Data-driven Controller Tuning for Attenuation of the Effect of Unknown Periodic Disturbance

Abstract

In this paper, we provide a novel method of controller parameter tuning for attenuating the effect of unknown exogenous periodic disturbances. Our approach is based on the direct use of one-shot experimental data obtained in the actual closed loop. By utilizing such a data, we tune the off-line disturbance observer with a nominal model of a plant to achieve the desired closed loop property. We also illustrate a numerical example to show the validity of our proposed method.