Homogeneous Hierarchical Composition of Areas in Multi-robot Area Coverage

Abstract

Multi-robot area coverage poses several research challenges. The challenge of coordinating multiple robots’ actions coupled with the challenge of minimizing the overlap in coverage across robots becomes even more complex and critical when large teams and large areas are involved. In fact, the efficiency critically hinges on the coordination algorithms used and the robot capabilities.

Multi-robot coverage of such large areas can be tackled by the divide-andconquer policy; decomposing the coverage area into several small coverage grids. It is fairly simple to devise algorithms to minimize the overlap in small grids by making simple assumptions. If the overlap ratio of these small grids can be controlled, one may be able to integrate them appropriately to cover the large grid.

In this paper, we introduce homogeneous hierarchical composition grids to decompose a coverage area into several small coverage primitives with appropriately sized robot teams. These coverage grids are viewed as cells at a Meta level and composed hierarchically with such teams functioning as a single unit. We state and prove an associated theorem that provides very good scaling properties to large grids. We have performed simulated studies to validate the claims and study performance.