Feedforward Tuning by Fitting Iterative Learning Control Signal for Precision Motion Systems
Luyao Dai, Xin Li, Yu Zhu, Ming Zhang
Abstract
In this article, a feedforward tuning approach by fitting iterative learning control (ILC) signal for precision motion systems is proposed. The idea is that, since ILC can achieve excellent tracking performance, ILC signal contains useful information about the plant dynamics and thus can be utilized for effective tuning of model-based feedforward controller. By this method, the high performance of ILC can be attained while the sensitivity of ILC to reference trajectory variation can be avoided. In the proposed algorithm, acceleration, jerk, and snap feedforward are tuned by fitting ILC signal through least squares first; then, the residual tracking error is compensated by an extra model-based feedforward controller, the structure of the controller is determined by Monte Carlo search and the parameters of the controller are also tuned by fitting ILC signal through least squares. Simulation and experiment on a precision motion system both validate the theoretical analysis and the proposed algorithm.