The Role of Artificial Noise in RSMA-based Secure Communications

‍ ‍This work focuses on investigating the role of artificial noise （AN） and the corresponding power allocation of different streams in enhancing the secrecy rate performance for the rate-splitting multiple access （RSMA） based wireless networks. In RSMA， user messages are split and encoded into common and private streams. Contrary to the conventional idea that regarding the multi-user interference （MUI） as performance bottleneck and fully treating it as noise， by partially decoding the MUI and partially treat it as noise， the common stream in RSMA is able to serve not only as useful data for intended legitimate users， but also AN to confuse the external eavesdroppers. Considering this， whether an additional AN is needed to enhance secrecy is worth investigating. Aiming to ensuring user fairness， the transmit beamforming design with or without additional AN is formulated to maximize the minimum achievable secrecy rate. To deal with the non-convexity of the formulated problem， a successive convex approximation-based optimization solution is proposed. Numerical results showed that under perfect CSI condition， whether in overloaded （including strictly overloaded） or underloaded scenarios， the achievable secrecy rate achieved by relying solely on the common stream in RSMA as the noise to interfere with external single-antenna eavesdroppers is almost the same as the secrecy rate of AN-assisted RSMA secure precoding scheme， and the additional AN cannot further improve the secrecy rate performance of the system.