《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (8): 2555-2566.DOI: 10.11772/j.issn.1001-9081.2025070920

• 网络与通信 • 上一篇    下一篇

面向空天地一体化网络的个性化联邦学习智能边缘聚合方法

金亚东1,2,3, 范菁1,2,3(), 郗恩康1,2,3, 董华1,2,3, 俞浩1,2,3, 孙伊航1,2,3   

  1. 1.云南民族大学 电气信息工程学院,昆明 650504
    2.云南省无人自主系统重点实验室(云南民族大学),昆明 650504
    3.云南省高校信息与通信安全灾备重点实验室(云南民族大学),昆明 650504
  • 收稿日期:2025-08-11 修回日期:2025-10-22 接受日期:2025-10-29 发布日期:2025-11-05 出版日期:2026-08-10
  • 通讯作者: 范菁
  • 作者简介:金亚东(1998—),男,云南曲靖人,硕士研究生,CCF会员,主要研究方向:联邦学习、空天地一体化网络
    范菁(1976—),女,云南景洪人,教授,博士,CCF会员,主要研究方向:联邦学习、信息网络安全
    郗恩康(2000—),男,山东枣庄人,硕士研究生,CCF会员,主要研究方向:联邦学习
    董华(2001—),男,山西运城人,硕士研究生,CCF会员,主要研究方向:联邦学习
    俞浩(2000—),男,湖北咸宁人,硕士研究生,CCF会员,主要研究方向:联邦学习
    孙伊航(2001—),男,河南许昌人,硕士研究生,CCF会员,主要研究方向:联邦学习。
  • 基金资助:
    国家自然科学基金资助项目(12361104);教育部-新一代信息技术创新项目(2023IT077);云南省教育厅科学研究基金资助项目(2025Y0670);云南省教育厅科学研究基金资助项目(2023Y0499);CCF-深信服“远望”科研基金资助项目(20240210);云南省吴中海专家工作站项目(202305AF150045);云南民族大学2025年硕士研究生科研创新基金资助项目(2025SKY027)

Personalized federated learning intelligent edge aggregation method for space-air-ground integrated networks

Yadong JIN1,2,3, Jing FAN1,2,3(), Enkang XI1,2,3, Hua DONG1,2,3, Hao YU1,2,3, Yihang SUN1,2,3   

  1. 1.School of Electrical and Information Engineering,Yunnan Minzu University,Kunming Yunnan 650504,China
    2.Yunnan Key Laboratory of Unmanned Autonomous Systems (Yunnan Minzu University),Kunming Yunnan 650504,China
    3.Key Laboratory of Information and Communication Security and Disaster Recovery in Universities of Yunnan Province (Yunnan Minzu University),Kunming Yunnan 650504,China
  • Received:2025-08-11 Revised:2025-10-22 Accepted:2025-10-29 Online:2025-11-05 Published:2026-08-10
  • Contact: Jing FAN
  • About author:JIN Yadong, born in 1998, M. S. candidate. His research interests include federated learning, air-space-ground integrated network.
    XI Enkang, born in 2000, M. S. candidate. His research interests include federated learning.
    DONG Hua, born in 2001, M. S. candidate. His research interests include federated learning.
    YU Hao, born in 2000, M. S. candidate. His research interests include federated learning.
    SUN Yihang, born in 2001, M. S. candidate. His research interests include federated learning.
  • Supported by:
    National Natural Science Foundation of China(12361104);Ministry of Education - New Generation Information Technology Innovation Project(2023IT077);Scientific Research Fund Project of Yunnan Provincial Department of Education(2025Y0670);CCF-SANGFOR “FarSight” Research Fund(20240210);Wu Zhonghai Expert Workstation Project of Yunnan Province(202305AF150045);2025 Postgraduate Research Innovation Fund of Yunnan Minzu University(2025SKY027)

摘要:

空天地一体化网络(SAGIN)作为支撑全球数据传输的综合性网络发挥着关键作用,而它高度动态异构的网络特性导致节点各数据呈现统计异构性,具体表现为非独立同分布(Non-IID)特性。联邦学习(FL)作为分布式机器学习,利用分布式数据训练模型,而FL对SAGIN中的Non-IID数据进行模型训练时,会损害每个节点上全局模型的泛化。针对这个问题,提出个性化联邦学习智能边缘聚合(FedIEA)方法,通过个性化联邦学习(PFL)捕获节点模型的全局模型中的所需信息。FedIEA方法的核心是一个智能边缘聚合(IEA)模块,它可以智能地将下载的全局模型和边缘模型聚合至每个节点上的边缘目标,以在每次迭代中训练之前的初始化本地模型。为了评估FedIEA方法的有效性,在MNIST、FashionMNIST和CIFAR-10这3个数据集上进行对比实验,结果表明:与联邦平均(FedAvg)、FedProx(Federated Proximal)、通过对数几率校准的联邦学习(FedLC)等典型算法相比,FedIEA在迪利克雷(Dirichlet)不同异构设置下的测试准确率都高于FedAvg等基准算法,在FashionMNIST和CIFAR-10数据集上基于卷积神经网络(CNN)模型比FedAvg分别高出25.30和41.28个百分点;在3个数据集上与9个最先进的PFL算法相比,FedIEA基于CNN和深度神经网络(DNN)模型的测试准确率都具有显著的优势。

关键词: 空天地一体化网络, 联邦学习, 个性化联邦学习, 统计异构性, 边缘聚合, 随机梯度下降

Abstract:

Space-Air-Ground Integrated Network (SAGIN) serves as a comprehensive network supporting global data transmission and plays a key role. However, its highly dynamic and heterogeneous network characteristics lead to an issue that data at each node exhibits statistical heterogeneity, specifically manifests Non-Independent and Identically Distributed (Non-IID) characteristics. Federated Learning (FL), as a distributed machine learning approach, uses distributed data for model training. When FL trains models on Non-IID data in SAGIN, the generalization performance of global model at each node is damaged. To address this issue, a Personalized FL-based Intelligent Edge Aggregation (FedIEA) method was proposed, which captured necessary information from global models of node models through Personalized Federated Learning (PFL). The core of FedIEA method is an Intelligent Edge Aggregation (IEA) module, which was able to aggregate downloaded global model and the edge models into the edge-side targets at each node intelligently to train the initialized local models in each iteration. To evaluate the effectiveness of FedIEA method, comparative experiments were conducted on three datasets, namely MNIST, FashionMNIST and CIFAR-10, with typical algorithms such as Federated Averaging (FedAvg), Federated Proximal (FedProx), and Federated learning via Logits Calibration (FedLC). The results show that the test accuracies of FedIEA under different Dirichlet heterogeneous settings are higher than those of benchmark algorithms like FedAvg; and based on Convolutional Neural Network (CNN) model, the test accuracies of FedIEA surpass FedAvg by 25.30 and 41.28 percentage points on FashionMNIST and CIFAR-10 datasets, respectively; on these three datasets, FedIEA method has significant advantages in the test accuracy based on CNN and Deep Neural Network (DNN) models over 9 state-of-the-art PFL algorithms.

Key words: Space-Air-Ground Integrated Network (SAGIN), Federated Learning (FL), Personalized Federated Learning (PFL), statistical heterogeneity, edge aggregation, Stochastic Gradient Descent (SGD)

中图分类号: