New Research


“Control”as the design of a dynamics

Control technology for the productivity growth and data-driven control theory

Professor Toshiyuki OHTSUKA,Associate Professor Atsumi OHARA,Assistant Professor Osamu KANEKO

 “Control”is one of the fundamental and important research areas in the sense that control technology enhances a greater comfort in daily-life like a car with ride quality, an easy-to-use home electric appliances, and so on, or in the sense that it also makes it possible that many industries manufacture higher-quality products more effectively. Thus, “control”is a meaningful and fundamental science for the design of a dynamics in order to make life and society more comfortable and convenient. Our research group addresses a study and a development of “control”from the theoretical and practical points of view. Here, we introduce one of our research activities on control technologies which contribute the productivity enhancement.
 Generally, information on the dynamics of a plant is required so as to design a controller with a desired property. This corresponds to that the first step for the design of a controller is to construct an appropriate mathematical model which represents the dynamics of a system by using the actual input and output data. At the same time, since this procedure takes a lot of time and costs, there are many cases in which it is also difficult to obain a mathematical model due to a cost-performance or an operational reason.
 On another front, the actual data of a system include fruitful information of the dynamics directly. Thus, it is expected that a new perspective in which the direct use of the data without mathemataical model yields a desired controller reflecting the dynamics. From this perspective, we develop and study “Data-driven control”that provides a new approach for the design of a controller based on the direct use of the data. Particulary, we also develop a new controller parameter tuning method called “FRIT”with only one-shot experimental data for the achievement of a given specfication, which contributes high cost performance in the sense that it is possible to drastically reduce the cost and time required for an appropriate controller with a desired property. Moreover, data-driven control theory is also theoretically new because the “language”for the design of a controller in this approach is the actual data. From these practical contributions and theoretical interests, we are now studying data-driven control.

Ohtsuka Laboratory web page
◎About this site
Go back to page top