Short-term traffic flow parameters prediction based on multi-scale analysis and artificial neural network

作者：Mei‐Ling Huang, Baichuan Lu · 年份：2010 · DOI：10.1109/car.2010.5456866 · 被引用次数：8 · 研究领域：Traffic Prediction and Management Techniques、Evaluation Methods in Various Fields、Advanced Sensor and Control Systems

In analyzing the nonlinearity characteristics and strong interference of traffic flow parameters, a new approach has been proposed for the prediction of traffic flow parameters. First, multi-scale analysis is used to decompose the sequences of traffic flow parameters into the low and high frequency ones and restore them according to the reconstruct principle of wavelet coefficients. Then artificial neural network is used in multi-scale forecast of these coefficients, with gene algorithm for optimization. Finally, some real detected traffic data are used to testify the precision of the model. The results show that the model can produce more accurate predictions than with traditional artificial neural network model.