Iterative learning control for a linear piezoelectric motor with a nonlinear input deadzone

In this work, we apply iterative learning control (ILC) approach to address control problems associated with a nonlinear, unknown and state-dependent deadzone for a linear piezoelectric motor. 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. The control problems become more complex when the input deadzone is nonlinear, unknown and state-dependent, as often encountered in circumstances where high precision actuation is required. 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. The simplest ILC must be able to accomplish the perfect tracking task over a finite interval, in the presence of system nonlinear uncertain dynamics, and without the accessibility of the system states. Experimental results clearly demonstrate the effectiveness of the ILC compensation scheme.