An Internal Model Based Iterative Learning Control for Wafer Scanner Systems

A new iterative learning control (ILC) is proposed, which embeds the reference model rather than the plant model to achieve a good transient performance and a fast convergence. First, the uncertain discrete-time linear system and the desired reference with an known internal model (IM) are introduced. Then, based on the IM principle, a new ILC scheme, named as IM-based ILC, is proposed by incorporating the reference IM. The IM-based ILC is independent of the plant model. Moreover, the IM-based ILC is carefully constructed as a state-feedback controller of a two-dimensional (2-D) ILC system, which is the coupling combination of an IM controller (IMC) and an ILC. The proposed coupling architecture combines the ability of an IMC (tracking the reference along the time axis) and that of an ILC (tracking the reference along the iteration axis), ensuring the synchronous convergence along the time and iteration axes. This is not trivial but the key to achieve the good transient performance and the fast convergence. In addition, two criteria are established to design a 2-D H∞ IM-based ILC and monotonically convergent one. Finally, the IM-based ILC is applied to the precision control of a wafer stage, where the experimental setup, the mathematical model, and the desired reference are discussed. And, the efficiency of the IM-based ILC is illustrated in the experiments.