Accurate vehicle positioning in urban areas

We present in this paper a method for land-vehicle navigation systems which fuses GPS and dead reckoning data especially in urban areas where GPS outages can happen. The DR sensor here computes the differential odometry of the vehicle. In order to improve the positioning accuracy, our algorithm also fuses the road map database of a GIS. The digital road map here is modeled as a measure described by a set of node locations and road bearings. This measurement equation allows a direct modeling of the map errors and uncertainties that can be embedded in the filter, and it is initialized by a matching procedure that uses the Mahalanobis metric as a statistical distance criterion. Based on a centralized fusion scheme, the filter sequentially processes the available measurements. We present experimental results from an urban transport network scenario in presence of GPS outages, which shows the benefits of the proposed method for the vehicle positioning.