Fast and robust map-matching algorithm based on a global measure and dynamic programming for sparse probe data

The location data from positioning devices such as those utilising global navigation satellite system (GNSS) provides vital information for the probe-car systems aiming at solving road-traffic problems. In the case of the Japanese Electronic Toll Collection System 2.0, a huge amount of probe data can be gathered at intervals of 200 m throughout the country. However, it is not easy for conventional map-matching algorithms to perform appropriately when they target this sparse probe data. Since for the sparser probe data of this range, it is required to check the reachability of the probe car between adjacent position fixes by using the Dijkstra's algorithm or A* algorithms. These algorithms, however, consume much computation power and can be a serious obstacle for map-matching processing, especially in real-time applications. The authors propose a new dynamic-programming-based map-matching algorithm, which can also reduce the calculation time for the reachability test by introducing a hash algorithm. The results of the evaluation confirm the robustness and the effectiveness of the proposed algorithm in terms of both accuracy and computational performance.