Self-tuning control of continuous-time two-input two-output system

Abstract

The paper presents a comparison of two approaches applied to self - tuning control of a continuous - time two - input two - output system. In first case a continuous - time controller is used and since derivatives of an input and output of a controlled process can not be directly determined, the differential filters and filtered variables are established to substitute primary variables. The filtered variables are then used in the recursive identification procedure. The second method approximates the continuous - time system by a discrete linear model of the same order and a discrete controller is used. The classical recursive least squares method with the directional forgetting is then used to identify the system. Both discrete and continuous - time controllers which were applied to control the system are based on polynomial methods.