智能语义隐蔽通信综述

冯照心; 徐逸凡; 邢成文; 徐煜华; 赵楠; 王金龙

doi:10.11999/JEIT260184

智能语义隐蔽通信综述

doi: 10.11999/JEIT260184 cstr: 32379.14.JEIT260184

1.
大连理工大学信息与通信工程学院大连 116024
2.
陆军工程大学通信工程学院南京 210000
3.
北京理工大学信息与电子学院北京 100081

基金项目: 国家自然科学基金(U23A20271, 62325103, 62271099)

详细信息

作者简介:
冯照心：男，博士生，研究方向为人工智能、语义隐蔽通信

徐逸凡：男，讲师，研究方向为动态频谱抗干扰、智能博弈决策

邢成文：男，教授，研究方向为机器学习、统计信号处理、凸优化、多元统计和阵列信号处理

徐煜华：男，教授，研究方向为无人集群通信、智能频谱对抗。赵楠：男，教授，研究方向为无人机通信与网络、安全与隐蔽通信等

王金龙：男，教授，研究方向为短波通信、认知无线电

通讯作者:
赵楠　zhaonan@dlut.edu.cn

中图分类号: TN915.0
计量
- 文章访问数: 8
- HTML全文浏览量: 3
- PDF下载量: 1
- 被引次数: 0
出版历程
- 收稿日期: 2026-02-11
- 修回日期: 2026-05-29
- 录用日期: 2026-05-29
- 网络出版日期: 2026-06-10

Survey on Intelligent Covert Semantic Communication

1.
School of Information and Communication Engineering, Dalian University of Technology, Dalian 116024, China
2.
College of Communications Engineering, Army Engineering University of PLA, Nanjing 210000, China
3.
School of Information and Electronics, Beijing Institute of Technology, Beijing 100081, China

Funds: The National Natural Science Foundation of China (U23A20271, 62325103, 62271099)

摘要

摘要: 面向第六代移动通信网络(6G)在超大规模连接、高可靠低时延等典型场景下的安全高效传输需求，语义隐蔽通信能够通过联合语义感知、语义压缩与隐蔽传输控制，在降低传输开销的同时实现信息的安全隐蔽传输，并提升低信噪比条件下的鲁棒性与资源利用率。此外，随着人工智能(AI)技术的快速发展，其具备的语义感知理解、智能决策优化与动态适应能力能够支撑语义重要性判别、自适应隐蔽策略优化以及语义驱动的智能隐蔽传输决策，为复杂动态场景下的智能语义隐蔽通信提供重要支撑。该文针对智能语义隐蔽通信展开综述。首先，介绍了语义通信、隐蔽通信以及语义隐蔽通信的基础概念。之后，探讨了人工智能赋能的智能语义隐蔽通信及其架构，并讨论了智能语义隐蔽通信的典型应用场景。然后，讨论了信道估计、频谱预测、波形设计等关键技术在智能语义隐蔽通信中的作用。最后，讨论了智能语义隐蔽通信面临的挑战和未来研究方向。
- 第六代移动通信 /
- 人工智能 /
- 语义隐蔽通信 /
- 语义通信 /
- 隐蔽通信
Abstract: Significance As the 6th generation mobile communication network (6G) evolves from the internet of everything to the intelligent connection of everything, the communication model is shifting from reliable bit transmission to effective semantic delivery. Semantic communication can extract and compress task-related semantics to reduce redundancy and resource consumption. However, with the highly structured nature, these semantics are vulnerable to eavesdropping and attacks. To address this challenge, covert communication is designed to prevent illegal monitoring and attacks by hiding the transmission behavior. With the support of artificial intelligence (AI), covert communication can leverage reinforcement learning to rapidly adjust power and resource allocation in the dynamic environment. Additionally, generative models can conceal signals within the environment by replicating their patterns. However, the stringent constraints imposed by covertness limit the transmission rates, making large-scale data transmission difficult. Intelligent semantic covert communication integrates semantic extraction with covert transmission, providing an effective and reliable solution for the secure and efficient 6G network. Progress With the development of AI, particularly the advances in deep learning for modeling complex features, semantic communication enables efficient semantic extraction and nonlinear compression of multimodal data. Moreover, research on semantic communication has shifted from separate source-channel coding to joint source-channel coding to achieve optimal performance through end-to-end training. For image transmission, convolutional neural networks can use local receptive fields to capture their spatial correlations. Moreover, for sequential data transmission, long short-term memory networks can use gating mechanisms to maintain temporal coherence. Additionally, generative AI, including generative adversarial networks and diffusion models, can learn the statistical patterns of environmental noise in the time, frequency and spatial domains to conceal the transmitted signals. Such approaches reduce the effectiveness of illegal monitoring and detection and enhance the adaptability of covert communication systems in the dynamic environment. AI has also significantly improved the autonomous decision-making in dynamic covert communication. By modeling covert transmission as a Markov decision process, deep reinforcement learning can learn to adjust resource allocation through interaction with the environment, thereby offering lower computational complexity than traditional convex optimization methods. Additionally, by integrating semantic extraction and covert transmission, intelligent semantic covert communication enables semantic-driven covert transmission, where large language models can assess semantic sensitivity and contextual risks, thereby enabling the selective covert transmission of sensitive semantic information. Conclusions Research on intelligent semantic covert communication demonstrates the advantages of close cooperation between semantic perception and physical-layer covert mechanisms. Moreover, the support of AI has significantly improved the efficiency of semantic extraction and the adaptability to dynamic and complex environment. Intelligent semantic covert communication guarantees efficiency and security for ubiquitous 6G services by integrating semantic understanding with covert transmission strategies. Prospects Future development of intelligent semantic covert communication should focus on key challenges, including AI-enabled detection, unified semantic metrics, model lightweighting, multimodal semantic alignment, system interpretability, and semantic hallucination. Active threat detection and adaptive strategies are essential to countering AI-driven surveillance. Moreover, causal reasoning within large multimodal models can mitigate semantic hallucinations and ensure reliable data transmission. In addition, advances in model compression technology and cloud-edge collaboration architectures are indispensable for deploying high-complexity AI models on resource-limited terminals. With the rapid development of AI, intelligent semantic covert communication can provide core support for the intelligent connection of everything, thereby helping build a more secure, efficient and reliable 6G network.
- The 6th generation Mobile Communication (6G) /
- Artificial Intelligence (AI) /
- Semantic Covert Communication /
- Semantic Communication /
- Covert Communication

