基于BP神经网络的高速公路短时交通流预测

Highway Dynamic Traffic Flow Prediction based on BP Neural Network

以高速公路交通流预测为研究对象,简化了高速公路宏观动态交通流模型,利用Matlab神经网络工具建立模型并进行网络训练与预测。对嘉兴站附近高速公路交通流数据进行了采集、建模和预测。从预测结果中得知,基于BP神经网络交通流预测模型具有很高的可靠度,该简化的交通流模型更为简练,预测结果亦可以点带面地面描述该站点一定空间及时间范围内的交通流情况。

This paper takes highway traffic f Iow prediction as object of study,simplifies the highway macroscopic dynamic traffic flow model and establishes a BP neural network model to train and predict based on Matlab neural network toolbox.It models and predicts real data collection from traffic flow nearby Jia Xing station. From the predicted results,this BP neural network model is proved to be highly reliable. It carries out that the simplified model of traffic flow is more concise,predictions can also describe a certain space and time of the site＇s traffic situation through a point to an area.