Prediction Of Traffic Flow Based On Grey Theory And BP Neural Networks

Recently with the progress of the traffic technology, the means of transportation is proved also and the application of ITS ( Intelligent Transportation System) is more extensive. The traffic inducement subsystem is the primary contents of transportation management subsystem. The precise prediction of short period traffic flow is as an important role in giving out the agreeable inducement and controlling of the next time to realize rather route choice and decrease the traffic jam. Previously the model of prediction of traffic flow have disadvantage that the time of operation is too long and the precision is low.In this paper, we aim at overcoming the disadvantage of the prediction model that referred above. The merit of grey model GM(1,1) is that the arithmetic is simple and it can use few data to construct the model, these bring the facility of constructing model but the grey model to predict the result of the fluctuating system cursorily and the precision rate decreasing with the time going on. In the end we construct a combination model. During the operation we improve the BP arithmetic to avoid to getting into the partial least point.By using traffic flow data of some road we construct a example and by use of Matlab to program .We contrast the new combination model to the single grey and BP neural network model. The result show that the combination model can take full advantage of every single model and avoid the disadvantage and the result of prediction is super than that single model to draw.