Research On Bus Arrival Time Prediction And Bus Transfer Mechanism

With the rapid development of intelligent city, intelligent traffic is also getting moreand more attention and bus arrival time prediction has become one hotspot to theresearchers in recent years. Accurate and real-time prediction of bus arrival time can notonly help travelers to choose a better route, but also provide scientific proof for the trafficdepartment to manage scientifically and make a reasonable scheduling.This dissertation focuses on the problems of bus arrival time prediction and bustransfer, and puts forward a dynamic prediction model of bus arrival time and a dynamicbus transfer model based on the time prediction, the main content includes several aspectsas follows:Firstly, it establishes a dynamic prediction model of bus arrival time based onvarieties of static and dynamic factors in order to explore a variety of static and dynamicfactors influencing bus arrival time. The experimental results show that the dynamic modelgets a higher accuracy than the baseline.Secondly, it proposes an adaptive model of bus arrival time prediction based on thevolatility of bus arrival time to increase efficiency and further improve the accuracy ofdynamic prediction model. It calculates bus arrival time volatility at three layers:1)different time periods in different dates;2) different sections in different dates; and3)different sections in different time periods to illustrate which factors affect the accuracy ofbus arrival time prediction. The further optimization for adaptive prediction divides theprediction route into different combinations, considering the volatility of adjacent sections.The experiment shows that our adaptive model successfully reduces the computation timeof the model and further improves the model’s prediction accuracy.Finally, it puts forward a bus transfer model based on the bus arrival time predictionaiming at the problem of real-time and low reliability of the existing bus transfer model.The model takes into account the impact of time periods on the transfer waiting time andthe transfer mode, effectively enhancing the reliability and real time of transfer model.