Analysis of Opportunistic Interactions for Multi-robot Exploration Under Communication Constraints

Multi-robot exploration (MRE) under communication constraints with high-level actions (i.e., deliver data to a base station) is a topic of research that recently received attention from the academic community. That stems from the fact that MRE can benefit interplanetary exploration, military, and search and rescue applications. Several researchers propose different methods that are taken into consideration to deliver data to a base station since it can help a human team to coordinate action plans efficiently and also to share information among robots. Among the most recent comparisons of different algorithms, we have not found one that addresses delivering data delivery situations. This paper extensively analyzes different frontier-based methods for MRE under communication constraints when the robot needs to deliver data to a base station. We measure the performance of the algorithms in terms of the amount of area explored per step to complete the exploration. For this, we developed a grid-world simulator that resembles a configuration space that can run hundreds of robots efficiently. With our analysis, we identified which method portrays the best performance and spent fewer computational resources, and provided a proof of concept in real robots showing qualitative results regarding its deployment.