Prediction of traffic flow based on the EMD and wavelet neural network

Abstract. The operation of the urban traffic exist a high degree of complexity and randomness. The key of Intelligent Transportation System is the real-time and accurate traffic flow prediction. Taking effective measures in a timely manner would prevent the occurrence of traffic accidents since traffic congestion brings much traffic inconvenience to people. Real-time traffic flow prediction plays a significant role in easing traffic congestion and guiding convenient travelling. Therefore, considering the characteristics of traffic flow, this paper presents a neural network, wavelet analysis method, and the EMD and wavelet neural network method respectively. Three methods are utilized in simulating the same set of traffic data and then the most effective way of solving traffic congestion is obtained by taking contrast analysis of simulation result.