Dynamic compression method for federated learning based on feature information in ISAC networks

Deng Bingguang; Peng Jiayin

doi:10.11959/j.issn.1000-436x.2026082

Chinese

您当前的位置：

首页 >

文章列表页 >

Dynamic compression method for federated learning based on feature information in ISAC networks

Papers | 更新时间：2026-05-07

- Dynamic compression method for federated learning based on feature information in ISAC networks
- Journal on Communications Vol. 47, Issue 4, Pages: 214-229(2026)
- 作者机构：
  
  重庆邮电大学通信与信息工程学院，重庆 400065
- 作者简介：
- 基金信息：
  
  Chongqing Natural Science Foundation Innovation and Development Joint Fund (China Star Network)(CSTB2023NSCQ-LZX0114)
- DOI：10.11959/j.issn.1000-436x.2026082
  CLC： TN929.5
- Received：04 January 2026，
  
  Revised：2026-03-20，
  
  Accepted：26 March 2026，
  
  Published：20 April 2026
- 稿件说明：
移动端阅览
邓炳光,彭佳音.ISAC网络中基于特征信息的联邦学习动态压缩方法[J].通信学报,2026,47(04):214-229.

Deng Bingguang,Peng Jiayin.Dynamic compression method for federated learning based on feature information in ISAC networks[J].Journal on Communications,2026,47(04):214-229.
邓炳光,彭佳音.ISAC网络中基于特征信息的联邦学习动态压缩方法[J].通信学报,2026,47(04):214-229. DOI： 10.11959/j.issn.1000-436x.2026082.

Deng Bingguang,Peng Jiayin.Dynamic compression method for federated learning based on feature information in ISAC networks[J].Journal on Communications,2026,47(04):214-229. DOI： 10.11959/j.issn.1000-436x.2026082.

摘要

针对6G通感一体化（ISAC）边缘网络中联邦学习（FL）在通信带宽受限、链路动态变化及设备异构条件下的通信效率问题，构建了一种面向ISAC场景的特征信息驱动压缩与聚合协同框架（FI-ISAC-FL）。不同于现有压缩方法，联邦更新压缩与聚合决策是为实现任务相关特征与系统状态之间的协同适配而联合设计。具体来说，设计Federated-AE在线训练机制对局部更新进行在线特征压缩，实现对高维更新的特征提取与高效传输，增强压缩表示对数据变化的适应能力。同时，结合ISAC场景下的信道状态与数据分布特征，构建状态感知的自适应聚合策略，改善通信受限条件下的训练稳定性与效率。仿真结果表明，在雷达目标识别任务中，所提方法在固定通信预算下能够取得较好的模型性能，在性能与资源开销之间实现更优折中，并对强感知干扰与Non-IID数据分布表现出较强的鲁棒性。

Abstract

To address the communication-efficiency problem of federated learning (FL) in 6G integrated sensing and communication (ISAC) edge networks under limited communication bandwidth

dynamic link conditions

and device heterogeneity

a feature-information-driven collaborative compression and aggregation framework

termed FI-ISAC-FL

was developed. Unlike existing compression methods

federated update compression and aggregation were jointly designed for coordinated adaptation between task-relevant features and system states. Specifically

a Federated-AE online training mechanism was introduced to compress local updates online

by which high-dimensional updates were effectively represented and efficiently transmitted

while the adaptability of compressed representations to data variations was enhanced. In addition

a state-aware adaptive aggregation strategy was developed by incorporating channel states and data-distribution characteristics in ISAC scenarios to improve training stability and efficiency under communication constraints. Simulation results show that

in a radar target recognition task

favorable model performance is achieved under a fixed communication budget

a better trade-off between performance and resource overhead is attained

and strong robustness against severe sensing interference and Non-IID data distributions is demonstrated.

关键词

Keywords

references

Letaief K B , Shi Y M , Lu J M , et al . Edge artificial intelligence for 6G: vision, enabling technologies, and applications [J ] . IEEE Journal on Selected Areas in Communications , 2022 , 40 ( 1 ): 5 - 36 .

Tan D K P , He J , Li Y C , et al . Integrated sensing and communication in 6G: motivations, use cases, requirements, challenges and future directions [C ] // Proceedings of the 2021 1st IEEE International Online Symposium on Joint Communications & Sensing (JC&S) . Piscataway : IEEE Press , 2021 : 1 - 6 .

Zhou Y , An Q C , Wang Z B , et al . Integrated sensing, computation, and communication enabled federated edge learning [J ] . IEEE Transactions on Wireless Communications , 2026 , 25 : 7117 - 7131 .

Wen D Z , Zhou Y , Li X Y , et al . A survey on integrated sensing, communication, and computation [J ] . IEEE Communications Surveys & Tutorials , 2025 , 27 ( 5 ): 3058 - 3098 .

Mcmahan H B , Moore E , Ramage D , et al . Communication-efficient learning of deep networks from decentralized data [PP ] . V4 . ( 2016-02-17 )[ 2026-01-04 ] . arXiv: arXiv. 1806 . 00582 .

