Data-Based Switching Feedforward Control for Repeating and Varying Tasks: With Application to an Ultraprecision Wafer Stage

In precision motion systems, well-designed feedforward control can effectively compensate for the reference-induced error. Compared to iterative learning control (ILC), data-based fixed-structure feedforward control (DFFC) possesses robustness against reference variations, but results in mediocre performance in exactly repeating tasks. In this paper, a novel data-based switching feedforward control (DSFC) approach is synthesized to well balance the tradeoff between the extrapolation capabilities and servo performance. Specifically, a theoretical framework is developed for the proposed DSFC approach, where the feedforward control is switched between ILC and DFFC according to whether the successive references are repeated or not. When operating in the DFFC mode, a new iterative parameter tuning algorithm is proposed to enable the performance enhancement compared to the pre-existing DFFC and overcome the limitation of the allowable reference variations. Furthermore, an unbiased estimate method for the convolution matrix of the (process) sensitivity function is developed based on the impulse response experiment. No parametric model is required throughout the proposed DSFC procedure, and the optimal parameters of the DFFC mode can be unbiasedly estimated. Experimental results on an ultraprecision wafer stage confirm that the proposed DSFC combines advantages of ILC and DFFC, and achieves high performance for both repeating and varying tasks.