The Bus Travel Time Prediction Based on Bayesian Networks

The prediction of bus travel time is one of the key of public traffic guidance, accurate bus arrival time information is vital to passengers for reducing their anxieties and waiting times at bus stop, or make reasonable travel arrangement before a trip. Research aim at bus travel time prediction is comprehensive at home and abroad. This paper proposes a model to combine road traffic state with bus travel to form the Bayesian network, with a lot of historical data, the parameter of network can be achieved, through estimating the real-time traffic status, so as to predict the bus travel time. We introduced Markov transfer matrix to forecast the traffic state, and substitute the estimate state value into the joint distribution of bus travel time and state, the real time bus travel time predicted value can be obtained. Bus travel time predicted by the proposed model is assessed with data of transit route 69 in Guangzhou between two bus stops, the results show that the proposed model is feasible, but the accuracy needs to be further improved.