Intelligent Collision-Free Formation Control of Ball-Riding Robots Using Output Recurrent Broad Learning in Industrial Cyber-Physical Systems

Abstract

This article presents an intelligent collision-free formation control method of ball-riding robots using an output recurrent broad learning strategy (ORBLS) in industrial cyber-physical systems (ICPS). A cyber ORBLS is incorporated with the backstepping sliding mode formation control (BSMFC) and potential field theory, called ICPS ORBLS-BSMFC, in order to attain collision-free formation control for the multiple ball-riding robots with uncertainties for ICPS, the proposed cyber ORBLS-BSMFC computing method is employed to address the robust self-balancing formation control problem of ICPS gyro-stabilized robots. A bi-directed and connected graph is used to mathematically model the inverse-atlas self-balancing robots with uncertainties in formation encountering unknown frictions, mass variations. Taking the feedback signals from the physical world, Lyapunov stability theory is utilized to prove that the cyber ORBLS-BSMFC control law makes the system asymptotically stable. Three simulations and two experimental results will manifest the effectiveness, superiority and merits of the proposed ICPS ORBLS-BSMFC with obstacle avoidance. Through comparative studies, the advantages of the proposed ICPS ORBLS-BSMFC computing are validated to accomplish collision-free formation control for ball-riding robots.