Decentralized control of reconfigurable robots using joint-torque sensing

In this paper, a decentralized controller for trajectory tracking of modular and reconfigurable robot manipulators is developed. The proposed control scheme uses joint-torque sensory feedback; also sliding mode control is employed to make both position and velocity tracking errors of robot manipulators globally converging to zero. Proposed scheme also guarantees that all signals in closed-loop systems will be bounded. In contrast to some of prior works in this scheme, each controller uses a smooth law to achieve its purposes. In this method, each controller uses only local information for producing control law hence separated controller can be used to control each module of manipulator and no information exchange between modules is required. Simulation results are provided for a reconfigurable robot with three modules to verify the performance of the proposed scheme. Results show that decentralized control of reconfigurable robots is feasible, in spite of strong dynamic coupling between modules.