Design and analysis of type-2 fuzzy logic systems

Abstract

Type-2 fuzzy logic systems (T2FLSs) are extensions of type-1 FLSs proposed to better handle uncertainties. This thesis aims at providing insights into the fundamental properties of T2FLSs and improving their performance. First, it shows T2FLSs can achieve a better compromise between accuracy/performance and interpretability. Then a simplified T2FLS structure is proposed to reduce the computational cost. Next, the original concept of Equivalent Type-1 Sets (ET1Ss) is introduced and used to show that a PI-like T2FLS may be equivalent to a PI T1FLS with adaptive PI gains in certain input ranges. This provides insights into why T2FLSs may be more robust.