Disturbance Estimation and Attenuation for Repetitive-Control Systems with Unknown Nonlinearity

摘要

This paper presents a framework for a class of single-input-single-output (SISO) nonlinear systems with periodic tasks and subject to aperiodic disturbances using linear-extended-state-observer (LESO)-based repetitive control method. A two-stage design procedure is devised to enhance the disturbance estimation and attenuation ability of repetitive-control system where design of LESO is separated from design of repetitive control law. An LESO is designed to estimate and compensate the nonvanishing lumped disturbance in real time fashion. The stability criterion of the LESO-based repetitive-control system is established in the presence of unknown nonlinear dynamics. The simulations demonstrate that the proposed LESO-based repetitive-control system can achieve satisfactory disturbance rejection and tracking control performances simultaneously. Furthermore, a slight correction to the amount of the delay of the repetitive controller leads to a substantial decrease in steady-state tracking error.