A Robust Fuzzy-Logic Technique for Computer-Aided Diagnosis of Microwave Filters

Abstract

This paper introduces an improved algorithm based on fuzzy logic for tuning microwave filters. The approach is demonstrated by considering slightly detuned and highly detuned eight-pole elliptic function filters with tuned resonators and four-pole Chebyshev filter with mistuned resonators. Employing Sugeno-type fuzzy-logic system (FLS) along with fuzzy subtractive clustering results in much fewer fuzzy rules. The parameters of the fuzzy system are methodically adjusted to provide an optimized system. Unlike previous published method, only one FLS is adequate to deal with both cases of slightly detuned and highly detuned filters. The achieved results demonstrate the validity of the proposed approach in identifying the filter elements that cause the detuning.