Distributed control of networked multirobot systems

Abstract

Distributed control of networked multirobot systems has recently attracted increased attention in the fields of both control and robotics. This research aims to investigate how to govern the behavior of the whole system by using distributed control strategies, which are synthesized by employing only local information and are implemented on individual robots. The coordination of multiple robots can improve the performance and complete tasks that are too difficult for a single robot to perform alone. Owing to this advantage, numerous potential engineering applications exist in military surveillance, rescue missions, space and ocean explorations, intelligent transportation systems, and other automated collaborative operations. However, how to design distributed control laws for individual agents and how to predict their group behavior remains poorly understood. This study addresses this challenging problem based on three perspectives. First, the study addresses the rendezvous problem of multiple nonholonomic mobile robots. Simple bearing-only distributed control laws are proposed. The convergence is first proven when the interaction topology between robots is connected. For the complete interaction case, we prove that under the proposed control law, the perimeter of the convex hull defined by the positions of robots decays all the time. Consequently, all the robots converge to a common point. Second, the study addresses the problem of steering a single or a group of nonholonomic mobile robots to enclose a target of interest. Bearing-only control schemes are developed, which deal with the encirclement of two types of targets: point target and disk target. When a single robot is used, circumnavigation schemes are proposed to achieve efficient encirclement of the target. We show that by using the proposed control schemes, the robot can circle the target from a prescribed distance without distance measurement. When multiple robots are deployed, coordination schemes are proposed to distribute the robots evenly around the target. Third, the study addresses the uniform circumnavigation problem of a team of ring-coupled nonholonomic mobile robots. A novel distributed solution is proposed based on cyclic pursuit/repelling strategies to achieve a circular motion around a target in an encircling formation. This new approach considers the minimum number of information flow links between the robots and uses local measurements only. The asymptotic collective behavior is analyzed based on the block diagonalization of circulant matrices by a Fourier transform. The main contributions of this study are as follows: First, the study has made an important theoretical contribution by revealing the possibility of achieving multirobot coordination when each robot uses control strategies that require bearing measurements only. Second, the study of the ring-coupled system shows how the collective behavior of multiple robots can be shaped when each robot uses different control weights. Third, the nonholonomic constraint of mobile robot is explicitly considered in control strategy design. The proposed control strategies are also applied to mobile robots to demonstrate their practical feasibility through extensive experiments.

Date of Award	15 Jul 2014
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Dong SUN (Supervisor)