A Development of DNN Beamforming for Satellite-Ground Systems with Graphene's Conductivity and Dispersion Relation

To integrate satellite communication with graphene-enhanced signal propagation and Deep neural network (DNN) beamforming to enhance functionality in the 6 GHz frequency band, we introduced the DDSGCD model for DNN Beamforming for Satellite-Ground Systems Using Graphene Conductivity and Dispersion. It analyses satellite coverage, including impacts from surrounding IMT-2020 base stations and various environmental losses. By graphene-based conductivity modelled through Kubo's formula, the model aims to improve antenna design for high-frequency usage. The challenge of beam-squinting, typical with phased array antennas due to angle variability, is addressed with a DNN-based algorithm simulating True-time delay (TTD) strategies to ensure precision while minimizing hardware demands. The results we discussed for the DDSGCD model of ECC, diversity gain, output signal voltage prediction, MSE, and probability distributions. By achieving an INR of around 6 dB with an SD of 0.2 dB, with a sample range of 5.8 to 6.2 dB. Over 180 epochs of MSE is decreased <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$1 \times 10^{\wedge}-6$</tex> to around <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$1.2 \times 10^{\wedge}-12$</tex>. With 2500 samples, the DNN beamforming shows -60 dB to 72 dB. A frequency range of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$2-5 \text{THz}$</tex>, with and without MTM, at 2 THz, of 9.95 dB compared to 8.0 dB, the ECC drops from 0.22 to around 0.035 with MTM.