Deep Learning Based Data-Driven Insights Into Microstrip Antenna Performance

Abstract

Microstrip antennas are widely used in modern communication systems due to their low profile, lightweight, and ease of integration with circuits. However, predicting their performance across various parameters such as bandwidth, gain, and VSWR remains a challenge, traditionally relying on empirical methods and complex simulations. Over the years, advancements in deep learning have opened new avenues for more accurate and efficient prediction models. The traditional approach to antenna design typically involved time- consuming and resource-intensive processes, including manual tuning, trial-and-error, and reliance on antenna simulation software. These methods, though effective, are limited in scalability and accuracy when dealing with vast design variations. In this research, we propose a data-driven approach using Multi-Layer Perceptron (MLP) regression models to predict key performance metrics of microstrip antennas. By training on a dataset containing diverse antenna parameters, this model offers improved accuracy and prediction speed over traditional methods. The proposed system aims to provide designers with a robust tool for optimizing antenna performance, reducing reliance on simulations, and expediting the design process. This deep learning- based solution demonstrates the potential for enhancing microstrip antenna design, making it more efficient, scalable, and precise.