Determining the path followed by a moving device is an important task in a number of fields. Map matching is the problem of obtaining the most likely trajectory of a device on the road network given a sequence of observed locations. Past work has demonstrated that it is possible to reconstruct the trajectory of a device with good accuracy even with sparse GPS positions. In this work, we show that similar results can be achieved using sparse sequences of cellular fingerprints. Compared to GPS positions, cellular fingerprints provide coarser spatial information, but they allow a significant reduction in power consumption. We propose a new map-matching algorithm, based on the well-known Hidden Markov Model construction, that successfully works with noisy and sparse cellular observations. The proposal has been tested on a urban environment of a medium-sized Italian city. Its robustness has been checked by varying the sampling of the observations and the density of the fingerprint map, and by using mixed sequences of GPS and fingerprints observations.
展开▼