BUS ARRIVAL TIME PREDICTION BASED ON ERROR WEIGHTED OF HISTORICAL AND REAL-TIME DATA

摘要

Transportation information distribute system is an important component of intelligent transportation system, bus arrival time prediction technique is difficult and is the keystone of the system. Provide better public transport information services, can improve the chances for choosing the bus and reduce the traffic congestion. This paper propose a method by analyzing historical data to judge public transport running stability, and analyzing real-time GPS data with Kalman filter to predict arrival time, the two predictions are merged to be the final release information at last, through error weighted method. Experimental results show the good accuracy of this method and the fast computing speed, it also easy to physical implementation and promotion.