抄録
The author's previous study proposed an efficient design method of time-varying gain type controller and its application to the positioning control of vibration systems. The responses of this method become similar to the responses of conventional nonstationary optimal regulators (NORs). This method calculates a time-varying Riccati equation and the responses of a controlled object simultaneously. Moreover, this study utilizes the random search technique for the tuning of the time-varying weights in the Riccati equation and realizes the efficient controller design under many control constraints. Following the previous study, this paper improves the previous method by taking the influences of parameter variations into account. This paper demonstrates an algorithm for the case which includes mass and friction variations. The effectiveness of the algorithm is verified by numerical calculations. Moreover, this paper also discusses some experimental investigation. Different from the use of NOR, the time-varying gains in this experiment are obtained by real-time computations.
