Performance prediction of a closely spaced dual port dual-band silicon-graphene antenna with high isolation for THz wireless applications using a machine learning algorithm

This article explains the design and optimization of a twin port silicon-graphene radiator in the THz frequency regime. The suggested radiator’s primary attributes are: 1) a plus-formed slot produces dual radiating mode (HEM11 δ and HEM12 δ ) inside the ceramic; 2) the presence of an absorber wall in between the twin port blocks the radiation between the ports and raises the separation level to above 35 dB as well; 3) using a graphene sheet over ceramic and metasurface (MS) provides the tunability in the proposed radiating structure; and 4) machine learning procedures, i.e., ANN and random forest, help to predict the expected S-parameters of the planned antenna. Actual (HFSS/CST software) and projected values (using the ML algorithm) confirm that the design radiator works effectively in the twin working spectrum, i.e., 3.3–3.98 and 4.9–5.45 THz. A good value of the far-field and MIMO parameters confirms its applicability in THz-built biomedical uses.