Spatial temporal mining for urban map matching

In recent years, the widespread of GPS devices produce huge amount of urban trajectories. Due to the complex structure of road network in urban area, many issues are not addressed in existing work, such as high volume of road network, diverse functionalities of roads, and high variance of road dynamics. In this paper, we propose a novel modularity-based map-matching algorithm called Urban Map-Matching (UrbMatch) utilizing urban GPS trajectories. UrbMatch considers (1) the spatial-temporal betweenness and (2) the marginal velocity of each road segment, to improve the accuracy of map-matching. Based on the results of spatial-temporal mining, a road network is decomposed into several sub-networks such that the map-matching task can be divided into several smaller sub-tasks and run in parallel. Accordingly, the running time can be reduced. We compare UrbMatch with an existing global map-matching algorithm using real-world dataset. The results show that our proposed approach not only are faster than state-of-the-art global map-matching method over 100 times but also outperforms other map matching techniques in terms of accuracy.