Improved GPS Positioning for Motor Vehicles Through Map Matching

Absolute positioning systems, especially the Global P o sitioning System (GPS), are playing an increasing role in the tracking of motor vehicles. Such systems, however were net designed to specifically track motor vehicles. Apart from the start and finish of journeys, motor vehicles are restricted to the mad network; a fact not reflected by the position measurements. Techniques exist for carrecting the measurements according to a map of the network but a mathematical framework for this correction process is lacking. The author has developed such a framework resulting in a map-aided estimation process that takes into account the measurement noise statistics to optimally translate the raw position measurements onto the road network. The theoratical performance of this estimator is derived and compared with that achieved using GPS measurements. The effect that Selective Availability (SA) has on the map-aided estimator is analysed and a technique for reducing the effect of SA is demonstrated In addition to the accuracy improvements, the map-aided positioning framework also provides a means for readily incorporationg other sources of positioning information. Driver bchaviour, mad type, and vehicle dynamics can be used to make further improvements in positioning accuracy and road identitication.’