HTML全文

图 1 语义通信基本架构

下载: 全尺寸图片幻灯片

图 2 人工智能关键技术分类概览

下载: 全尺寸图片幻灯片

图 3 智能语义隐蔽通信网络架构

下载: 全尺寸图片幻灯片

图 4 智能语义隐蔽通信应用场景

下载: 全尺寸图片幻灯片

表 1 AI驱动的语义隐蔽通信简要总结

研究方向	AI模型	关键作用机制	优点	缺点
智能语义通信	深度学习	从原始数据中学习语义特征与压缩映射关系	无需人工语义建模	语义度量不统一
智能隐蔽通信	强化学习与生成式AI	通过环境交互学习隐蔽策略、拟合背景噪声生成伪装信号	降低计算复杂度，提升实时性	训练过程的不稳定性
智能语义隐蔽通信	大语言模型	基于上下文理解与推理能力，进行语义感知和隐蔽策略决策	支持语义驱动的跨层联合优化	推理时延和算力需求较高

下载: 导出CSV

表 2 不同通信范式之间的对比与总结

对比维度	传统通信	语义通信	隐蔽通信	智能语义隐蔽通信
传输目标	比特无差错传输	语义信息准确恢复	通信行为的低可检测性	语义高效传输与通信行为隐蔽
传输效率	受信道容量约束	通过语义压缩提升效率	受平方根定律约束，传输速率较低	语义压缩缓解隐蔽约束，效率优于隐蔽通信
安全机制	无安全保障	无专门安全机制	通过降低可检测性抵御监测和干扰	语义重要性驱动的选择性隐蔽，实现差异化保护
资源开销	无额外开销	语义编解码计算开销	无额外开销	语义编解码的计算开销与隐蔽决策的AI推理开销
智能化水平	无	依赖数据驱动语义建模	以数值优化为主	AI驱动跨层联合优化
适用场景	通用通信场景	资源受限、低时延场景	高安全性与抗干扰场景	语义高效传输与隐蔽性协同保障的复杂通信场景

下载: 导出CSV

表 3 智能语义隐蔽通信关键技术简要总结

关键技术	参考文献	面临挑战	AI模型	优势
智能信道估计	文献[32]	低信噪比下难以精准获取信道信息	深度残差网络、生成式AI	非线性的信道特征重构
智能频谱预测	文献[33]	动态环境频谱规律难捕捉	深度学习	时序频谱特征的精准预测
多维波形设计	文献[26]	语义强结构性易暴露特征	生成对抗网络、扩散模型	拟合背景噪声统计分布
智能波束赋形	文献[34, 35]	时变信道下波束的动态调整	深度强化学习	实时的波束设计与优化
智能协同干扰	文献[36]	分布式节点的协调与同步	多智能体强化学习	去中心化的动态干扰策略
智能资源分配	文献[37, 38]	多维资源耦合的非凸优化	深度强化学习	低复杂度下实现实时决策
自适应传输调度	文献[39]	跨层参数耦合决策复杂	大语言模型智能体	语义信息驱动的传输决策

