Secure MIMO Communication via Pilot-Based Codebook Artificial Noise: Unknown Eavesdropper Channel

Pilot-Based Codebook Artificial Noise (PCAN), which plays a crucial role in ensuring secure communication, is an emerging physical layer security technique. PCAN faces a significant challenge due to the unknown eavesdropper channel. To address this issue, this letter re-models the eavesdropper channel using random matrix theory. By applying the η-transform and Shannon-transform, the asymptotic expression for the eavesdropper channel capacity is derived under the condition of an unknown eavesdropper channel. The simulation result shows that when the number of transmitter antennas equals the number of eavesdropper antennas and the system dimension is large, the model can effectively characterize the eavesdropper channel capacity.Within this framework, the secrecy capacity is optimized using the Difference of Convex functions - Successive Convex Approximation (DC-SCA) algorithm, providing a new approach to improve the performance of PCAN.