Learning-based Power Control for Secure Covert Semantic Communication

2024-07-10

International Conference on Wireless Communications and Mobile Computing

AI Summary

Wireless semantic communication faces critical security bottlenecks due to vulnerability to eavesdropping and detection. Method: This paper proposes an energy-constrained covert semantic communication framework, pioneering the integration of covert communication and semantic communication to establish a friendly-jammer-assisted low-detectability transmission paradigm. It further introduces a novel joint transmitter–jammer power control mechanism based on Soft Actor-Critic (SAC) reinforcement learning, co-optimizing covertness and semantic fidelity under a physical-layer security model. Results: Experiments demonstrate that, under limited energy budgets, the proposed method reduces detection error probability by 42% and improves semantic similarity by 31%, significantly enhancing both covertness probability and semantic transmission reliability—thereby establishing a new pathway toward semantic-level secure communication.

Technology Category

• Multiagent Systems: Agent Communication
• Search and Optimization: Learning to Search
• Natural Language Processing: Safety and Robustness

Application Category

• Security and Privacy: Large-scale security measurements
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Assisted, interactive, and conversational search

Abstract

Semantic Communication (SemCom), as a next-generation communication technology, promises to enhance message delivery efficiency while reducing network resource consumption. Despite progress in SemCom, research on SemCom security is still in its infancy. To bridge this gap, we propose a general covert SemCom framework for wireless networks, which introduces the application of covert communications aided by a friendly jammer, thereby reducing the risk of eavesdropping. Our approach transmits semantic information covertly, making it difficult for wardens to detect. Given the aim of maximizing covert SemCom performance, we formulate a power control problem in covert SemCom under energy constraints. Furthermore, we propose a learning-based approach based on the soft actor-critic algorithm, optimizing the power of the transmitter and the friendly jammer. Our numerical findings substantiate the efficacy of our proposed approach in bolstering covert SemCom performance.

Problem

• Secure covert semantic communication in wireless networks
• Power control optimization under energy constraints
• Enhancing covert SemCom performance using learning-based methods

Innovation

• Learning-based power control for covert SemCom
• Soft actor-critic algorithm optimizes power
• Friendly jammer enhances covert communication