Image Processing-Based Center Calculation Method for General and Interval Type-2 Fuzzy Systems

Abstract

Despite being a good method against uncertainties, the high complexity of the operations in type-2 fuzzy system makes it slow. So, it should be reduced to use type-2, especially in real-time problems. In this study, a method based on image digitization has been proposed for type-reduction and defuzzification steps, which are forming the actual computational complexity in type-2 fuzzy systems. Instead of solutions using iterative and complex methods, the problem is converted into image processing problem by digitizing the fired outputs. The deduction methods such as the center of gravity are computed easily on digitized images. Example scenarios show that the work done is superior to KM algorithm in terms of both precision and computational complexity. In addition, different simulations were conducted to demonstrate the superiority of the study over real problems, and the study was tested for a complex human model motion control problem.