Consensus-Based Protocol for Distributed Exploration and Mapping

Abstract

Distributed exploration in multi-agent systems requires agents to retain a consistent view of the environment, even under limited communication ranges. We propose a consensus-based protocol for distributed exploration that enables agents to synchronize their local maps so that their collective global views are consistent. With the proposed protocol, agents can dynamically form communication networks and synchronize their local environment maps. In contrast to the existing consensus-based solutions, our work is computationally efficient and provides a convergence guarantee under communication loss. Through our extensive experiments, we show that the proposed protocol enables agents to build consistent environment maps collaboratively and efficiently. We also show that agents can significantly save their communication bandwidth and reach optimal solutions in the presence of communication loss.