A real-time update method for bus stops and routes based on historical bus data

Precise bus stops and trajectories are crucial for public transport. However, acquiring accurate information on bus trajectories and stop locations faces the situation of point drift of the global positioning system (GPS), low sampling frequency, and the absence of the latest bus information. Thus, this study developed a method, based on Dumpster Shafer evidence theory and K-Shortest Paths, to update the latest bus stops and routes and automatically remove abandoned bus stops by taking advantage of historical GPS data. To speed up computation, we divided the road network into meshes, dynamically generating a local road network. The effectiveness of the method is validated on Qingdao bus data, and the experiment demonstrates that the method is robust in pinpointing the bus stops and the trajectories of different bus routes.