Parametric Modeling of Third-Order Waveguide Filter Using Combined Pole-Residue-Based Transfer Functions with Neural Networks

Microwave filters are essential passive components in contemporary wireless communication systems, and the design process based on electromagnetic (EM) simulations is a crucial aspect of filter development. Artificial neural networks (ANNs) methods have found widespread application within the scope of parametric modeling techniques. This paper introduces an advanced technique that combines neural networks with pole-residue-based transfer function (Neuro-TF) to perform parametric modeling of a third-order waveguide filter. The experimental results demonstrate that, compared to alternative methods, the pole-residue-based neuro-TF model significantly improves modeling accuracy.