Extended self-tuning algorithm

Abstract

This article concerns the practical implications of an extended self-tuning property. The extension represents a refinement of existing asymptotic self-tuning properties, in the sense that it allows the designer to make arbitrary assumptions concerning the disturbance dynamics. The important practical consequence of this is that a simple pole-assignment self-tuner may now be designed which combines servo-tracking and regulatory closed-loop criteria. In the pure regulation case a priori parametrization of the noise dynamics can be used to aid convergence of the self-tuning algorithm. In addition, an implicit pole-assignment regulator is easily formulated from the extended self-tuner.