Machine learning technique based highly efficient slotted 4-port MIMO antenna using decoupling structure for sub-THz and THz 6G band applications

This article presents the results of our investigation into a machine learning (ML) method for 6G MIMO antenna performance improvement in the THz band. Regarding the lowest frequency, the MIMO antenna is developed on a polyimide substrate with dimensions of 0.614 lambda 0 in length and 0.614 lambda 0 in width. Besides its compact size, the suggested design boasts a maximum gain of 15 dB, an isolation of more than 31, and an efficiency rating of 97.40% at its maximum setting. When a DG exceeds 9.99, the ECC value falls below 0.0012. Moreover, a wide frequency band of 85.71%, which covers the sub-THz and THz bands, has been achieved. The variance score, R square, mean square error (MSE), mean absolute error (MAE), and root mean square error (RMSE) are some of the criteria that can be used to evaluate the performance of ML models. Out of the five ML models, random forest regression (RFR) has the smallest error and is the most accurate. For gain prediction, the accuracy is more than 95%, as shown by the R-square and var score. The suggested antenna is a serious competitor for 6G THz band applications, as shown by the results of the CST simulations and the predictions made by the machine learning methods.