Model reference adaptive sliding control for linear systems

Abstract

This paper proposes a controller design for continuous linear systems with unknown parameters. A model reference adaptive sliding controller is introduced, which is differentiator free and requires the measurement of system’s input and output (I/O) signals only. The controller is based on the combination of an output feedback sliding control and an adaptation scheme for estimating the upper bound of the magnitude of the unknown plant parameters. Exponential stability is achieved for plants with arbitrary relative degree and is shown using Lyapunov’s stability analysis. The proposed controller avoids chattering and does not suffer from parameter drift due to adaptation. The proposed controller is easy to implement and has superior transient behavior as compared to the standard adaptive controller.