Multi-hypothesis map-matching on 3D navigable maps using raw GPS measurements

For many road transport applications, maps of the environment where the vehicles evolve are available. This information can contribute to the positioning process itself. In this paper, global positioning on navigable maps is formalized in a general Bayesian framework. Using a tight integration of map data, a generic solution to multi-hypothesis map-matching is described. This method is then applied to the use of raw GPS measurements: pseudoranges and Dopplers. State space equations are given and a marginalized particle filter is proposed to solve efficiently the problem. Experimental results are presented and show that this approach can provide good results even if few satellites are visible.