基于层次联邦与动态权重的卫星网络异常检测方法

doi:10.19678/j.issn.1000-3428.0069214

摘要/Abstract

摘要：

随着全球卫星网络的发展与覆盖, 卫星网络攻击的影响逐渐扩大, 开展卫星网络攻击或异常行为检测研究, 对保障空天地一体化网络安全至关重要。针对联邦学习(FL)在卫星网络分布式异常检测中存在通信连接不稳定、聚合时间过长的问题, 提出一种基于层次FL与改进动态半异步聚合算法的卫星网络异常检测方法。首先, 融合层次FL进行分布式异常检测模型训练与分层聚合, 提出一种3层架构的空天地协同异常检测框架, 将高空通信平台站(HAPS)作为边缘计算层用于局部聚合, 解决星地通信范围有限的问题; 然后, 针对FL在卫星网络中存在的通信连接不稳定、聚合时间过长的问题, 提出并实现一种改进动态半异步联邦聚合算法, 结合卫星状态信息构造改进聚合时机序列和动态权重, 缓解由于星地通信间歇性而导致的聚合时间过长、模型精度受影响的问题。在STIN和CIC-IDS-2017数据集上的对比实验结果表明, 该方法在保证良好的异常检测模型性能的同时, 有效提高了卫星网络分布式异常检测中的FL聚合效率。

关键词: 层次联邦学习, 动态权重, 卫星网络, 异常检测, 近地轨道

Abstract:

Satellite network attacks have gradually increased with the development and widespread coverage of global satellite networks. Research on satellite network attacks or anomaly behavior detection is paramount for ensuring the security of space-air-ground integrated networks. Considering the issues of unstable communication and prolonged aggregation time in the application of Federated Learning (FL) for distributed anomaly detection in satellite networks, this paper proposes a satellite network anomaly detection method based on hierarchical FL and an improved dynamic semi-asynchronous aggregation algorithm. First, by integrating the hierarchical FL for distributed anomaly detection model training and hierarchical aggregation, a three-tier architecture for a space-air-ground collaborative anomaly detection framework is proposed. The proposed framework utilizes High-Altitude Platform Stations (HAPS) as an edge computing layer for local aggregation, addressing the limited satellite-ground communication range. Subsequently, to address unstable communication and prolonged aggregation times in FL within satellite networks, an improved dynamic semi-asynchronous federated aggregation algorithm is proposed and implemented. This algorithm constructs an improved aggregation timing sequence and dynamic weights by incorporating satellite status information, thereby mitigating the issues of prolonged aggregation time and compromised model accuracy caused by intermittent satellite-ground communication. Comparative experiments on the STIN and CIC-IDS-2017 datasets reveal that the proposed method effectively enhances FL aggregation efficiency in distributed anomaly detection for satellite networks while maintaining the excellent performance of the anomaly detection model.

Key words: hierarchical Federated Learning (FL), dynamic weights, satellite network, anomaly detection, Low Earth Orbit (LEO)

牛渲文, 杜晔, 杨明松, 李昂, 黎妹红. 基于层次联邦与动态权重的卫星网络异常检测方法[J]. 计算机工程, 2025, 51(12): 210-220.

NIU Xuanwen, DU Ye, YANG Mingsong, LI Ang, LI Meihong. Anomaly Detection Method for Satellite Networks Based on Hierarchical Federation and Dynamic Weights[J]. Computer Engineering, 2025, 51(12): 210-220.

https://www.ecice06.com/CN/Y2025/V51/I12/210

图/表 19

图1 空天地协同异常检测框架层次结构

Fig.1 Hierarchical structure of the space-air-ground collaborative anomaly detection framework

图2 模型训练和检测流程

Fig.2 Model training and detection process

图3 聚合时机序列

Fig.3 Aggregation timing sequence

图4 卫星网络中采用固定聚合时机序列的示意图

Fig.4 Schematic diagram of fixed aggregation timing sequence in satellite network

