DCP: Distributed Collaborative Planner for Multi-robot Autonomous Exploration in Large-scale Unknown Environments

Exploring large-scale unknown environments with a multi-robot system is a challenging task. Existing methods of collaborative exploration mostly rely on centralized systems. This paper proposes a distributed collaborative planner for multi-robot autonomous exploration. We formulate the collaborative exploration as a min-max optimization problem. The planner aims to decrease the total exploration distance by minimizing the maximum distance of all robots, and to reduce the variance among all the distances to further decrease the overall exploration time. Simulation experiments in different environments and with various numbers of robots demonstrate that the proposed method not only ensures the exploration completeness but also enhances the exploration efficiency.