A Hierarchical Searching Approach for Laser SLAM Re-localization on Mobile Robot Platform

摘要

Simultaneous localization and mapping (SLAM) is an important issue for mobile robot system. Re-localization, as a part of close loop checking of SLAM, focuses on localize the position of a robot agent when it is restarted. Traditional localization methods couldn't reach both a high accuracy and a little time cost. In this paper, we proposed a newly hierarchical searching approach for laser SLAM re-localization. We do hierarchically searching to match an input laser frame with the whole 2-D map, instead of comparing it with each frame in the map. By using such approach, we could accelerate re-localization speed, so the performance of SLAM is improved. Our SLAM system is well organized on a mobile robot platform. After field test, our approach reaches a average precision (AP) of 90.6% on a 20m×20m map area, and the time cost is averagely 30ms. The result shows our hierarchical searching approach is effective for re-localization problems on laser SLAM.