Robust Secure Precoding for UAV-Aided Multi-Beam Satellite NOMA Communications

The wide coverage and broadcasting characteristics of satellite communications lead to multi-beam downlinks being vulnerable to security threats, such as eavesdropping, hacking and illegal access. This article takes into account the case of multiple users and an eavesdropper (Eve) in the target beam. In particular, we consider the deployment of an unmanned aerial vehicle to generate artificial noise in order to confuse Eve, while acting as a relay for the legitimate users. In addition, a non-orthogonal multiple access (NOMA) strategy is used to support multi-user communication and to improve the transmission rate. Considering the constraints due to quality of service, total and per-beam transmit power of the satellite, two robust secure precoding algorithms are presented to maximize the minimal achievable secrecy rate of the legitimate users for both non-critical and critical applications. Since the formulated optimization problems are non-convex, we first use the arithmetic-geometric mean inequality to solve the non-convex constraint of the successive interference cancellation decoding order, and the logarithmic parameter form is addressed by using the first-order Taylor series expansion. Besides, the secure outage probability constraint of the critical case is effectively resolved by applying the Bernstein-type inequality/decomposition-based large deviation inequality. Moreover, semi-definite relaxation and penalty function optimization methods are adopted to design the transmit power of the satellite in two cases, respectively. Simulation results verify the effectiveness and superiority of the proposed robust precoding design methods.