The Model of Short-time Traffic Flow Prediction on High-grade Highway

In order to study the short-time traffic flow prediction on high-grade highway, the article proposed a model based on wavelet analysis and RBF neural network. Aiming to the traffic flow's characteristic of highway, the study focus on three facet: network topology, the difference of continuous flow and discontinuous flow, the flow of lanes' uplink and downlink are not equal. Thus the article use the wavelet analysis to do data preprocessing, then structure the model of short-term traffic flow prediction based on RBF neural network. The experiment result shows that the new hybrid model adapt to high-grade highway, and model considering traffic flow characteristic is better than the model which is not. Meanwhile the model has the higher precision of prediction.