A Novel Design of Energy Efficient & Spectral Efficient with Deep Secured Symbol Level Autoencoder Integration with NOMA & Deep Neural Network Model

This paper proposes a novel approach for device-to-device communication using NOMA and Deep Neural Network (DNN). This novel proposal focus to enhance the security of millimetre-wave (mmWave), massive non-orthogonal multiple access (NOMA) communication systems with integration of Deep Secure Symbol Level Autoencoder (DSSLA) with deep learning capabilities for secure symbol-level encoding and decoding. This method not only boosts the confidentiality and reliability of information transmission but also adds a layer of security to the communication process. To further optimize signal transmission and mitigate channel impairments characteristic of mmWave environments, the method employ hybrid precoding techniques that combine the strengths of digital and analog precoding. This implementation is carried out in MATLAB, utilizing its computational process to conduct extensive simulations and performance analyses. The performance factors considered are Spectral efficiency, Energy efficiency, Code book size, and Outage probability. The outcomes show how successful the DSSLA-based security measures and hybrid precoding strategies. It is inferred that the proposed method outperforms the classical methods namely OMP Precoding, PMA Precoding, and Deep Learning Hybrid Precoding in terms of Spectral efficiency and Energy efficiency.