On implementation of feedback-based PD-type iterative learning control for robotic manipulators with hard input constraints

Abstract

It is well-known that a robotic manipulator with the position output has a relative degree 2. A standard feed-forward iterative learning control (ILC) requires the second derivative of the tracking error to achieve a perfect tracking performance. This leads to implementation issues when the output signals are noisy. To address such an issue, this paper presents a feedback-based PD-type ILC with the consideration of the actuator saturation, in which only the first derivative of the tracking error is needed. With the help of a composite energy function, under mild assumptions, Theorem 1 shows that the proposed feedback-based PD-type ILC can ensure perfect tracking performance in the presence of actuator saturation. The simulation and experimental results show the effectiveness of the proposed method.