Decentralized Task Allocation in Multi-robot Exploration with Position Sharing Only

. This paper is concerned with multi-robot exploration in environments with limited communication that is supposed to be unreliable. The communication is low bandwidth, sufficient only for sharing positions of the robots, and does not allow sharing maps of the environment. Under these constraints, we propose to address the multi-robot exploration as a decentralized task allocation to coordinate robots in the exploration mission while exchanging only their positions. The feasibility of the proposed approach has been validated in various simulated scenarios within the virtual cave circuit environments of the DARPA Subterranean Challenge. The reported results indicate the proposed approach yields shorter or competitive average longest traveled paths than the decentralized MinPos method with the allowed sharing of full environment maps. The introduced method thus represents a suitable choice for decentralized task allocation in limited communication scenarios.