Development of a map matching method using the multiple hypothesis technique

Abstract

This paper proposes a map matching method using the multiple hypothesis technique (MHT) to determine a road in probabilistic approach. The MHT is a method to track multiple targets under the clutter environment using a likelihood function. To realize a map matching method using the MHT, pseudo-measurements are generated utilizing adjacent roads of GPS position and the MHT is reformulated as a single target problem. Since pseudo-measurements are generated using digital maps, topological properties such as road connection, direction, and road facility information can be considered by calculating the probabilities of hypotheses. In order to reduce the degradation of the map matching performance by bias errors in the road data in digital maps, a Kalman filter is employed to estimate the bias. Field experimental results show that the proposed map matching method provides a consistent performance even in complex downtown areas, overpass/underpass areas, and in the areas that roads are adjacent in parallel.