下载: 导出CSV

参考文献(43)

[1]	CUI Qimei, YOU Xiaohu, WEI Ni, et al. Overview of AI and communication for 6G network: Fundamentals, challenges, and future research opportunities[J]. Science China Information Sciences, 2025, 68(7): 171301. doi: 10.1007/s11432-024-4337-1.
[2]	WANG Yining, HAN Han, FENG Yixiao, et al. Semantic communication empowered 6G networks: Techniques, applications, and challenges[J]. IEEE Access, 2025, 13: 28293–28314. doi: 10.1109/ACCESS.2025.3532797.
[3]	CALVANESE S E and BARBAROSSA S. 6G networks: Beyond Shannon towards semantic and goal-oriented communications[J]. Computer Networks, 2021, 190: 107930. doi: 10.1016/j.comnet.2021.107930.
[4]	YAN Shihao, ZHOU Xiangyun, HU Jinsong, et al. Low probability of detection communication: Opportunities and challenges[J]. IEEE Wireless Communications, 2019, 26(5): 19–25. doi: 10.1109/MWC.001.1900057.
[5]	LI Zan, SHI Jia, SI Jiangbo, et al. Intelligent covert communication: Recent advances and future research trends[J]. Engineering, 2025, 44: 101–111. doi: 10.1016/j.eng.2024.12.007.
[6]	YIN Chengyu, LIU Yiliang, and SU Zhou. Semantic covert communication: Concealing sensitive image information via covertness-oriented artificial noise[J]. IEEE Journal on Selected Areas in Communications, 2026, 44: 1545–1560. doi: 10.1109/JSAC.2025.3648706.
[7]	IMT-2030(6G)推进组. 面向6G的语义通信技术研究报告[R]. 2025. IMT-2030(6G) Promotion Group. Research report on semantic communication technology for 6G[R]. 2025. (查阅网上资料, 未找到本条文献报告编号和英文翻译, 请确认).
[8]	刘方彧, 丁家润, 冯禹硕, 等. 大模型使能的语义通信研究现状与发展趋势[J]. 信号处理, 2025, 41(6): 993–1014. doi: 10.12466/xhcl.2025.06.002. LIU Fangyu, DING Jiarun, FENG Yushuo, et al. Research status and development trends in foundation model-enabled semantic communication[J]. Journal of Signal Processing, 2025, 41(6): 993–1014. doi: 10.12466/xhcl.2025.06.002.
[9]	YANG Wanting, DU Hongyang, LIEW Z Q, et al. Semantic communications for future internet: Fundamentals, applications, and challenges[J]. IEEE Communications Surveys & Tutorials, 2023, 25(1): 213–250. doi: 10.1109/COMST.2022.3223224.
[10]	WEAVER W. Recent contributions to the mathematical theory of communication[J]. ETC: A Review of General Semantics, 1953, 10(4): 261–281.
[11]	CHACCOUR C, SAAD W, DEBBAH M, et al. Less data, more knowledge: Building next-generation semantic communication networks[J]. IEEE Communications Surveys & Tutorials, 2025, 27(1): 37–76. doi: 10.1109/COMST.2024.3412852.
[12]	HUANG Jianhao, LI Dongxu, HUANG Chuan, et al. Joint task and data-oriented semantic communications: A deep separate source-channel coding scheme[J]. IEEE Internet of Things Journal, 2024, 11(2): 2255–2272. doi: 10.1109/JIOT.2023.3293154.
[13]	XU Jialong, TUNG T Y, AI Bo, et al. Deep joint source-channel coding for semantic communications[J]. IEEE Communications Magazine, 2023, 61(11): 42–48. doi: 10.1109/MCOM.004.2200819.
[14]	倪磊, 达新宇, 胡航, 等. 基于MP-WFRFT的多维联合调制卫星隐蔽通信研究[J]. 电子与信息学报, 2023, 45(4): 1183–1191. doi: 10.11999/JEIT220121. NI Lei, DA Xinyu, HU Hang, et al. Research on multi-dimensional joint modulation for satellite covert communication based on MP-WFRFT[J]. Journal of Electronics & Information Technology, 2023, 45(4): 1183–1191. doi: 10.11999/JEIT220121.
[15]	SIMMONS G J. The prisoners’ problem and the subliminal channel[M]. CHAUM D. Advances in Cryptology. New York: Springer, 1984: 51–67. doi: 10.1007/978-1-4684-4730-9_5.
