A Novel SMMS Teleoperation Control Framework for Multiple Mobile Agents With Obstacles Avoidance by Leader Selection

• Published in: IEEE transactions on systems, man, and cybernetics. Systems man, and cybernetics. Systems, 2023-03, Vol.53 (3), p.1-13
• Main Authors: Huang, Fanghao, Chen, Xuanlin, Chen, Zheng, Pan, Ya-Jun
• Format: Article
• Language: English
• Subjects: Agents (artificial intelligence); Algorithms; Controllers; Feedback; Force feedback; Formation control; Leader selection; multiagent; Multiagent systems; Neural networks; Nonlinearity; Obstacle avoidance; obstacles avoidance; Radial basis function; Robot sensing systems; Robots; Sliding mode control; Stability; Subsystems; Task analysis; teleoperation control; Trajectory; Uncertainty; virtual force feedback
• ISSN: 2168-2216; 2168-2232
• DOI: 10.1109/TSMC.2022.3199112

Teleoperation of multiple agents has the unique advantage to complete tasks with wide range and is an effective solution to help agents avoid obstacles with human intelligence, especially when encounters the local-minima problem. In this article, a novel nonlinear single-master-multislave (SMMS) teleoperation control framework is proposed for multiple mobile agents to achieve obstacles avoidance under delays, nonlinearities, various uncertainties, and nonholonomic constraints. Namely, the slave trajectory planner is designed to cope with the nonholonomic constraints caused by underactuated characteristics of slave agents, while the slave obstacle avoidance planner is designed to cope with obstacles in the environment, which can avoid the obstacles by artificial potential function (APF)-based obstacle avoidance algorithm. Particularly, considering that the APF usually encounters the local-minima problem, a leader selection algorithm is designed for the slave obstacle avoidance planner and a virtual force feedback is designed for the master subsystem, where the slave agents can get rid of local-minima points while teleoperated by human operator with confident force feedback. The global stability of the overall system can be guaranteed under the proposed radial basis function neural network (RBFNN)-based adaptive sliding mode master controller and slave formation controller under delays, nonlinearities and various uncertainties. The comparative experiment is implemented, and the results show the effectiveness of proposed control framework in the achievement of good performance including position tracking, force feedback, and formation and obstacles avoidance while the stability is guaranteed.