Design of UWB MIMO Antenna Using Connectivity-Aware Graph Neural Network for Smart Fabric Communications

Abstract

The design of an ultra-wideband (UWB) multiple-input multiple-output (MIMO) antenna for smart fabric communications can significantly improve the reliability of wireless connections and overcome challenges such as channel fading and shadowing caused by the human body. This technology proves valuable for security applications, location tracking, and patient monitoring. Cotton fabric serves as the substrate material for the proposed textile antenna. With a connectivity-aware graph neural network (CAGNN), a UWB MIMO antenna is designed with four octagonal radiators, each loaded with multiple slots, offering a frequency range of 2.9–12 GHz. The unit cell measures $0.36{\lambda }_0\times 0.164{\lambda }_0\times 0.014{\lambda }_0$, and the MIMO antenna measures $0.48{\lambda }_0\times 0.48{\lambda }_0\times 0.014{\lambda }_0$. Performance metrics, like channel capacity loss (CCL), diversity gain (DG), total active reflection coefficient (TARC), envelope correlation coefficient (ECC), and specific absorption rate (SAR), are considered. The efficiency of the proposed approach is compared with existing approaches. The proposed CAGNN-DUWBMIMO technique is implemented. The proposed approach achieves 26.36%, 20.69%, and 30.29% better ECCs; 19.12%, 28.32%, and 27.84% better DGs; and 12.04%, 13.45%, and 22.80% better SARs compared with existing models: design with the analysis of a UWB MIMO antenna for smart fabric communications (DA-UWBMIMO), design with the analysis of a UWB MIMO antenna for smart fabric communications (DDPT-UWBMIMO), development including the comprehensive assessment of a double-port textile UWB MIMO antenna for biomedical use (DDBW-MIMOA), respectively.