Distributed Inference-Based Multi-Robot Exploration

This work proposes a technique for distributed multi-robot exploration that leverages novel map inference to increase the team’s cumulative exploration efficiency. The multirobot team uses a distributed algorithm to coordinate the exploration using both the inferred and observed portions of the map. Individual robots select exploration poses by accounting for expected information gain and travel costs. Robots resolve conflicts between exploration goals with local auctions of expected travel costs. The benefits of inference-informed exploration are demonstrated in both simulated explorations and hardware trials. The proposed technique is compared against frontierbased and information-based exploration approaches. These comparisons evaluate the performance of the three exploration methods with decaying communication and a varied number of agents. Including inference in the coordination leads to a 13.15% reduction in the cumulative exploration path length in the trials conducted.