Dual-band graphene-based THz MIMO antenna for 6G-enabled biomedical application using supervised regression machine learning validation

The exponential increase in wireless data traffic and the growing demand for biomedical sensing have driven the advancement of sophisticated antenna technologies, particularly within the terahertz (THz) frequency range. This research presents an innovative graphene-based microstrip patch antenna featuring a slotted design and MIMO configuration, specifically designed for the high-speed needs of 6G communications and sensitive biomedical applications. The proposed antenna features a compact dual-band structure with a narrowband of 0.2839 THz (1.7654 to 2.0493 THz) and a wideband of 3.8201 THz (2.6412 to 6.4613 THz), achieving a peak gain of 12.83 dB, high efficiency (93.1%), and excellent isolation (-38.8 dB). A novel slot geometry, including hexagonal, triangular, and circular trous slots, enhances impedance matching, bandwidth, and radiation properties. To address the limitations of conventional THz antenna design, the study integrates an RLC equivalent circuit model and machine learning (ML)-driven regression algorithms to optimize performance. The ML-based prediction model validated against simulated data demonstrates high accuracy and significant reductions in design iteration time. Material analyses confirm graphene and polyimide as optimal choices for high-efficiency, miniaturized THz devices. The design progression from single-element to MIMO configurations with silver-based decoupling structures ensures low mutual coupling, making the antenna suitable for high-data-rate, low-latency wireless systems. The comprehensive evaluation encompassing ECC, DG, CCL, TARC, MEG, and radiation characteristics highlights the antenna's potential in both wireless communications and biomedical applications.