High Performance Quad Port Compact MIMO Antenna for 38 GHz 5G Application with Regression Machine Learning Prediction

Combining machine learning with multiple-input multiple-output (MIMO) antennas requires a careful approach that includes the latest advancements in wireless communication for 5G technology. This antenna is built using Rogers 5880 material, known for its excellent high-frequency performance. It achieves a strong isolation level of 28 dB, which reduces interference between channels and improves signal clarity. The operative bandwidth ranges from 35.739 to 39.289 GHz, critical for high data rates in 5G while keeping a return loss of − 10 dB or better. The antenna has a maximum gain of 8.5 dB and an efficiency of 97.41%, meaning it effectively translates power into strong signals. Its small size of 21 mm × 21 mm makes it ideal for compact devices without sacrificing performance. This article explores methods for evaluating the antenna’s fitness for 5G, including advanced simulations and an RLC circuit model. We use the Advanced Design System (ADS) to create a detailed model and compare the results with CST Microwave Studio (CST MWS), directing on return loss metrics. After simulations, we apply regression machine learning techniques to improve predictive accuracy using a dataset from CST MWS. Among the tested methods, decision tree regression is particularly effective, providing accurate efficiency predictions. Overall, this antenna design is strong for modern 5G communication systems, ensuring reliable performance and advancing wireless connectivity.