Artificial Noise-Aided Secure Precoding for RSMA Satellite Downlinks Under Channel Uncertainty

Abstract

Rate-splitting multiple access (RSMA) has recently emerged as a promising transmission paradigm for enhancing spectral efficiency through interference management in satellite-terrestrial integrated networks. However, the broadcast nature of satellite downlinks makes them inherently vulnerable to eavesdropping, especially when the channel state information (CSI) of both legitimate users and potential eavesdroppers (Eves) is imperfect. In this paper, we investigate a robust secure precoding design for an RSMA-based satellite system serving multiple legitimate ground users in the presence of multiple Eves, with imperfect CSI for both the legitimate users and the Eves. To enhance physical layer security, artificial noise (AN) is deliberately injected into the satellite transmit signal to degrade the eavesdropping links while preserving the quality-of-service requirements of legitimate users. A robust secure precoding framework is developed by jointly optimizing the information-bearing and the AN precoding matrices subject to the total transmit power constraint at the satellite. The objective is to maximize the sum secrecy rate, while accounting for CSI uncertainty of both the legitimate users and the Eves. The resulting non-convex optimization problem is reformulated into a tractable form and efficiently solved via an iterative penalty function. Numerical results demonstrate that the proposed AN-aided RSMA secure precoding scheme significantly outperforms conventional multiple access schemes in terms of secrecy performance and robustness against the Eve CSI uncertainty.