论文部分内容阅读
交通流预测已成为智能交通的重要组成部分,针对短时交通流的非线性和不确定性,文中根据实际交通流中存在的混沌,利用C-C方法和小数据量法对交通流混沌进行了分析,在交通流混沌时间序列相空间重构的基础上构建了基于粒子群优化神经网络的单点单步预测模型,运用该模型对实际采集的美国加州城市快速路交通流数据进行了仿真研究,结果表明,该预测模型具有较高的预测精度,能够满足智能交通控制和诱导的需求。
Traffic flow prediction has become an important part of intelligent traffic. Aiming at the nonlinearity and uncertainty of short-term traffic flow, this paper analyzes the chaos of traffic flow based on the chaos in the actual traffic flow and the CC method and small data volume method , A single-point single-step prediction model based on particle swarm optimization neural network was constructed on the basis of phase space reconstruction of traffic flow chaotic time series. The model was used to simulate the real-time traffic flow data collected from California urban expressway in the United States. The results show that the prediction model has high prediction accuracy and can meet the needs of intelligent traffic control and induced.