Prediction of Traffic Flow Based on Cellular Automaton

Abstract

Traffic flow forecasting is an important foundation for intelligent traffic system control and guidance, while microscopic traffic flow model plays an important role to reproduce the basic characteristics of traffic flow and to be an important part of traffic control. Based on the NS cellular automaton model, using grey model of Markov residual modification, and introducing the prediction theory of grey envelope, a new grey envelope prediction model has been established. Through simulation experiment, the predicted value of average speed of every minute has been obtained by the proposed model, and meanwhile compared with Kalman filtering model and traditional grey prediction, the results have shown that there is better precision in the proposed prediction model, which can solve problems of prediction accuracy, such as time series of strong randomness and volatile sequence.