交通流实时预测的混沌时间序列模型  

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

作　　者：臧利林[1] 贾磊[1] 杨立才[1] 刘涛[1] 

机构地区：[1]山东大学控制科学与工程学院,山东济南250061

出　　处：《中国公路学报》2007年第6期95-99,共5页China Journal of Highway and Transport

基　　金：国家自然科学基金项目(60674062);教育部高等学校博士学科点专项科研基金项目(20060422054)

摘　　要：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。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 realtime prediction of traffic flow completely and provides the massive foundation for traffic signal control and induction of traffic flow.

关 键 词：交通工程 交通流 混沌时间序列 实时预测 

分 类 号：U491.14[交通运输工程—交通运输规划与管理]