Experimentally Verified Robustness Properties of a Class of Model Inverse ILC Agorithms
In this paper, the subject is the robustness and noise rejection
properties of an inverse type iterative learning control algorithm.
As a new result it is shown that by adapting the learning gain as a
function of trial number it is possible to produce a more accurate
limiting error even when the plant is subject to measurement noise.
This result is experimentally verified on an industrial-scale gantry