Bounded adaptive output feedback tracking control for flexible-joint robot manipulators

Abstract

This paper presents a bounded adaptive output feedback tracking control approach for flexible-joint robot manipulators with parametric uncertainties and bounded torque inputs, from a systematic perspective of different (weak or strong) joint flexibilities. The singular perturbation theory and integral manifold concept are applied to decouple the dynamics of flexible-joint robot manipulators into a slow subsystem and a fast subsystem. A class of saturation functions is used to make the control law bounded, ensuring the torque control inputs are within the output limitation of the joint actuators. An adaptive control law of the projection type is adopted to handle the feed-forward term of the slow sub-controller with parametric uncertainties. Meanwhile, an approximate differential filter and a high-gain observer are utilized in the slow and fast subsystems, respectively, to estimate the unmeasurable states, making the complete closed-loop control with only position measurements of motors and links. Importantly, a corrective control scheme is proposed to break through the traditional singular perturbation approach and to make it feasible for robot manipulators with strong joint flexibility. Furthermore, an all-round and strict stability analysis of the whole control system is given. Finally, simulation results verify the superior dynamic performance of the proposed approach.