Machine learning-based novel-shaped THz MIMO antenna with a slotted ground plane for future 6G applications

Abstract

This study discusses the results of using a regression machine learning technique to improve the performance of 6G applications that use multiple-input multiple-output (MIMO) antennas operating at the terahertz (THz) frequency band. This research evaluates an antenna’s performance using various methodologies, such as simulation and RLC equivalent circuit models. The suggested design has a broad bandwidth of 2.5 THz and spans from 6.2 to 8.7 GHz, a maximum gain of 14.59 dB, and small dimensions (100 × 300) µm2. It also has outstanding isolation exceeding − 31 dB with 96% efficiency. The ADS allowed us to confirm the accuracy of the CST results by creating a simulated version of the same RLC circuit. Reflection coefficients obtained from the CST and ADS simulators are similar. The supervised regression ML approach is employed accurately to predict the antenna’s potential gain. Several metrics, such as the variance score, R square, mean square error (MSE), mean absolute error (MAE), and root mean square error (RMSE), can evaluate machine learning (ML) models. Out of the six machine learning models analyzed, the Extra Tree Regression model demonstrates the lowest error and achieves the highest level of accuracy in predicting gain.