Affiliation:
1. Oriental College of Technology
2. Barakatullah University Institute of Technology, Bhopal
3. Bangalore Technological Institute
4. Madhyanchal Professional University
Abstract
In this work, a two-port graphene-ceramic-based antenna is designed and discussed. Some
important features for the proposed two-port radiator are (a) the feed
layout produces dual radiating modes, i.e., HEM11δ and HEM12δ inside the ceramic at 2.6 and
5.2 THz, respectively; (b) the asymmetrical swastik-shaped
aperture creates the circularly polarized waves within dual operating
bands, i.e., 2.71–2.84 and 5.15–5.42 THz; (c) a
coating of graphene over ceramic makes the radiator frequency tunable
by varying its chemical potential; and (d) deep neural network (DNN)
and XGBoost-based machine learning (ML) algorithms are used to predict
the |S11| parameter of the designed antenna in
order to reduce the computational complexity. By comparing the
optimized outcome obtained from HFSS with the CST EM simulator and
predicted values from ML algorithms, it is verified that the designed
two-port radiator works in dual frequency bands, i.e., 2.4–3.2
and 5.0–5.43 THz. Stable values of MIMO, as well as the
far-field parameter, confirm the applicability of the designed antenna
for THz wireless applications.