交通流实时预测的混沌时间序列模型

针对城市交通流普遍存在的混沌特性,介绍了一种改进的加权一阶局域预测模型,并将其应用于交通流实时预测中。为了进一步提高算法的精度与速度,对最优邻域的点数进行动态选择,通过改进,使之成为一种鲁棒性强、预测精度高的实时预测算法,并能有效地用于短时交通流的预测问题中。仿真结果表明:该算法完全满足实时交通流预测的需要,为交通信号智能控制和交通流诱导奠定了坚实的基础。

Chaotic Time Series Model of Real-time Prediction of Traffic Flow

Aimed at the chaotic characteristic in urban traffic flow,authors introduced an improved prediction model of weighted one-rank local-region and applied the model to real-time prediction of traffic flow.In order to improve accuracy and speed of the algorithm,the number of optimum neighbor point was selected dynamically.By improvement,it became a real-time prediction algorithm with strong robustness and high accuracy.The method can be used for a brief time prediction of traffic flow.Simulation results indicate that the algorithm can meet real-time prediction of traffic flow completely and provides the massive foundation for traffic signal control and induction of traffic flow.