Dynamic Output-Only Iterative Learning Control Design

Iterative learning control applies to systems that repeatedly execute the same finite duration task. The distinguishing feature of this form of control action is that all data generated on a previous execution of the task is available to compute the control action for the subsequent execution. This paper uses the repetitive process stability analysis and optimization techniques to design a dynamic controller that, in contrast to previous designs in the repetitive process/2D systems setting, does not require measurement of the state dynamics or observer-based estimation. Examples to demonstrate the application of the new design are given.