Adapting to sensing and actuation variations in multi-robot coverage

Abstract

This article considers the problem of multi-robot coverage control, where a group of robots has to spread out over an environment to provide coverage. We propose a new approach for a group of robots carrying out this collaborative task that will adapt online to performance variations among the robots. Two types of performance variations are considered: variations in sensing performance (e.g. differences in sensor types, calibration, or noise), and variations in actuation (e.g. differences in terrain, vehicle types, or lossy motors). The robots have no prior knowledge of the relative strengths of their performance compared to the others in the team. We present an algorithm that learns the relative performance variations among the robots online, in a distributed fashion, and automatically compensates by assigning the weak robots a smaller portion of the environment and the strong robots a larger portion. Using a Lyapunov-type proof, we show that the robots converge to a locally optimal coverage configuration. The algorithm is also demonstrated in both MATLAB simulations and experiments with Pololu m3pi ground robots.