Distributed Monocular Multi-Robot SLAM

In this paper, we propose a distributed multi-robot SLAM system, where each robot estimates its pose and reconstructs the environment simultaneously using the same monocular SLAM algorithm, while sharing the results of their incremental maps by streaming keyframes through the Robot Operating System (ROS)messages and the wireless network. Subsequently, the multi-robot group can obtain the global map with high efficiency and robustness. To build this multi-robot SLAM architecture, we propose a novel vision based multi-robot relative pose estimating and map merging method which uses the appearance-based place recognition method to determine multi-robot relative poses and build the large-scale global map by merging each robot's local map. Extensive experiments have been conducted and the experimental results show that the proposed distributed monocular multi-robot SLAM system can be used in outdoor large-scale environments.