A map matching method for restoring movement routes with cellular signaling data

Cellular signaling data is a valuable and abundant data source to explore human mobility. Yet challenges remain to restore movement routes from signaling data due to its coarse positioning information. We propose an efficient map matching method based on road network topology. First, a customized spatial-temporal clustering algorithm ST-DBSCAN was employed to find stationary point clusters, which were later used to segment trips into sub-trips. The search space was then clipped with a fixed buffer zone along the line that connects the whole trip. Two optional strategies were provided to find the best matching routes with distance costs. Experiments on real-world data showed that both strategies achieved high map matching accuracies (88.2% and 94.3%). With Deep Mode, the method reached higher accuracy, while with longer computation time. The proposed method has the potential in solving practical problems, in the sense that it could be easily parallelized to deal with mass data.