A Simplified Data-driven Optimal Iterative Learning Control based on Iterative Extended State Observer

摘要

Inspired by Active Disturbance Rejection based Iterative Learning Control (ADR-ILC) and Data-Driven Optimal Iterative Learning Control (DDOILC), this paper proposes a simplified data-driven optimal iterative control method based on iterative extended state observer (IESO). Accurate estimation of the system uncertainties is observed by IESO during the iterative process. Though considering the uncertainties on iterative dynamic linearization method, it is not needed to deduce a new form of the original iterative pseudo partial derivative (PPD). IESO, undertaking as the tool to estimate the whole uncertainties, is added into the DDOILC control law as a separate part. The whole control law is more intuitive and concise than other IESO based DDOILC method which has modified PPD updating law and control law. At the same time the variable gain control mechanism makes the proposed method demonstrate superiority over ADR-ILC in the case of strong nonlinearity. Simulation shows it that can achieve better performance than DDOILC and the other IESO based DDOILC.