TIVMM: An effective map algorithm for low-sampling-rate GPS trajectories in road networks

Abstract

In the process of map matching, we often encounter such a scenario. When there are two parallel paths between two points and the diver actually chooses the path that is longer. This scenario can be handled very well by high-sampling-rate based map matching algorithm. However, due to their limited energy budgets of the handheld GPS devices, they are more capable to provide low-sampling-rate trajectories. And the existing low-sampling-rate map matching algorithm cannot be applied directly to the scenario discussed before. Hence this paper proposes a Time Interactive Voting Map Matching (TIVMM) algorithm to resolve the problem. TIVMM algorithm considers the spatial geometry and topological structures of the road network, the speed constraints of the paths, the actual traveling time and the strength of the mutual influence. Extensive evaluations are performed and the results show that the proposed algorithm can provide higher accuracy compared to the existing approaches.