图5 卫星网络中采用改进聚合时机序列的示意图

Fig.5 Schematic diagram of improved aggregation timing sequence in satellite network

图6 Walker-Delta星座与HAPS结构

Fig.6 Walker-Delta constellation and HAPS structure

图7 Walker-Delta星座卫星轨道周期

Fig.7 Orbital period of the Walker-Delta constellation satellite

图8 HAPS与卫星连接全覆盖示意图

Fig.8 Schematic diagram of full coverage for HAPS and satellite connections

图9 基于STIN数据集的异常检测混淆矩阵

Fig.9 Anomaly detection confusion matrix based on STIN dataset

图10 基于CIC-IDS-2017数据集的异常检测混淆矩阵

Fig.10 Anomaly detection confusion matrix based on CIC-IDS-2017 dataset

图11 STIN数据集上不同FL算法的准确率曲线

Fig.11 Accuracy curves of different FL algorithms on STIN dataset

图12 CIC-IDS-2017数据集上不同FL算法的准确率曲线

Fig.12 Accuracy curves of different FL algorithms on CIC-IDS-2017 dataset

参考文献 29

1	KODHELI O , LAGUNAS E , MATURO N , et al. Satellite communications in the new space era: a survey and future challenges. IEEE Communications Surveys & Tutorials, 2021, 23 (1): 70- 109.
2	GUO H Z , LI J Y , LIU J J , et al. A survey on space-air-ground-sea integrated network security in 6G. IEEE Communications Surveys & Tutorials, 2022, 24 (1): 53- 87.
3	YAO H P , WANG L Y , WANG X D , et al. The space-terrestrial integrated network: an overview. IEEE Communications Magazine, 2018, 56 (9): 178- 185. doi: 10.1109/MCOM.2018.1700038
4	蒋长林, 李清, 王羽, 等. 天地一体化网络关键技术研究综述. 软件学报, 2024, 35 (1): 266- 287.
	JIANG C L , LI Q , WANG Y , et al. Survey on key technologies in space-ground integrated network. Journal of Software, 2024, 35 (1): 266- 287.
5	CAO H, WU L L, CHEN Y, et al. Analysis on the security of satellite Internet[EB/OL]. [2023-10-05]. https://link.springer.com/chapter/10.1007/978-981-33-4922-3_14.
6	ASHRAF I , NARRA M , UMER M , et al. A deep learning-based smart framework for cyber-physical and satellite system security threats detection. Electronics, 2022, 11 (4): 667. doi: 10.3390/electronics11040667
7	AGRAWAL S , SARKAR S , AOUEDI O , et al. Federated learning for intrusion detection system: concepts, challenges and future directions. Computer Communications, 2022, 195, 346- 361. doi: 10.1016/j.comcom.2022.09.012
8	ZHAO J Y, ZOU J X, LI Z P, et al. High Altitude Platform (HAP) communication system based on satellite communication theory[C]//Proceedings of the 22nd IEEE International Conference on Communication Technology (ICCT). Washington D.C., USA: IEEE Press, 2022: 446-450.
9	GUO W , XU J , PEI Y K , et al. A distributed collaborative entrance defense framework against DDoS attacks on satellite Internet. IEEE Internet of Things Journal, 2022, 9 (17): 15497- 15510. doi: 10.1109/JIOT.2022.3176121
10	CAO S , DANG S X , ZHANG Y , et al. A blockchain-based access control and intrusion detection framework for satellite communication systems. Computer Communications, 2021, 172, 216- 225. doi: 10.1016/j.comcom.2021.03.023
11	LAVAUR L , PAHL M O , BUSNEL Y , et al. The evolution of federated learning-based intrusion detection and mitigation: a survey. IEEE Transactions on Network and Service Management, 2022, 19 (3): 2309- 2332. doi: 10.1109/TNSM.2022.3177512
12	刘金硕, 詹岱依, 邓娟, 等. 基于深度神经网络和联邦学习的网络入侵检测. 计算机工程, 2023, 49 (1): 15-21, 30. doi: 10.19678/j.issn.1000-3428.0065155
	LIU J S , ZHAN D Y , DENG J , et al. Network intrusion detection based on deep neural network and federated learning. Computer Engineering, 2023, 49 (1): 15-21, 30. doi: 10.19678/j.issn.1000-3428.0065155
13	LI K , ZHOU H C , TU Z , et al. Distributed network intrusion detection system in satellite-terrestrial integrated networks using federated learning. IEEE Access, 2020, 8, 214852- 214865. doi: 10.1109/ACCESS.2020.3041641
