Traffic Flow Information Prediction Based On Information Fusion And Neural Network

Intelligent Transportation System (ITS) is an important sign to enter the information age. Accurate traffic information collection and forecasting is the key to ITS.This paper summarizes the advantages and disadvantages of traffic flow collection equipments, and designs the traffic flow testing equipment based on anisotropic magneto resistive (AMR). Because locations of traffic flow testing equipments are scattered and some of equipments are in the distance, traffic collection device is designed based on general packet radio service (GPRS). Followed by a discussion of the neural networks and information integration in the relationship and integration algorithm, a detailed study of radial basis function (RBF) neural network prediction model is established based on the spatial correlation of the RBFNN forecasting model integration. The measured traffic flow data through the integration of prediction model for training and simulation shows that this approach is reasonable, so that neural network prediction accuracy and stability of traffic flow is improved. Finally, this paper completes the collection of traffic flow information to predict the preliminary design of system' software.