Interval type-2 fuzzy logic systems optimized by central composite design to create a simplified fuzzy rule base in image processing for quality control application

Abstract

A novel method that uses a Mandami interval singleton type-2 fuzzy logic system (IT2 SFLS) with the support of the central composite design (CCD) technique and the classic approach of composed base inference (CBI) is made to enhance the modeling and the construction of the fuzzy rule base. The IT2 SFLS has the potential to outperform the singleton type-1 fuzzy logic systems (T1 SFLS). The IT2 SFLS systems accounts for the uncertainties that can be added during the system modeling and construction: the uncertain rules created using noisy data. There is no way to take into account this uncertainty in the antecedent and consequent membership functions of a singleton type-1 fuzzy logic systems. Due to this uncertainty, an additional process is required to filter the measured data, but the uncertainty is still present in the structure of the T1 system. The main goal of the proposed model is to enhance the performance obtained in the dimensional features evaluation in a quality assurance process of the manufacturing of product parts. The experiments developed in a real facility show that the application of the proposed method produced better results than that obtained by the T1 SFLS benchmarking system.