Occupancy Grid Map Merging Using Feature Maps

摘要

Multi-robot exploration and mapping studies have demonstrated that it is often more efficient to explore unknown areas in parallel rather than with a single agent. However, the question of how to integrate individual maps into a consistent global map remains an open research area. This problem, known as map merging, comprises the establishment of a frame of reference for multiple mobile robots, the identification of regions of map overlap, and the integration of individual maps to produce a global result. In this work, we build a hybrid map which integrates occupancy grid and feature data to solve this problem. This integrated representation permits fast and effective map merging. Experimental results are presented that demonstrate algorithm performance in a realistic scenario.