Uncertainty modeling for interval Type-2 Fuzzy Logic Systems based on sensor characteristics

In the past decade Type-2 Fuzzy Logic Systems (T2 FLSs) gained increased research attention due to their potential to outperform Type-1 FLSs in applications with dynamic uncertainties. This advantage is typically attributed to the capability of T2 Fuzzy Sets (FSs) to better model the dynamic uncertainty and cope with its negative impacts. However, the accuracy, correctness and interpretation of such uncertainty modeling using the T2 FLSs have been rarely addressed or taken into account during the design of the respective fuzzy controller. The contribution of this paper is in analyzing the uncertainty modeling capabilities of the commonly used Interval T2 (IT2) FSs with uncertain parameters. In addition, a novel method for incorporating the experimentally measured input uncertainty into the design of the IT2 FLS is proposed. It is demonstrated that the novel IT2 FLS design method improves the accuracy of the input uncertainty model in the specific problem domain. As a consequence, the modeled uncertainty is then more accurately reflected in the output domain as the geometry of the typereduced centroid.