TORQUE REDISTRIBUTION AND TIME REGULATION METHODS FOR ACTUATOR SATURATION AVOIDANCE OF FAULT-TOLERANT PARALLEL ROBOTS

In this article, we investigate and employ torque redistribution and t ime regulation approaches to track a preplanned trajectory while avoid ing actuator saturation for redundantly actuated, fault-tolerant paral lel robots encountering actuator failure. The torque redistribution me thod utilizes actuation redundancies (if available) after joint failur e and/or saturation to alter the torque requirements of the joints so as to avoid saturation. The time regulation method is based upon the a lteration of the time to completion of the defined joint-space traject ory. By searching for a suitable time scaling factor to change (reduce ) the inertia load requirements, the robot attempts to slow down or sp eed up as it moves along the defined trajectory to avoid saturation. T he preferred implementation strategy is to use the torque redistributi on approach when redundancy is available and actuation capacity is suf ficient, otherwise the time regulation approach is employed. For both methods, the computed torque technique, consisting of a model referenc e algorithm in the feedforward process, and a proportional plus integr al plus derivative (PID) controller, to deal with modeling errors and disturbances in the feedback process, are utilized. The results of sim ulations of a planar 2-DOF 5-bar linkage encountering both full and pa rtial actuation failure show the viability (at least in this case) of these approaches to trajectory tracking with redundantly actuated robo ts in the presence of actuator failure. (C) 1995 John Wiley & Sons, In c.