MIMO controller synthesis for LTI and LPV systems using input-output models

This paper considers the problem of designing MIMO fixed-structure controllers based on a polynomial framework for both Linear Time-Invariant (LTI) and Linear Parameter- Varying (LPV) systems having input-output models. For SISO systems, a polynomial method has been proposed recently to solve such problems, where the a priori selection of a central polynomial is carried out heuristically, and for LPV systems the synthesis problem is solved using a sum-of-squares relaxation approach. To reduce the complexity of the design procedure, we propose a combined gradient and LMI-based algorithm for optimizing the choice of the central polynomial matrix in terms of the closed-loop performance. Moreover, for LPV systems, the computational complexity of the sum-of-squares approach is reduced by introducing polytopic input-output LPV representations. Quadratic stability and performance is guaranteed in the admissible scheduling range. The proposed methods are illustrated with simulation examples. © 2011 IFAC.

DDC Class

600: Technik

620: Ingenieurwissenschaften

Options

MIMO controller synthesis for LTI and LPV systems using input-output models