基于小波分析的随机交通流组合预测方法研究

针对单一交通流预测方法对短时随机交通流预测适用性不强的问题,为适应交通诱导和信号控制的需要,提出了基于小波分析的短时交通流RBF神经网络及Markov链组合预测方法。在充分分析短时交通流随机特性的基础上,建立了短时交通流组合预测模型,并进行了实例仿真分析,结果表明组合预测较单纯神经网络预测方法预测精度显著提高,在交叉口实时交通控制中具有较好的适应性。

Stochastic Traffic Series Time Combination Forecasting Based on Wavelet Analysis

In order to adapt to the needs for traffic guidance and signal control,a short-time traffic flow RBF neural network and Markov chain combination prediction method based on wavelet analysis was put forward to solve the problem that a Single traffic flow prediction method possessed weak applicability for short-time stochastic traffic flow prediction.Under the intensive analysis of stochastic characteristics of short-time traffic flow,the short-time traffic flow combination prediction model was built,and moreover,example simulation analysis was done.The results show that the precision of combination prediction method compared to pure neural network prediction one is significantly improved,and the combination prediction method has better adaptability in the real-time traffic control of intersections.