Security optimization and beamforming design for active RIS-assisted UAV relaying NOMA networks

Physical layer security (PLS) in non-orthogonal multiple access (NOMA) networks faces critical challenges, especially for eavesdropping risk of cell-edge users in conventional passive reconfigurable intelligent surface (RIS)-assisted unmanned aerial vehicles (UAV) relaying systems, which may be increased because of channel sharing and multiplicative fading. Unlike ground-based RIS deployments limited by installation constraints and multi-path signal degradation, this work proposes an active RIS-assisted UAV (U-ARIS) for multi-user multiple-input single-output (MU-MISO) NOMA systems with two legitimate users and one eavesdropper. Moreover, we formulate a joint optimization problem to maximize sum secrecy rate by coordinating base station beamforming and U-ARIS reflection beamforming. Due to the multivariate nature of optimization variables and complex non-convexity, an alternating optimization (AO) algorithm is designed to decompose the problem into two sub-problems, which can be solved by semidefinite relaxation (SDR) and successive convex approximation (SCA). Simulations demonstrate that the proposed U-ARIS-assisted NOMA scheme achieves secrecy rate improvements of 12.5% and 50% compared to UAV-passive RIS (U-PRIS) and conventional OMA systems, respectively.