Study on Traffic Flow Base on RBF Neural Network

Abstract

According to charging lathing standard type classification volume of traffic blowing the expense at enormous and the data is inaccurate, But the volume of traffic may obtain easily according to the traffic survey classification from each kind of intelligent toll station. The function neural networks method will charge with because of radial base the rate of flow data change becomes volume of traffic inquiring into a data, Building-up changes a model, Using the MATLAB programming, Obtaining each kind of vehicle type error distinguish ratio and total error distinguish ratio. It can utilize fully thereby advantage of the fee-collecting station, Every fee-collecting station all has the precise writer that various motorcycle type passes in the process charging, it can cut down the cost getting the traffic survey data, Enables the traffic survey the data to obtain the full use.