Physical Layer Security Authentication Based on Channel Multi-feature Integration for Downlink

The research on physical layer security authentication schemes for downlink transmission in massive Multiple Input Multiple Output (Massive MIMO) systems is relatively limited, and the existing research has mainly focused on the authentication problem in massive MIMO uplink transmission, while the research on the downlink has not been sufficient. Although some methods originally designed for conventional MIMO are also applicable to massive MIMO systems, these algorithms are not optimized, and thus the authentication performance may suffer. In order to effectively defend against attacks from illegal devices, it is urgent to design secure authentication schemes at the physical layer for downlink transmission. Based on the above problems, the security of downlink transmission of Massive MIMO systems was studied. A novel downlink physical layer security authentication scheme based on channel multi-feature fusion is proposed, which exploits the arrival time and arrival angle characteristics of the channel at the user device side. The authentication problem is formulated as an image single classification recognition problem, and the image entropy and structural similarity metric (SSIM) of the channel image are extracted and learned by the Support Vector Machine (SVM) algorithm. The proposed scheme is evaluated by simulations and shows good performance.