An auxiliary system discretization approach to Takagi-Sugeno fuzzy models.

Abstract

This paper proposes a new procedure for discretizing nonlinear systems described by Takagi-Sugeno fuzzy models. The discretization procedure consists of obtaining a linear auxiliary system that approximates the Takagi-Sugeno model over a sampling instant. By discretizing this auxiliary system, a norm bounded uncertain linear discrete-time system is found, which is capable of representing the fuzzy model. This auxiliary system, as well as the norm bounded uncertainty, is found by solving an optimization problem with Linear Matrix Inequality (LMI) constraints. To illustrate the discretization procedure, a constant state observer is synthesized based on simple LMI conditions and then applied to a real nonlinear Chua’s circuit. Additionally, a state-feedback controller based on our discretization approach is synthesized and we obtain larger maximum sampling periods than other tested strategies from the literature.