Lim W Y B , Luong N C , Hoang D T , et al . Federated learning in mobile edge networks: a comprehensive survey [J ] . IEEE Communications Surveys & Tutorials , 2020 , 22 ( 3 ): 2031 - 2063 .

Alistarh D , Grubic D , Li J , et al . QSGD: communication-efficient SGD via gradient quantization and encoding [PP ] . V4 . ( 2016-10-07 )[ 2026-01-04 ] . arXiv: arXiv. 1610 . 02132 .

Lin Y J , Han S , Mao H Z , et al . Deep gradient compression: reducing the communication bandwidth for distributed training [PP ] . V3 . ( 2017-12-05 )[ 2026-01-04 ] . arXiv: arXiv. 1712 . 01887 .

Zhuansun Y , Li D D , Huang X H , et al . Communication-efficient federated learning with adaptive compression under dynamic bandwidth [PP ] . V1 . ( 2024-05-06 )[ 2026-01-04 ] . arXiv: arXiv. 2405 . 03248 .

Beitollahi M , Lu N . FLAC: federated learning with autoencoder compression and convergence guarantee [C ] // Proceedings of the GLOBECOM 2022 - 2022 IEEE Global Communications Conference . Piscataway : IEEE Press , 2022 : 4589 - 4594 .

Gündüz D , Qin Z J , Aguerri I E , et al . Beyond transmitting bits: context, semantics, and task-oriented communications [J ] . IEEE Journal on Selected Areas in Communications , 2023 , 41 ( 1 ): 5 - 41 .

Wang H , Li H R , Chen H M , et al . FedSC: federated learning with semantic-aware collaboration [C ] // Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining . New York : ACM Press , 2025 : 2938 - 2949 .

Dong X N , Zeng J , Wen J H , et al . SFL: a semantic-based federated learning method for POI recommendation [J ] . Information Sciences , 2024 , 679 : 121057 .

马千飘 , 贾庆民 , 刘建春 , 等 . 异构边缘计算环境下异步联邦学习的节点分组与分时调度策略 [J ] . 通信学报 , 2023 , 44 ( 11 ): 79 - 93 .

Ma Q P , Jia Q M , Liu J C , et al . Client grouping and time-sharing scheduling for asynchronous federated learning in heterogeneous edge computing environment [J ] . Journal on Communications , 2023 , 44 ( 11 ): 79 - 93 .

Wen D Z , Xie S J , Cao X W , et al . Integrated sensing, communication, and computation for over-the-air federated edge learning [J ] . IEEE Transactions on Wireless Communications , 2026 , 25 : 2748 - 2762 .

Liu P X , Zhu G X , Wang S , et al . Toward ambient intelligence: federated edge learning with task-oriented sensing, computation, and communication integration [J ] . IEEE Journal of Selected Topics in Signal Processing , 2023 , 17 ( 1 ): 158 - 172 .

王毅 , 瞿治国 , 孙乐 . 结合自注意力机制和软阈值降噪的对比联邦学习特征聚合算法 [J ] . 重庆邮电大学学报(自然科学版) , 2024 , 36 ( 5 ): 966 - 973 .

Wang Y , Qu Z G , Sun L . Self-attention mechanism and noise reduction techniques combined with federated comparative learning feature aggregation algorithm [J ] . Journal of Chongqing University of Posts and Telecommunications (Natural Science Edition) , 2024 , 36 ( 5 ): 966 - 973 .

张晓琴 , 金西兴 , 陆艳军 , 等 . 注意力机制驱动的个性化联邦学习特征分离方法 [J ] . 计算机应用研究 , 2025 , 42 ( 4 ): 1102 - 1107 .

Zhang X Q , Jin X X , Lu Y J , et al . Attention-driven feature separation method for personalized federated learning [J ] . Application Research of Computers , 2025 , 42 ( 4 ): 1102 - 1107 .

Zhao Y , Li M , Lai L Z , et al . Federated learning with non-IID data [PP ] . V2 . ( 2022-07-21 )[ 2026-01-04 ] . arXiv: arXiv. 1806 . 00582 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

HFLGuard: clustering and autoencoder-based defense method against heterogeneous federated poisoning

DPBR-Adapt: a hierarchically adaptive differential privacy defence scheme for federated learning

Personalized differential privacy federated learning method for collaborative spectrum sensing

Related Author

LI Yongfei

FANG Chen

CHANG Chaowen

GUO Yuanbo

FU Chunhui

Hu Ronglei

Bai Chenyang

Wei Zhanzhen

Related Institution

Department of Cryptogram Engineering, Information Engineering University

School of Cyberspace Security, Hainan University

Department of Electronic and Communication Engineering, Beijing Electronic Science and Technology Institute

School of Information Engineering, Minzu University of China

School of Automation, Guangdong University of Technology

AI问答

Postal code：100079
Tel：（010）53879206 Email：tmw@bjxintong.com.cn
Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09082226号-64 京公网安备11010602201714号
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