A Fusion Map Matching Method for Low-Cost Ground Autonomic Positioning System

Abstract

Map matching technology is an effective way to improve the accuracy of ground inertial positioning systems by similarity comparison of vehicle trajectory and road shape characteristics. Traditional point matching algorithm can correct the inertial positioning errors in time through continuous projection operation, but it is prone to mismatch in urban complex road networks with similar shape features. Line matching algorithm can reduce the probability of false matching, but it demands to waiting obvious road bending characteristics, leading to the problem of untimely error correction. Therefore, an innovative fusion map matching method is proposed in this paper, aiming to make an organic combination of the advantages of point and line map matching algorithms, especially for accuracy improvement of low-cost ground autonomic positioning systems. Moreover, the paper discusses the influence of inertial positioning error characteristics on map matching accuracy, pointing out the original positioning errors with gentle fluctuation is more conducive to further improve map matching performance. The correctness and feasibility of proposed fusion matching method are demonstrated by a vehicle experiment, proving that positioning accuracy could be improved by 22.5% compared with traditional line map matching method.