Sliding Mode Based Robust Learning Control For Uncertain Linear System With Mismatched Disturbance

A robust learning control strategy is presented for uncertain linear system with mismatched disturbance. A linear model with both mismatched uncertainty and disturbance is proposed firstly, in which the information of uncertainty and disturbance are unknown. Based on a reference model, a tracking error based linear model is built. What's more, the input matrix of the linear model is assumed to be suffering from uncertainty. An extended state observer is built to approximate the uncertainty and disturbance online, and then a novel state tracking error based sliding surface is proposed. The stability of the sliding surface is guarantee by the designing parameter. A sliding mode based learning controller is designed for the closed-loop uncertain system. Moreover, a numerical simulation result on hypersonic flight vehicles is proposed to test the presented method.