POP: A Passenger-Oriented Partners matching system

Sharing one taxi by more than one person is treated promising, since it enables us to take a taxi in rush-hour more conveniently. Hence, we develop POP, a prototype system to find appropriate partners to share a taxi with a given passenger. The framework of POP includes two phases, namely offline preprocessing and online matching. During the offline preprocessing phase, it constructs an R-tree index for road network to speedup data access and computes average travel time for each road segment based on history trajectory data, while during the online matching, it tries to find appropriate partners to a given passenger which aims to save time as much as possible. We also propose a simple pricing method to allocate fee between passengers.