Data‐Driven Tuning for Two‐Degree‐of Freedom Control Using Initial Control Signals and Initial Controller Parameters

Abstract

A new data‐driven approach based on initial controller parameters and initial experimental data was proposed to tune a controller such that the response of the closed‐loop and open‐loop systems are close to the desired control response. This method is called virtual internal model tuning (VIMT). In VIMT, one‐shot closed‐loop output data are required. Practically, there are cases where it is important to focus on the input rather than on the output. This study extends VIMT to the case where only one‐shot input data and initial controller parameters are obtained through closed‐loop experiments. Furthermore, we discuss the advantage of the proposed method compared to the conventional VIMT via frequency response analysis of weight function. In this study, I investigated the control performance of the proposed methods through experimental verification. © 2022 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.