Uplink Secrecy in RIS-Aided MIMO-NOMA Networks with User-Centric Artificial Noise

This paper investigates the uplink secrecy performance of a reconfigurable intelligent surface (RIS)-aided multiple-input multiple-output non-orthogonal multiple access (MIMO-NOMA) network. In the proposed system model, users communicate with the base station (BS) via a strategically partitioned RIS. However, a silent passive eavesdropper overhears the data through a direct link. As part of the proposed security design, we propose a user-centric artificial noise (AN) transmission strategy that degrades the decoding ability of the eavesdropper without sacrificing the power budget. The AN is nullified in the composite channel utilizing singular value decomposition. After analyzing the channel statistics, which consider all channels experiencing Nakagami- <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$m$</tex> fading, we derive theoretical closed-form expressions for the ergodic secrecy rate (ESR) for each user in the NOMA pair using Gauss-Chebyshev quadrature and Taylor-Maclaurin expansions. Monte Carlo simulations validate the correctness of our analysis. The numerical results confirm that our proposed scheme consistently achieves a positive ESR. We also explore the impact of channel fading and RIS-partitioning ratio on the uplink ESR and comment on optimal power allocation. Notably, the results also confirm that the proposed system outperforms the benchmark MIMOorthogonal multiple access-based scheme.