摘要：Nanopositioning stages are an example of motion systems that are required to accurately perform high frequency repetitive scanning motions. The tracking performance can be significantly increased by iteratively updating a feedforward input by using a nonparametric inverse plant model. However, in this paper it is shown that current approaches lack systematic robustness considerations and are suffering from limited design freedom to enforce satisfying convergence behavior. Therefore, inspired by the existing Iterative Learning Control approach, robustness is added to the existing methods to enable the desired convergence behavior. This results in the Robust Iterative Inversion-based Control method, whose potential for superior convergence is experimentally verified on a Nanopositioning system.
