Publication

Abstract : Designing antennas that can perform at high levels for modern wireless communication is something tedious and demanding. Traditional methods rely on iterative simulations, requiring lots of processing engines and expert fine-tuning. These normal processes run a number of electromagnetic simulations while trying out different configurations of the design-an inefficient yet cost-heavy process. For these challenges, the work is machine-learning based for the design and optimization of microstrip patch antennas within the MATLAB site. The proposed method will very much improve the existing traditional simulation-heavy antenna design and optimization-through advanced machine learning mechanisms like Random Forest Regression and Bayesian optimization, predicting performance parameters like return loss (S11), gain, bandwidth, and efficiency of the antennas without much simulation. The approach is easily adaptable to all frequency bands, thus making it highly practical for next-generation wireless technologies, such as IoT, 5G, and satellite communication. With reduced iterative counts and computational expense, the proposed machine-learning driven framework can emerge as an efficient, scalable, and cost-effective solution for the rapid advancement of high-performance antennas in increasingly dynamic wireless communication spheres.