[16]	BASH B A, GOECKEL D, and TOWSLEY D. Limits of reliable communication with low probability of detection on AWGN channels[J]. IEEE Journal on Selected Areas in Communications, 2013, 31(9): 1921–1930. doi: 10.1109/JSAC.2013.130923.
[17]	SHAHZAD K, ZHOU Xiangyun, and YAN Shihao. Covert communication in fading channels under channel uncertainty[C]. 2017 IEEE 85th Vehicular Technology Conference (VTC Spring), Sydney, Australia, 2017: 1–5. doi: 10.1109/VTCSpring.2017.8108525.
[18]	HE Biao, YAN Shihao, ZHOU Xiangyun, et al. On covert communication with noise uncertainty[J]. IEEE Communications Letters, 2017, 21(4): 941–944. doi: 10.1109/LCOMM.2016.2647716.
[19]	LI Zan, HUI Pei, GUAN Lei, et al. A frequency hopping scheme for covert communication with noise uncertainty[J]. IEEE Transactions on Communications, 2026, 74: 3132–3146. doi: 10.1109/TCOMM.2025.3649718.
[20]	LU Xingbo, YAN Shihao, YANG Weiwei, et al. Covert communication with time uncertainty in time-critical wireless networks[J]. IEEE Transactions on Wireless Communications, 2023, 22(2): 1116–1129. doi: 10.1109/TWC.2022.3201872.
[21]	WANG Chao, XIONG Zehui, ZHENG Meng, et al. Covert communications via two-way IRS with noise power uncertainty[J]. IEEE Transactions on Communications, 2024, 72(8): 4803–4815. doi: 10.1109/TCOMM.2024.3381672.
[22]	CHEN Yuhao, YANG Qianqian, SHI Zhiguo, et al. The model inversion eavesdropping attack in semantic communication systems[C]. 2023 IEEE Global Communications Conference, Kuala Lumpur, Malaysia, 2023: 5171–5177. doi: 10.1109/GLOBECOM54140.2023.10436996.
[23]	LUO Xuewen, CHEN H H, and GUO Qing. Semantic communications: Overview, open issues, and future research directions[J]. IEEE Wireless Communications, 2022, 29(1): 210–219. doi: 10.1109/MWC.101.2100269.
[24]	YANG Ke, WANG Sixian, DAI Jincheng, et al. SwinJSCC: Taming swin transformer for deep joint source-channel coding[J]. IEEE Transactions on Cognitive Communications and Networking, 2025, 11(1): 90–104. doi: 10.1109/TCCN.2024.3424842.
[25]	GHOURAB E M, ALHUSSEIN O, MI De, et al. Covert IRS-UAV networks empowered by deep reinforcement learning[J]. IEEE Journal on Selected Areas in Communications, 2026, 44: 1813–1826. doi: 10.1109/JSAC.2025.3642218.
[26]	于季弘, 林子砚, 叶能, 等. 基于三方生成对抗网络的隐蔽通信方法[J]. 通信学报, 2023, 44(11): 225–236. doi: 10.11959/j.issn.1000-436x.2023215. YU Jihong, LIN Ziyan, YE Neng, et al. Covert communication method based on tripartite generative adversarial network[J]. Journal on Communications, 2023, 44(11): 225–236. doi: 10.11959/j.issn.1000-436x.2023215.
[27]	GUO Shuaishuai, WANG Yanhu, YE Jia, et al. Semantic importance-aware communications with semantic correction using large language models[J]. IEEE Transactions on Machine Learning in Communications and Networking, 2025, 3: 232–245. doi: 10.1109/TMLCN.2025.3530875.
[28]	邓娜, 邢成文, 赵楠. 卫星隐蔽通信技术综述[J]. 中兴通讯技术, 2024, 30(5): 55–67. doi: 10.12142/ZTETJ.202405009. DENG Na, XING Chengwen, and ZHAO Nan. A review of satellite covert communication technologies[J]. ZTE Technology Journal, 2024, 30(5): 55–67. doi: 10.12142/ZTETJ.202405009.
[29]	FENG Lei, LIU Huaide, ZHAO Yikun, et al. Semantic communication for low-altitude economy: Harnessing the power of generative AI and edge intelligence[J]. IEEE Wireless Communications, 2026, 33(1): 36–44. doi: 10.1109/MWC.2025.3629007.
