Coverage Maintenance in Multi-Robot Systems Under Sensor Failures

Abstract

This study focuses on the challenges that arise following sensor failures in multi-robot systems reliant on relative position sensing for cooperative tasks. As a robot with a failed sensor cannot directly observe its neighbors, it would be unable to participate in coordinated coverage control. To address this issue, we propose a two-part method. First, a relative position estimation strategy enables the failed robot to infer the positions of its neighbors by using their shared observations and a dead-reckoning process based on communicated velocity data. Second, a motion control strategy guides the failed robot to a location where the overall sensor coverage area can be maximally recovered. The proposed strategies are designed to operate in a distributed manner by using only local communication. The effectiveness of the two-part method is demonstrated through experiments with physical robots and numerical simulations. The results reveal that sensor coverage can be maintained or restored even after sensor failures. These findings support the feasibility of robust multi-robot coordination in scenarios involving sensor failures.