FRITによるスミス補償器の最適制御パラメータとプラントモデルの同時獲得について

抄録

In this paper, we consider the data driven approach to the parameter tuning of a controller used for time-delay systems. The first author once proposed the fictitious reference iterative tuning (which is abbreviated as the FRIT) which yields the desired parameter of a controller by using only one-shot experimental data instead of mathematical models of a plant. Thus the FRIT is an effective and useful tuning method from the practical points of view in the sense that the required time and cost can be drastically reduced. This paper extends the FRIT to the parameter tuning of the Smith compensator for time-delay systems. It is well known that the Smith compensator is an effective controller for time-delay systems when we have mathematical models of a plant with time-delay. We show that the FRIT is effectively applicable to the parameter tuning of the Smith compensator under the situation in which we do not have mathematical models of a plant. Particularly, by introducing the special structure for the feedback controller used in the Smith compensator, we show that such a structure enables us to obtain not only the desired parameter of a controller but also the mathematical model of a plant. Finally, we also give a numerical example and an experimental one in order to show the validity of the proposed methods.