[30]	刘传宏, 郭彩丽, 杨洋, 等. 人工智能物联网中面向智能任务的语义通信方法[J]. 通信学报, 2021, 42(11): 97–108. doi: 10.11959/j.issn.1000-436x.2021214. LIU Chuanhong, GUO Caili, YANG Yang, et al. Intelligent task-oriented semantic communication method in artificial intelligence of things[J]. Journal on Communications, 2021, 42(11): 97–108. doi: 10.11959/j.issn.1000-436x.2021214.
[31]	XU Rui, LI Gaolei, WU Jun, et al. Toward covert and reliable communication for anti-eavesdropping transmission in V2X networks[J]. IEEE Transactions on Wireless Communications, 2025, 24(8): 6429–6442. doi: 10.1109/TWC.2025.3553132.
[32]	FENG Yunqi, SHAN Zhendong, SHEN Hesheng, et al. Digital semantic communication system with a learnable channel estimator for OFDM transmission[J]. IEEE Transactions on Cognitive Communications and Networking, 2025, 11(5): 3361–3370. doi: 10.1109/TCCN.2025.3533008.
[33]	ZHANG Hao, ZHOU Fuhui, WU Qihui, et al. Spectrum cognition: Semantic situation for next-generation spectrum management[J]. IEEE Network, 2025, 1–10. doi: 10.1109/MNET.2025.3604901.
[34]	HUANG Guojun, AN Jiancheng, YANG Zhaohui, et al. Stacked intelligent metasurfaces for task-oriented semantic communications[J]. IEEE Wireless Communications Letters, 2025, 14(2): 310–314. doi: 10.1109/LWC.2024.3499970.
[35]	ZHANG Yifei, GAO Zhen, ZHAO Jingjing, et al. AI empowered channel semantic acquisition for 6G integrated sensing and communication networks[J]. IEEE Network, 2024, 38(2): 45–53. doi: 10.1109/MNET.2024.3354264.
[36]	WANG Yining, HU Ye, DU Hongyang, et al. Multi-agent reinforcement learning for covert semantic communications over wireless networks[C]. 2023 IEEE International Conference on Acoustics, Speech and Signal Processing, Rhodes Island, Greece, 2023: 1–5. doi: 10.1109/ICASSP49357.2023.10095748.
[37]	WANG Lingyi, WU Wei, ZHOU Fuhui, et al. Adaptive resource allocation for semantic communication networks[J]. IEEE Transactions on Communications, 2024, 72(11): 6900–6916. doi: 10.1109/TCOMM.2024.3405355.
[38]	WANG Lingyi, WU Wei, ZHOU Fuhui, et al. IRS-enhanced secure semantic communication networks: Cross-layer and context-awared resource allocation[J]. IEEE Transactions on Wireless Communications, 2025, 24(1): 494–508. doi: 10.1109/TWC.2024.3495720.
[39]	HUANG Yudong, DU Hongyang, ZHANG Xinyuan, et al. Large language models for networking: Applications, enabling techniques, and challenges[J]. IEEE Network, 2025, 39(1): 235–242. doi: 10.1109/MNET.2024.3435752.
[40]	FU Qiang, YU Mingkai, SHI Yongpeng, et al. Tradeoff between covertness and transmission in NR V2X: Traditional and LLM-assisted cases[J]. IEEE Journal on Selected Areas in Communications, 2026, 44: 1561–1576. doi: 10.1109/JSAC.2025.3638734.
[41]	ZHANG Xuefei, WANG Yangfan, SUN Yao, et al. SeER: A unified transmission performance metric for wireless semantic communication[J]. IEEE Network, 2026, 40(3): 239–245. doi: 10.1109/MNET.2025.3578970.
[42]	XU Rui, LI Gaolei, YANG Zhaohui, et al. Covert UAV data transmission via semantic communication: A DRL-driven joint position and power optimization method[C]. 2024 IEEE/CIC International Conference on Communications in China, Hangzhou, China, 2024: 66–71. doi: 10.1109/ICCC62479.2024.10681776.
[43]	秦志金, 赵菼菼, 李凡, 等. 多模态语义通信研究综述[J]. 通信学报, 2023, 44(5): 28–41. doi: 10.11959/j.issn.1000-436x.2023105. QIN Zhijin, ZHAO Tantan, LI Fan, et al. Survey of research on multimodal semantic communication[J]. Journal on Communications, 2023, 44(5): 28–41. doi: 10.11959/j.issn.1000-436x.2023105.