A Method of Map Matching for Personal Positioning System

ABSTRACT: To track human’s movements with high accuracy in certain areas, where GPS system is not available, complimentary positioning methods are needed. Personal positioning system using a sensor of acceleration and a gyrocompass together with GPS is now being developed as one of promising methods. Accumulation of positioning errors, however, cannot be avoided. One of the solutions to reduce or modify these errors is “Map matching” technique, which modifies estimated user’s position assuming that users are always along road networks. In this paper, a method of Map Matching for Personal Positioning Systems is proposed. While it has already been in practical use for car navigation systems, its simple algorithm based on distance and direction estimation is not sufficient for tracking of man’s complicated movement. In addition, matching should be done for three-dimension maps because human movement consists of both vertical and horizontal motion. The algorithm proposed in this paper combines two of the current algorithms, one of which is based on matching one’s shift patterns with shapes of road network, and the other is for a frequent step-by-step adjustment of one’s estimated position against outlines of objects around it. Using this integrating algorithm, we propose a new Map Matching method that would give good performance in terms of accuracy of positioning system for human movement in 3D space.