14th INTERNATIONAL CONFERENCE ON ENGINEERING & NATURAL SCIENCES, Sivas, Türkiye, 18 - 19 Temmuz 2022, ss.628-639
Abstract
In fuzzy logic applications, a Fuzzy Inference System (FIS), which gives outputs for inputs
by using various membership functions, plays an important role. The accuracy of the outputs
of the FIS correlates with the use of inference methods as well as the rules suggested as an
expert experience. In this direction, it is an important contribution to evaluate the FIS from
different perspectives to get better results. In a former study, Fractional Fuzzy Inference
System (FFIS) was suggested as an approach that claims to make a more meaningful
evaluation using the traditional fuzzy inference system. Unlike the traditional FIS, fractional
membership functions which consist of constant or dynamic fractional indices, are used in
FFIS. In addition to the accuracy of the membership function, FFIS operates the evaluation
process by considering the volume of the information depending on the accuracy degree. For
this purpose, a horizontal membership degree is used together with the vertical membership
degree used in the traditional fuzzy logic system. This horizontal membership function is
defined in terms of fractional indices and vertical membership functions. The volume of the
information is taken into account by using both vertical and horizontal membership functions.
Fractional indices are values that range from 0 to 1. When it takes the value of 1, the
traditional FIS is obtained. The FFIS, which uses the degree of accuracy of the information
together with the volume of information, is an expansion of the traditional FIS. In this study,
FFIS is used for position control of a DC motor. The results obtained with FFIS were
compared with the results obtained with traditional FIS. It has been shown that the
application with FFIS by selecting different fractional indices provides fine tuning. As a
result, more satisfactory results were obtained with the use of FFIS.