Satellite communication is a cornerstone for achieving global connectivity in future 6G networks. However, wireless transmissions are inherently vulnerable to eavesdropping due to their broadcasting nature, and satellite networks are no exception. Phy...
Satellite communication is a cornerstone for achieving global connectivity in future 6G networks. However, wireless transmissions are inherently vulnerable to eavesdropping due to their broadcasting nature, and satellite networks are no exception. Physical layer security (PLS) offers a complementary approach to cryptography by exploiting channel characteristics. Yet, fast time variation, Doppler shifts, intermittent links, and tight power budgets in low Earth orbit (LEO) satellite constellations make PLS design complicated. Prior studies in PLS have mainly focused on single-satellite links or fixed access patterns, while the multi-satellite scheduling problem has received relatively less attention. Particularly, satellite scheduling to enhance data transmission with protective interference (i.e., artificial noise) for security purposes in the LEO satellite network has not yet been investigated.
In this thesis, we propose an integrated PLS framework that jointly leverages artificial noise (AN), satellite scheduling, and deep learning. We formulate an optimization problem that selects satellites for data transmission as well as artificial noise generation to prevent potential eavesdropping while preserving the quality of legitimate links. Further, we develop a learning-based satellite scheduling algorithm to guarantee low-latency scheduling decisions without heavy online optimization, applicable to a dynamic channel environment and limited CSI availability. We evaluate our proposed methods in terms of secrecy rate and secrecy outage probability under realistic power and bandwidth constraints. We further validate that the proposed methods can achieve robust physical-layer security by coupling coordinated AN with learning-driven multi-satellite scheduling. Our results suggest a promising direction of future PLS design for 6G non-terrestrial networks, which can reduce dependence on heavy key management and enable globally secure satellite connectivity.