on the road. In this paper, an index, level of safety (LOS), is defined to indicate the road safety extent of traffic state. It is a continuous number and should much fit human perception on safety. In this paper, a new method is present to evaluate LOS on freeway based on ANFIS (adaptive-network-based fuzzy interference system). ANFIS applied in this paper is based on zero-order Sugeno fuzzy model embedded into a framework of adaptive networks. In the evaluating system the inputs are mean velocity, mean density and the plane visibility on the road detected, and the output is LOS. After a simulation with MATLAB, the traffic safety evaluation system based on ANFIS is put up and a series of fuzzy logic rules are trained. As a result, the evaluating system based on ANFIS has the adaptive capacity and will be helpful to enhance the safety level on the road.