Predicting Bus Real-time Travel Time Basing on both GPS and RFID Data

Abstract

Instantaneous and accurate prediction of bus arrival time can help improve the quality of bus-arrival-time information service, and attract additional ridership. On the basis of bus running processes, a self-adaptive exponential smoothing algorithm is proposed to predict the bus running speed based on the short-term running speeds of taxis and buses available. And a bus travel time prediction model is proposed, in which the delay caused by the signal control and the acceleration and deceleration are considered. The research results show that there is a significant linear correlation between speeds of buses and taxis on the same link during the same time period, and the overall performance of the RFID-data-based model is superior to that of AVL- data-only-based model, regardless of whether the traffic congested or not.