Fuzzy modeling based on L2 gain criterion

This paper presents a robust fuzzy modeling based on L₂ gain criterion. The most important thing is that fuzzy modeling executes using LMI conditions. We derive an LMI condition to identify the parameters of a Takagi-Sugeno fuzzy model (T-S fuzzy model). The LMI guarantees to minimize the summation of the upper bound of the identification error (SUE) between outputs of a real plant and those of a T-S fuzzy model. More importantly, we derive L₂ gain based fuzzy modeling conditions. It achieves robust parameter identification for the data contaminated by noise. An example shows the utility of the proposed iterative LMI approach to L₂ gain based fuzzy modeling