14	MOUSTAFA N , KHAN I A , HASSANIN M , et al. DFSat: deep federated learning for identifying cyber threats in IoT-based satellite networks. IEEE Transactions on Industrial Informatics, 2024, 8, 1- 8.
15	SALIM S , MOUSTAFA N , HASSANIAN M , et al. Deep-federated-learning-based threat detection model for extreme satellite communications. IEEE Internet of Things Journal, 2024, 11 (3): 3853- 3867. doi: 10.1109/JIOT.2023.3301626
16	RAZMI N, MATTHIESEN B, DEKORSY A, et al. On-board federated learning for dense LEO constellations[C]//Proceedings of the IEEE International Conference on Communications. Washington D.C., USA: IEEE Press, 2022: 4715-4720.
17	ELMAHALLAWY M, LUO T, IBRAHEM M I. Secure and efficient federated learning in LEO constellations using decentralized key generation and on-orbit model aggregation[C]//Proceedings of the 2023 IEEE Global Communications Conference. Washington D.C., USA: IEEE Press, 2023: 5727-5732.
18	SO J, HSIEH K, ARZANI B, et al. FedSpace: an efficient federated learning framework at satellites and ground stations[EB/OL]. [2023-12-21]. https://doi.org/10.48550/arXiv.2202.01267.
19	MCMAHAN B, MOORE E, RAMAGE D, et al. Communication-efficient learning of deep networks from decentralized data[EB/OL]. [2023-12-21]. https://proceedings.mlr.press/v54/mcmahan17a/mcmahan17a.pdf.
20	LIU L M, ZHANG J, SONG S H, et al. Client-edge-cloud hierarchical federated learning[C]//Proceedings of the 2020 IEEE International Conference on Communications (ICC). Washington D.C., USA: IEEE Press, 2020: 1-6.
21	MEI C L, GAO C, XING Y X, et al. An energy consumption minimization optimization scheme for HAP-satellites edge computing[C]//Proceedings of the 22nd IEEE International Conference on Communication Technology (ICCT). Washington D.C., USA: IEEE Press, 2022: 857-862.
22	LI T , SAHU A K , ZAHEER M , et al. Federated optimization in heterogeneous networks. Proceedings of Machine Learning and Systems, 2020, 2, 429- 450.
23	CHEN Y , SUN X Y , JIN Y C . Communication-efficient federated deep learning with layerwise asynchronous model update and temporally weighted aggregation. IEEE Transactions on Neural Networks and Learning Systems, 2020, 31 (10): 4229- 4238.
24	XIE C, KOYEJO S, GUPTA I. Asynchronous federated optimization[EB/OL]. [2023-12-21]. https://arxiv.org/abs/1903.03934.
25	ZHANG H, WU T T, CHENG S Y, et al. Aperiodic local SGD: beyond local SGD[C]//Proceedings of the 51st International Conference on Parallel Processing. Washington D.C., USA: IEEE Press, 2022: 1-10.
26	张泰江, 李勇军, 赵尚弘. 基于GEO/LEO双层卫星网络的路由算法优化设计. 计算机工程, 2020, 46 (7): 198- 205. doi: 10.19678/j.issn.1000-3428.0055504
	ZHANG T J , LI Y J , ZHAO S H . Optimization design of routing algorithm based on GEO/LEO double-layer satellite network. Computer Engineering, 2020, 46 (7): 198- 205. doi: 10.19678/j.issn.1000-3428.0055504
27	VEMURI S S , DAPPURI B . Walker-Delta constellation design for LEO-based navigation using small satellites. IET Conference Proceedings, 2022, 2022 (1): 250- 253. doi: 10.1049/icp.2022.0627
28	郑鹏飞, 陈宏宇, 郭崇滨. 低轨巨型链形星座解析设计及效能分析. 西北工业大学学报, 2022, 40 (1): 148- 157.
	ZHENG P F , CHEN H Y , GUO C B . Analytical design and performance analysis of LEO mega chain satellite constellation. Journal of Northwestern Polytechnical University, 2022, 40 (1): 148- 157.
29	DAKIC K, CHAN C C, AL HOMSSI B, et al. On delay performance in mega satellite networks with inter-satellite links[C]//Proceedings of the 2023 IEEE Global Communications Conference. Washington D.C., USA: IEEE Press, 2023: 4896-4901.

