Identification of Interpolating Affine LPV Models for Mechatronic Systems with one Varying Parameter

Abstract

This paper presents a technique to identify an affine parameter-dependent model based on a set of Linear- Time-Invariant (LTI) models that are identified in local operating points of the varying system. The systems under investigation are mechatronic systems in which the dynamic behaviour is depending on a single varying parameter. By fitting specific pole and zero loci on the poles and zeros of the local LTI models, an affine state-space model can be constructed. Since the resulting affine model is well conditioned and has a low complexity, it is suitable to be used for Linear- Parameter-Varying (LPV) control. The applicability of the presented technique is shown on an industrial pick-and-place machine with position-dependent dynamics.