On the State-Space Representation of Local Affine Models in Takagi–Sugeno–Kang Fuzzy Systems (Zur Zustandsdarstellung von Takagi–Sugeno–Kang-Fuzzy-Systemen mit Lokal Affinen Modellen)

Abstract

Abstract Considering the constant terms in the local models of dynamic Takagi–Sugeno–Kang fuzzy models can provide for improved prediction quality. However, the constant terms have typically not been treated analytically in analysis and controller design. In this paper, minimal state-space representations are derived from local affine input/output models for usage in TSK fuzzy models. Transformation algorithms for structure, parameters and initial values are proposed.