Machine learning-based four-port MIMO antenna design with U-shaped geometry for ultra-fast wireless systems and 6G high speed communication

In this research, a high-performance MIMO antenna system is designed and analysed for advanced wireless communication applications. The proposed antenna features a compact structure with four radiating elements optimized for operation in the THz frequency range. The compact size of the structure is 100 × 100 µm 2 . The structure is giving a high bandwidth of 20 THz. The results are conducted to evaluate key performance metrics ECC, MEG, TARC, DG, and CCL. The results demonstrate excellent impedance matching with S11 values below –10 dB across the desired band, along with ultra-low ECC (<0.003), indicating minimal inter-element correlation. The antenna achieves a high diversity gain close to 10 dB and low TARC, low CCL of 0.01 bits/Hz, ensuring robust multi-port excitation performance. Machine learning was employed as a predictive model to estimate antenna performance, achieving a high accuracy with an R 2 value of 0.98. The bandwidth between 27 and 33 THz has an operational bandwidth of the proposed four-port MIMO antenna. Additionally, the metamaterial-inspired design enhances the gain and isolation while maintaining a compact footprint. These characteristics make the proposed MIMO antenna highly suitable for next-generation high-speed communication systems such as 6G, THz wireless networks, and wearable or IoT platforms.