Triple band Yagi antenna design and performance prediction for next generation (6G) wireless applications using supervised regression models

Abstract

This study presents a high-performance Yagi antenna designed for sub-terahertz (sub-THz) and terahertz (THz) wireless communication applications, addressing the growing demands of emerging technologies such as high-resolution imaging, short-range secure communication, sensing, spectroscopy, and next generation 6G systems. Developed on a polyimide substrate with compact dimensions of 53 μm × 47 μm, the proposed antenna achieves exceptional directional radiation characteristics across the targeted frequency bands. Simulation results reveal a maximum gain of 14.9 dB and an outstanding efficiency of 99.65%, highlighting its suitability for advanced high-frequency applications. To ensure reliable performance evaluation, the antenna is analyzed and validated using multiple approaches, including full-wave simulations through CST, HFSS, and Altair FEKO, as well as curve-fitting techniques and an RLC equivalent circuit model. In addition, machine learning techniques are employed to predict antenna efficiency based on extensive simulation datasets, incorporating both neural network models and supervised regression algorithms. Among the evaluated models, Random Forest Regression exhibits superior prediction accuracy compared to other neural network and regression methods. The integration of high gain, near-perfect efficiency, and robust predictive modelling establishes the proposed Yagi antenna as a promising candidate for future sub-THz and THz wireless systems, particularly for 6G communication infrastructure.