Centralized tensor product model form

The main objective of the paper is to propose the centralization of the tensor product model form and HOSVD based canonical form of linear parameter-varying models. We show the advantages of the proposed centralization in the domain of control theory with mathematical analysis as well as with numerical examples.

I. Introduction

Modern control theory mainly focuses on analysis and control design based on Linear Matrix Inequalities (LMI) in Linear Parameter Varying (LPV) system representations. Currently, it seems to be the most widespread approach to achieve robust and efficient nonlinear control results. These approaches usually prepare the system model for LMI design based on the analytical derivations. This requires a series of idiosyncratic and rather sophisticated analytical methods. Although we have the analytical solutions, the final result of the LMI design and the implementation of the controller based on numerical computations is approximative from the perspective of the final realization.