[1]	吴杰辉, 柳毅. 面向多元时间序列的联合优化异常检测模型[J]. 计算机工程, 2025, 51(9): 166-176.
[2]	徐式芃, 王雷, 盛捷. 基于知识图谱的异常个体提前识别模型研究[J]. 计算机工程, 2025, 51(9): 59-70.
[3]	苗茹, 李祎, 周珂, 张俨娜, 常然然, 孟更. 一种改进的Faster R-CNN遥感图像多目标检测模型研究[J]. 计算机工程, 2025, 51(8): 292-304.
[4]	王光明, 李冬青, 蒋从锋. 不平衡数据集下的数据中心网络流量异常检测[J]. 计算机工程, 2025, 51(8): 227-237.
[5]	王亚, 甘青松, 沈琦, 宋余庆, 刘毅, 韩凯, 刘哲. 基于动态联合加权的带钢表面缺陷分类方法[J]. 计算机工程, 2025, 51(6): 286-296.
[6]	蓝章礼, 邢彩卓, 张洪. 基于改进YOLOv5s的被遮挡交通标志检测算法[J]. 计算机工程, 2025, 51(5): 361-369.
[7]	莫定涛, 俱莹, 李文进, 张亚生, 何辞, 董飞虎. 面向卫星网络的流量分类方法研究[J]. 计算机工程, 2025, 51(5): 9-19.
[8]	张安勤, 丁志锋. 融合动态图嵌入和Transformer自编码器的网络异常检测[J]. 计算机工程, 2025, 51(4): 47-56.
[9]	魏德宾, 乔维维, 张怡. 基于麻雀搜索算法的软件定义卫星网络控制器部署[J]. 计算机工程, 2025, 51(3): 172-179.
[10]	邸钦渤, 陈劭力, 时良仁. GRD: 基于GNN和扩散模型的多变量时序数据异常检测算法[J]. 计算机工程, 2025, 51(11): 35-44.
[11]	周炫余, 吴莲华, 郑勤华, 肖天星, 王紫璇, 张思敏. 联合语义提示和记忆增强的弱监督跳绳视频异常检测方法[J]. 计算机工程, 2024, 50(7): 87-95.
[12]	李永飞, 李铭洋, 常鑫, 曹可欣. 基于可解释性深度学习的物联网水质监测数据异常检测[J]. 计算机工程, 2024, 50(6): 179-187.
[13]	赵博超, 马嘉骏, 崔磊, 栾文鹏, 朱静. 基于改进VMD-XGBoost-BiLSTM组合模型的光伏发电异常检测[J]. 计算机工程, 2024, 50(3): 306-316.
[14]	王俊, 赖会霞, 万玥, 张仕. 基于角度的图神经网络高维数据异常检测方法[J]. 计算机工程, 2024, 50(3): 156-165.
[15]	宋航, 周凤, 熊伟. 基于自相关-变分对抗学习的物理系统异常检测[J]. 计算机工程, 2024, 50(12): 358-366.

选择文件类型/文献管理软件名称

选择包含的内容