Middle-Frequency Limited Model Reduction Techniques for T-S fuzzy discrete-time systems

This paper deals with the problem of Middle-Frequency (MF) approach H
_∞
model reduction for nonlinear discrete-time systems in the Takagi-Sugeno (T-S) fuzzy models. The objective is to propose a new design sufficient condition via linear matrix Inequality (LMIs) formulations, ensuring that the error system is stable and has a minimized H
_∞
performance when frequency ranges of noises are known beforehand. Less conservative results are obtained by using the new finite frequency H
_∞
performance index is firstly defined, Finsler's lemma and some independent matrices. Simulation example demonstrate the technique and its advantage.