Parameter Conditions and Least Squares Identification of Single-input Single-output Convex Fuzzy System

摘要

This paper presents fuzzy parameter conditions and a least squares (LS) based method to identify fuzzy systems that have convex input-output relationships. We first suggest sufficient conditions on parameters of zeroth order Takagi-Sugeno-Kang (TSK) type convex fuzzy systems and first order TSK-type convex fuzzy systems. By using the suggested conditions, we propose an LS based fuzzy system identification method whose system preserves the convexity property. Simulation results show that the proposed identification method preserves the convexity property and has less error than the conventional LS method.