Flocking-based distributed terrain coverage with dynamically-formed teams of mobile mini-robots

We consider the problem of distributed terrain coverage within an unknown environment by teams of mobile, resource-constrained mini-robots. Previous research on this topic mainly focuses on multiple robots that cover the environment individually while sharing limited information about their actions using local heuristics. However, these methods do not address the issue of dynamically building robot teams and restructuring the team formations based on the operational constraints within the environment. In this paper, we combine a flocking-based mechanism for robot team formation with a utility-based mechanism that enables robots to change teams to dynamically to improve the coverage of the environment. We have tested our techniques empirically on accurate models of e-puck robots on the Webots simulator within different environments. Our experimental results show that our team-based coverage technique performs significantly better in terms of coverage, by about 25%, and slightly better in terms of redundancy in area covered, than comparable distributed terrain coverage strategies for resource-constrained mini-robots.