A Method for Distributed Multi-Robot Simultaneous Localization and Mapping

Abstract

A novel method was presented to realize the distributed multi-robot simultaneous localization and mapping for the efficient exploration of unknown environments and partial map merging. According to the robot's explored information, a single robot built its partial map and estimated its location in the next step, it could actively communicate with other robots. The communication robots formed a dynamic exploration cluster. A method of correlation matching of two right angles with less calculation was given to address the partial map merging of the robots in a cluster. The robots in a cluster share the merged map to coordinate their exploration actions and estimate their relative locations. The method enabled teams of robots to efficiently build highly accurate maps of unknown environments, even when the initial locations of the robots were unknown. Simulation results showed that the map generated by our approach was same with the standard map, and verify the high accuracy and robustness of this method.