Data-Based Robust Tracking Control for Learning Systems Under Disturbance Observers

The high precision tracking is a fundamental objective for iterative learning control (ILC) systems, which may be challenging in the presence of the iteration-varying disturbances. This article aims to address the robust tracking control problem for ILC systems having the iteration-varying disturbances, where the accurate model information is unavailable. Based on the input and disturbed output data collected from the test iterations, the nominal model of the ILC system is first constructed, under which a disturbance observer (DOB) is established to estimate not only the iteration-varying disturbance but also the model uncertainty. Further, a DOB-based ILC updating law is developed to achieve the robust tracking objective through inserting the estimation of the iteration-varying disturbance and the model uncertainty, such that the tracking error is dependent continuously on the bound of the second-order variation rate of the disturbance. Particularly, the perfect tracking objective can be realized under the iteration-varying disturbance subject to the convergent variation rate. As a result, the tracking performance of the ILC system is improved under the iteration-varying disturbance, where the design of the DOB and the DOB-based ILC updating law depends only on the input and output data from the test iterations in the absence of the accurate model information.