Short-time Traffic Flow Prediction Based on Wavelet Neural Network

Abstract

With the rapid development of social economy, traffic congestion, frequent traffic accidents, and traffic pollution have become the focus of attention. As an important part of intelligent transportation system, traffic guidance system and traffic control system play a great role in improving road congestion, Therefore, the short-time traffic flow prediction is of great significance to urban traffic system. Based on the investigation of traffic flow characteristics, we analyze the advantages and disadvantages of available short-time traffic flow forecasting methods. In view of the nonlinear, time-varying and uncertain characteristics of traffic flow, wavelet neural network is selected as the traffic flow forecasting model combining the self-learning and self-adaptive characteristics of neural network. To solve the problem of short-time prediction and accuracy of traffic conditions, wavelet basis is used as the characteristic function of pattern recognition to optimize the parameter and error space of traditional BP neural network, the short-time prediction of traffic state is realized. In this paper, the wavelet neural network model is established, the traffic flow sample set is constructed, the sample data is trained to convergence, and a certain number of sample data are selected for testing. The test results show that the wavelet neural network not only has high prediction accuracy but also has fast convergence speed, good real-time performance, and has certain application value.