Machine Learning Inspired Impedance Matching Prediction of Fractal Gasket Antenna

Abstract

This work aims to provide a machine learning inspired design technique of fractal gasket antenna (FGA). Considering the diverse applications of fractal gasket antenna in providing high pattern diversity, low cross-polarization, etc., it has become a popular candidate in next generation planar pcb technology. The planar structure of the fractal gasket provides the benefit of easier integration with other circuit components. However, the design and analysis of FGA is dominantly done through full-wave electromagnetic simulation which has a certain amount of computation and memory requirement. In this regard, the present work adopts the multi layered perceptron (MLP) neural network approach to design an impedance matched FGA. The proposed data driven approach extensively investigates the variation study of different parameters of the neural network. The proposed technique has the future scope of extensibility for other types of antenna geometries.