A fictitious reference iterative tuning method with simultaneous delay parameter tuning of the reference model

Fictitious reference iterative tuning (FRIT) is a method for designing the feedback controller of an unknown plant model. FRIT searches control parameters so that the plant output follows the reference model output by optimizing the performance index that consists of the fictitious reference output computed from one-shot experimental input-output data. The reference model gives design specifications in the FRIT method. However, if a mismatched reference model is selected, e.g., it requires too fast response regarding the controlled plant, FRIT may possibly derive improper control parameters that make the performance of the closed-loop system undesirable. The paper proposes a modified FRIT method with simultaneous delay parameter tuning of the reference model in order to overcome the problem due to the mismatched reference model. The proposed method can adjust the rise time of the reference model output by tuning the delay parameter. The paper also proposes pre-filter design for the one-shot experimental input-output data, which makes the controller obtained in the proposed FRIT become a good approximation to the controller optimizing the original performance index evaluating deviation of the plant output from the reference output. Finally, a numerical example is shown to demonstrate the effectiveness of the proposed method.