Beam alignment (BA) is a crucial process in millimeter-wave (mmWave) communications for precise directional transmission and efficient link establishment, but its open nature makes it vulnerable to eavesdropping and identity impersonation attacks. To this end, we propose a novel security framework named CovertAuth, designed to provide a unified defense against both threats. For eavesdropping attacks, we advance the existing covert communication design by extending its joint optimization of the beam training budget and transmission power to a more practical BA scenario that incorporates the mutual coupling (MC) effect in antenna array impairments and imperfect eavesdropper channel state information. For impersonation attacks, the MC effect is explored as a device feature to design a novel adaptive weight-based physical layer authentication mechanism. Theoretical models for authentication metrics like detection and false alarm probabilities are first provided to conduct performance analysis. Based on these models, an optimization problem is constructed to determine the optimal weight value that maximizes authentication accuracy. With these optimal weights, a weighted-sum energy detector is employed to achieve identity validation. The resulting adaptive authentication scheme is then integrated with the covert communication to enhance the security of the BA phase. Finally, simulation results demonstrate that CovertAuth achieves high detection accuracy while satisfying the covertness requirement, offering a comprehensive security solution for the mmWave BA stage.

Authentication in millimeter-Wave (mmWave) Multiple-Input Multiple-Output (MIMO) systems is a critical issue due to the unique characteristics of mmWave communication, such as highly directional beamforming and the ability to support massive device connectivity. To address this challenge, this paper proposes a novel low-complexity decision-level-based Distributed Physical Layer Authentication (DPLA) framework to combat identity-based impersonation attacks in mmWave MIMO systems. The DPLA framework leverages Beam Pattern (BP) deviation, which arises from hardware-specific gain errors, as a key authentication feature. A fusion center is introduced to make the final authentication decision by aggregating local decisions from multiple collaborative nodes, enabling multi-directional perception. Specifically, a low-complexity hybrid combining fusion rule is carefully designed to accommodate the fully connected structure of mmWave MIMO systems, balancing computational efficiency and authentication performance. A rigorous performance analysis is conducted by deriving closed-form analytical expressions for the probabilities of correct detection and false alarm. Furthermore, the asymptotic detection and discrimination performance are systematically analyzed in the large-scale antenna regime. To further enhance authentication accuracy, digital signaling matrices are designed using the deflection coefficient maximization principle. The feasibility of the proposed framework is validated through a comprehensive evaluation, demonstrating its superior robustness and efficiency compared to benchmark methods.