基于信道容量约束的物理层安全语义通信OA

With advancements in artificial intelligence,semantic communication has emerged as a key research direction for 6G networks.However,the broadcast nature of wireless channels introduces privacy risks,necessitating secure solutions.Existing methods often rely on network-layer encryption,which suffers from operational complexity and potential key leakage.Physical layer security,leveraging the channel's time-variation and randomness,can circumvent complex cryptographic mechanisms.In response,this paper proposes a physical layer secure semantic communication framework based on channel capacity constraints.The core design goal is to maximize the legitimate channel's capacity while restricting the overall security capacity to a predefined threshold.A differentiable mutual information estimator is introduced to construct a dual-channel capacity estimation model that includes both legitimate users and eavesdroppers,which can be jointly optimized with the semantic communication system.Additionally,an innovative security sentence similarity metric is proposed to quantify the security performance of semantic communication.Simulation results demonstrate that the proposed scheme significantly enhances communication security at the cost of a slight decrease in transmission accuracy.