Communication constraints multi-agent territory exploration task

摘要

A common assumption made in multi-robot research is the connectedness of the underlying network. Although this seems a valid assumption for static networks, it is not realistic for mobile robotic networks, where communication between robots usually is distance dependent. Motivated by this fact, we explicitly consider the communication limitations. This paper extends the LFIP based exploration framework previously developed by Pal et al. (Cogn. Comput. doi:10.1007/s12559-012-9142-7, 2012), to address the Multi-Agent Territory Exploration (MATE-n k ) task under severe communication constraints. In MATE-n k task agents have to explore their environment to find and visit n checkpoints, which only count as “visited” when k agents are present at the same time. In its simplest form, the architecture consists of two layers: an “Exploration layer” consisting of a selection of future locations for the team for further exploring the environment, and “Exploration and CheckpointVisit layer”, consisting of visiting the detected checkpoints while continuing the exploration task. The connectivity maintenance objective is achieved via two ways: (1) The first layer employs a leader-follower concept, where a communication zone is constructed by the leader using a distance transforms method, and (2) In the second layer we make use of a graph theory for characterizing the communication, which employs the adjacency and Laplacian matrices of the graph and their spectral properties. The proposed approach has been implemented and evaluated in several simulated environments and with varying team sizes and communication ranges. Throughout the paper, our conclusions are corroborated by the results from extensive simulations.