A hybrid neural model for the characterization of a single layer SIW waveguide

Abstract

In this work, a hybrid neural model (HNM) able to characterize accurately the dispersion behavior of the fundamental TE₁₀ mode of a single layer SIW waveguide, is developed. The HNM combines the analytical expression that models the dispersion characteristics of this guiding structure, with a Multi Layer Perceptron Neural Network (MLPNN) which operates as an estimator of the cutoff angular frequency ω_t of the fundamental mode. The comparison among HNM computations, numerical results obtained with methods proposed in literature and full-wave data validate both the accuracy and the effectiveness of the proposed approach.