With the rapid advancement of 6G, identity authentication has become increasingly critical for ensuring wireless security. The lightweight and keyless Physical Layer Authentication (PLA) is regarded as an instrumental security measure in addition to traditional cryptography-based authentication methods. However, existing PLA schemes often struggle to adapt to dynamic radio environments. To overcome this limitation, we propose the Adaptive PLA with Channel Extrapolation and Generative AI (APEG), designed to enhance authentication robustness in dynamic scenarios. Leveraging Generative AI (GAI), the framework adaptively generates Channel State Information (CSI) fingerprints, thereby improving the precision of identity verification. To refine CSI fingerprint generation, we propose the Collaborator-Cleaned Masked Denoising Diffusion Probabilistic Model (CCMDM), which incorporates collaborator-provided fingerprints as conditional inputs for channel extrapolation. Additionally, we develop the Cross-Attention Denoising Diffusion Probabilistic Model (CADM), employing a cross-attention mechanism to align multi-scale channel fingerprint features, further enhancing generation accuracy. Simulation results demonstrate the superiority of the APEG framework over existing time-sequence-based PLA schemes in authentication performance. Notably, CCMDM exhibits a significant advantage in convergence speed, while CADM, compared with model-free, time-series, and VAE-based methods, achieves superior accuracy in CSI fingerprint generation.

As a typical scenario for the 6th Generation mobile communication systems (6G), Hyper Reliable Low Latency Communication (HRLLC) is expected to ensure extremely low delay and high reliability, while supporting wireless transmission of large-scale massive data. However, existing communication networks face the dual challenges of inadequate performance metrics and limited network resources. Therefore, this paper proposes the Flexible Bit and Semantic on-demand Transmission (FBST) framework, including three key technologies: adaptive transmission mode decision, flexible transmission time interval scheduling, adjustable semantic compression ratio. The FBST framework could satisfy the strict QoS requirements of users and provide on-demand services for users. Based on the Stochastic Network Calculus (SNC) modeling method, we conduct precise delay analysis and provided a general expression for the delay violation probability of the α - κ - μ channel, which could be extended to various complex channels. In addition, the Knowledge-base Parameterized Deep Q-Network (KP-DQN) algorithm is proposed to solve the resource allocation issue, which is a mixed action space problem with complex calculations caused by SNC. Finally, the simulation results show that FBST framework could satisfy extremely strict delay and reliability requirements of users, and the KP-DQN algorithm improving operational efficiency by over 76.8%.