Expected time of arrival model for school bus transit using real-time global positioning system-based automatic vehicle location data

The school bus is a major transportation mode for students in Canada. Unexpected delay of a school bus may be a major source of inconvenience for students and their parents. Accordingly, the provision of timely and reliable information on the expected arrivals of school buses would be of great benefit to them. This study develops an expected time of arrival (ETA) model for school buses. The model predicts arrival time from the input of two categories: the last several days' historical data and the current day's operational conditions. An operational strategy is additionally incorporated into the model to reduce the risk that an overestimated arrival time can result in missing the bus. This study evaluates the model using data collected from real-world operations of school buses on which a global positioning system-based automatic vehicle location (AVL) system is installed. The proposed model consistently shows lower levels of prediction error than moving average and regression approaches. With the operational strategy, the model provides a sufficiently reliable service in which approximately 99%-100% of students do not miss the bus, with the tolerable wait time of 162-177.