Research on Traffic Flow Base on Neural Network

The traffic flow according to toll vehicle type classification spends greatly and the data is inaccurate. But the traffic flow according to traffic investigation classification may obtain easily from intelligent toll station. Using neural networks ways transform charge flow into traffic investigation flow, building-up transformation model, programming in MATLAB, Obtaining error ratio of each kind of vehicle type and total error ratio. It can utilize fully the advantage of the toll-gate. Every toll-gate has the charge record to all kinds of vehicles type in the process charging. The research can cut down the cost of the traffic survey,and make full use of the traffic survey data.