Distributed Receding Horizon Coverage Control by Multiple Mobile Robots

Abstract

This paper presents a distributed receding horizon coverage control algorithm to control a group of mobile robots having linear dynamics with the assumption that the robot dynamics are decoupled from each other. The objective of the coverage algorithm considered here is to maximize the detection of the occurrence of the events. First the authors introduce a centralized receding horizon coverage control and then they introduce a distributed version of it. To avoid the common disadvantages that are associated with the centralized approach, the problem is then decomposed into several RHCC problems, each associated with a particular robot, that are solved using distributed techniques. In order to solve each RHCC, each robot needs to know the trajectories of its neighbors during the optimization time interval. Since this information is not available, an algorithm is presented to estimate the trajectory of the neighboring robots. To minimize the estimation error, a compatibility constraint, which is also a key requirement in the closed-loop stability analysis, is considered. Moreover, the proof of the close-loop stability of this distributed version is provided and shows that the location of the robots will indeed converge to the centroids of a Voronoi partition. Simulation results validate the algorithm and the convergence of the robots to the centroidal Voronoi configuration.