Iterative learning control for systems with input deadzone

Abstract

In this work, we apply Iterative Learning Control (ILC) approach to address control problems associated with a nonlinear, unknown and state-dependent deadzone. Input deadzone is a kind of non-smooth and non-affine-in-input factor. It gives rise to difficulty in control due to its presence in the system input channel as well as the singularity. Since many control tasks in automated industrial processes are repeated or run-to-run in nature, we can apply ILC methods to deal with the input deadzone. Unlike many existing deadzone compensation schemes which are highly complex and hard to implement, in this work the ultimate objective is to apply the simplest, easy-to-go ILC method and meanwhile achieve a satisfactory deadzone compensation.