Fuzzy neuroconformal analysis of multilayer elliptical cylindrical and asymmetrical coplanar striplines

Abstract

In this paper, accurate and compact analytic closed-form expressions are presented in order to calculate the quasi-static parameters of symmetric/asymmetric multilayer planar coplanar stripline (MACPS), multilayer cylindrical (MCCPS), and multilayer elliptical (MECPS) by using the conformal mapping technique (CMT). The general form of the developed expressions can take into account an arbitrarily number of dielectric layers above and/or below the strips’ interface. Models based on artificial neural networks (ANNs) and fuzzy systems (FSs) are implemented where the CMT being the reference in training process. The ANNs and FSs are trained using the back-propagation algorithm together with Levenberg–Marquardt (LM) and Takagi–Sugeno–Kang (TSK) methods, respectively. By writing efficient Matlab® R13 coding for implementation of the ANNs and FSs and performing adequate sampling of the input variables, the size of the input training matrix reaches 10,000 by 14 which ensures high accuracy of the two models. The results of the ANNs and FSs trained with their respective algorithms for the quasi-static parameters of the MACPS, MCCPS, and MECPS are in very good agreement with the results available in the literature.