T-reduce: Route-aware mobile trajectory data reduction in transportation management systems

Due to the increasingly popular e-commerce and rapidly growing global market of delivery services, transportation management system (TMS) has been playing a critical role in minimizing transportation delay and cost as well as improving reliability. However, how to efficiently keep track of running vehicles and judiciously store their corresponding trajectories become a key issue in the presence of scarce wireless bandwidth and limited storage space. In this paper, we propose a novel mobile trajectory reduction scheme, called T-Reduce, to reduce the size of trajectory data stored in a TMS server. We first develop a path refinement operation to adjust raw location update data transmitted from vehicles to a given road topology as close as possible to improve their location accuracy. Second, we develop a route matching operation consisting of three sub-operations, route extraction, route recognition, and trajectory generation, for the server to identify and extract a set of routes by decomposing the received location update data. Then, the server can store each trajectory as a set of corresponding route ids instead of storing entire location update data. We set up a small-scale testbed, implement the proposed scheme with a real-world trace data collected from a logistic company, and conduct extensive experiments for performance evaluation and comparison. The experimental results show that the proposed scheme can significantly reduce the average trajectory error and route information up to 26.4% and 88%, respectively, compared to that of two prior trajectory-based and corner-based approaches. The proposed approach can also achieve up to 5.72 times cost efficiency compared to the prior approaches.