Real time estimation of bus arrival time under mobile environment

Recently, several city governments of Korea developed Bus Information Sytem (BIS) as one of initiatives of ITS deployment efforts. The core function of BIS is to deliver the arrival time information of incoming transit bus at every bus station. In this research, we propose a computational model for estimating transit vehicle's link travel time from data transmitted through wireless communication channel and outline major components of real BIS server program architecture. Our algorithm is based on a neural network algorithm tailored for each period of the day under mobile environment. In this model, link weights are updated every 7 days in order to provide feedback control based on bias between estimate and real measurement. Our system can be implemented using reasonable space resources and should provide good estimates of bus arrival time.