An extended probabilistic self-localization algorithm using hybrid maps

Map-matching algorithms integrated with GPS/DR are widely used for high precise self-localization in everyday tasks. However, GPS signal is not available in many places, e.g. tunnels or jammed signals. A state-of-the-art solution is a GPS-free map-matching algorithm with a probabilistic model as well as an efficient approximate inference algorithm. Although that algorithm is computationally efficient, it still puts high demands on the computation and driving tracks. This paper adopts the model and presents an extended probabilistic self-localization algorithm using hybrid maps which include terrain maps and road network maps. Experiments show that the extended probabilistic algorithm enhances the real-time performance and relaxes the requirement of the shape of the routes. The proposed vehicle self-localization algorithm meets positioning requirements of ITS and it can provide references for actual use of map-matching algorithm in ITS.