Data-Driven State Feedback Gain Update Based on Model Matching in Frequency Domain

This paper presents a method for updating the state feedback gain of an LTI single-input nth-order system using data-driven techniques. The goal is to match the closed-loop system to a desired reference model, which is represented by a complementary sensitivity function. To achieve this, a new model matching error is introduced. The appropriate gain can be determined using the least squares method in the frequency domain, based on closed-loop data obtained from an improperly tuned system due to model errors. Although the proposed method is specialized for updating the state feedback gain, it can perform gain updates that take into account the effects of noise at an extremely low computational cost. The method’s effectiveness is demonstrated through numerical experiments.