Traffic jams prediction method based on two-dimension cellular automata model

Urban traffic jams is a prevalent problem affecting many cities around the world. One challenge in this research is how to predict the traffic jam accurately and in real time effectively. This paper proposes a traffic jam prediction method based on two-dimension cellular automata model, which is inspired by the famous Biham, Middleton and Levine (BML) model. This method is supposed to be effective in describing the different characteristics of the urban traffic networks, so as to predict the accurate positions of the traffic jams at the intersections. The main research includes that: (1) we propose a practical approach to mapping the urban traffic topological structure into a modified BML (M-BML) model; (2) we propose the solutions to the conflict points and the fuzzy points in the mapping strategy from the M-BML model to the urban traffic road networks. Extensive experiments are carried out, which reveal that when the vehicle flow density ranges between 0.3 and 0.7, the traffic jams prediction accuracy is 81.25% by the proposed M-BML. A real project example is also exploited with our method, which further proves our method´s accuracy and correctness.