2015 年 51 巻 9 号 p. 597-604
This paper provides a closed-loop identification method based on generalized minimum variance (GMV) evaluation. The method assumes that the system model is represented as the CARMA (Controlled Auto-Regressive and Moving Average) model, and identifies parameters of a disturbance model as well as plant parameters. Closed-loop input and output measurements on the nominal GMV controll system are used for the identification. The present method has a distinctive point in terms of using the data-driven control specification for minimizing the variance of the generalized output. The paper proves that minimization of the proposed cost function results in identifying plant and disturbance model parameters uniquely. Finally, the effectiveness of the proposed method is demonstrated through numerical examples for both stable and unstable processes.