Regression-based approach for bus trajectory estimation

The path travel time between trips are proved to be strongly linear correlated, and the regression parameters are gradually converging to a steady value with buses travel on the route. A regression-based method for bus trajectory estimation at downstream stops was proposed, and its performance on estimation accuracy was compared with the historical-based model. The research results show that the proposed regression-based model shows a conditional higher accuracy than the historical-based model on trajectory estimation, and a two-stage model for trajectory estimation was suggested.