Modeling and tracking of public transit in urban environments

Advanced Traveler Information Systems provide real-time transit information concerning location of vehicles and the predicted arrival times. Such information is based on estimates of the state of the vehicles from on-board sensors, and models of vehicle motion through traffic. In this paper, we propose new models and algorithms to track public transit in urban environment. The new models incorporate side information such as average speed on a segment of road, quality of GPS estimation, desired path and road interconnectivity. This side information is either known apriori or can be learned in real time through communication among vehicles. We develop a model of interconnected road segments, each of which has a different dynamical model for the possible motion. The resulting model is a hybrid system, with state dependent transitions as vehicles traverse from segment to segment. We develop new estimation algorithms that incorporate the state-dependent switching among models by using adaptive predictors to estimate switching times. We show through simulation the relative advantages of the different estimation approaches.