Design and Synthesis of a Dual Band Series Feed Microstrip Patch Antenna Arrays using Machine Learning and Deep Learning Algorithms

This work explores the application of machine learning (ML) and deep learning (DL) techniques in the predictive modeling of antenna parameters, specifically for a binomial triangular microstrip patch antenna designed for 5G communication. Rather than relying solely on computationally expensive electromagnetic simulations, this study leverages Physics-Informed Neural Networks (PINNs) and the Koopman Theory Algorithm to establish accurate relationships between the antenna’s design parameters and its resonant frequencies and S-parameters. The ML models were trained to predict key performance metrics, such as the return loss and resonant frequencies , based on the antenna’s physical dimensions. This approach significantly reduces simulation time while maintaining high accuracy in performance estimation. The results confirm that data-driven methodologies can effectively capture the complex, nonlinear dependencies in antenna design, enabling rapid optimization. By integrating computational electromagnetics with AI-driven modeling, this work highlights the potential of ML and DL in revolutionizing the design and analysis of advanced wireless communication systems.