Neural Analysis of Top Shielded Multilayered Coplanar Waveguides
Artificial neural networks (ANNs) have been promising tools for many applications. In recent years, a computer-aided design approach based on (ANNs) has been introduced to microwave modelling, simulation and optimization. In this work, the characteristic parameters of top shielded multilayered coplanar waveguides (CPWs) have been determined with the use of ANN models. These neural models were trained with Levenberg-Marquardt, resilient propagation, Bayesian regulation, quasi-Newton, and backpropagation learning algorithms. Better performance and learning speed with a simpler structure were achieved from these models. The results have shown that the estimated characteristic parameters are in very good agreement with the computed results by using conformal mapping theory. The Levenberg-Marquardt learning algorithm was found to be the best algorithm among all. As a result, ANN models presented in this work can be used easily, simply and accurately to determine the characteristic parameters of the top shielded multilayered CPWs.
Coplanar Waveguides, Effective Relative Permittivity, Characteristic Impedance, Artificial Neural Networks