Adaptive learning control of flexible robot manipulators with joint flexibility

This paper presents a learning controller for flexible robot manipulators which is based on the computed torque error compensation method. The learning controller learns the desired torque input for tracking of link and elastic force dynamics. The uncertain system parameters and elastic force bound are also learned by the learning rules derived in the learning system. The link position error and the coupled elastic force error are shown to converge asymptotically, which, in turn, is confirmed via computer simulations.