Prediction and evaluation of traffic states at signalized intersections

Abstract

The parallel transportation system based on the method of ACP (Artificial systems, Computing experiments, Parallel Control) will promote the level of city traffic intelligent decision and scientific management. A key problem in the system is how to design a computing experiment method to predict and evaluate the traffic state by real-time and accuracy. This paper introduces the discrete-time queuing model to analyze the traffic flow at the signalized intersection and gives the evaluation conditions of the traffic state. Then, the evaluation conditions are applied to judge the traffic state based on the prediction data of traffic flows from the grey model. Experiments show the method is effective and feasible.