基于混沌和RBF神经网络的短时交通流量预测

A Short-term Traffic Flow Forecasting Method Based on Chaos and RBF Neural network

针对传统的应用数学模型方法在短时交通流预测精度和实时性方面存在的问题,论文从非线性时间序列的角度对短时交通流量预测进行探讨,提出采用基于混沌理论的RBF神经网络预测方法。首先在采用小数据量的Lyapunav指数计算方法判定交通流存在混沌的前提下,对交通流量数据进行相空间重构。构建了RBF神经网络,并对模拟产生的Lorenz和Rossler混沌时间序列数据以及实际采集的高速公路交通流量数据进行了仿真研究。结果表明,该方法对模拟产生的混沌时间序列具有很好的预测效果,在交通流量的短时预测上也具有较高的预测精度。

Aiming at the prediction precision and real time problem using mathematical model for short-time traffic flow, from the point of nonlinear time sequence, this paper discussed short-time traffic flow prediction and presented a prediction method using RBF neural network based on chaos theory. On the premise that small data Lyapunav＇s exponent was used to decide that chaos exists in traffic flow system, we first performed phase space reconstruction for traffic flow data. RBF neural network was then constructed. Finally, we performed simulation using chaotic time sequence data generated by Lorenz and Rossler and the real measured expressway traffic flow data. Simulation results show that the proposed method hag effective prediction results for simulated chaotic time sequence and highly prediction precision in short-time traffic flow.