A Lightweight Intrusion Detection Method for UAV Networks Based on Hybrid Knowledge Distillation

ZHANG Lizhe; ZHANG Pengyu; WU Zhijun

doi:10.11959/j.issn.1000-0801.DXKX260281

Chinese

您当前的位置：

首页 >

文章列表页 >

A Lightweight Intrusion Detection Method for UAV Networks Based on Hybrid Knowledge Distillation

更新时间：2026-09-22

- A Lightweight Intrusion Detection Method for UAV Networks Based on Hybrid Knowledge Distillation
- Telecommunications Science (2026)
- 作者机构：
  
  中国民航大学安全科学与工程学院，天津东丽 300300
- 作者简介：
- 基金信息：
  
  National Natural Science Foundation of China(62571531)
- DOI：10.11959/j.issn.1000-0801.DXKX260281
  CLC： TP393
- Received：30 April 2026，
  
  Revised：2026-08-09，
  
  Accepted：25 August 2026，
- 稿件说明：
移动端阅览
ZHANG Lizhe, ZHANG Pengyu, WU Zhijun. A Lightweight Intrusion Detection Method for UAV Networks Based on Hybrid Knowledge Distillation[J/OL]. Telecommunications Science, 2026.
DOI：

ZHANG Lizhe, ZHANG Pengyu, WU Zhijun. A Lightweight Intrusion Detection Method for UAV Networks Based on Hybrid Knowledge Distillation[J/OL]. Telecommunications Science, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260281.

摘要

随着低空经济的快速发展，无人机网络面临的安全威胁日益增多，传统的入侵检测方法难以直接部署到无人机端。针对这一问题提出了一种基于混合知识蒸馏的无人机网络轻量级入侵检测方法。该方法采用双向门控循环单元与Transformer串行结合的方式，构建高性能教师模型，然后使用混合知识蒸馏策略，获得具有高检测性能的无人机网络轻量级入侵检测学生模型。在UAV-NIDD数据集上的实验结果表明，学生模型在参数量和模型大小相较于教师模型均降低约98%的情况下，仍取得了0.9616的宏平均F1值，并在树莓派端实现了1.198ms的推理时延。

Abstract

With the rapid development of the low-altitude economy

security threats facing Unmanned Aerial Vehicle (UAV) networks have increased

and traditional intrusion detection methods are difficult to deploy directly on UAV terminals. To address this problem

a lightweight intrusion detection method for UAV networks based on hybrid knowledge distillation was proposed. This method adopted the serial combination of Bidirectional Gated Recurrent Unit and Transformer to construct a high-performance teacher model

and then used a hybrid knowledge distillation strategy to obtain a lightweight intrusion detection student model for UAV networks with high detection performance. Experimental results on the UAV-NIDD dataset show that

while the number of parameters and model size of the student model are both reduced by approximately 98% compared with those of the teacher model

the student model still achieves a macro-average F1 score of 0.9616 and an inference latency of only 1.198 ms on the Raspberry Pi.

关键词

Keywords

references

Adil M , Jan M A , Liu Y , et al . A systematic survey: Security threats to UAV-aided IoT applications, taxonomy, current challenges and requirements with future research directions [J ] . IEEE Transactions on Intelligent Transportation Systems , 2023 , 24 ( 2 ): 1437 - 1455 .

杜雪茹 , 王迪 , 曹越 , 等 . 无人机网络入侵检测与信任体系研究综述 [J ] . 武汉大学学报(理学版) , 2023 , 69 ( 05 ): 665 - 676 .

DU X R , WANG D , CAO Y , et al . Survey on intrusion detection and trust system for UAV networks [J ] . Journal of Wuhan University (Natural Science Edition) , 2023 , 69 ( 05 ): 665 - 676 .

Imrana Y , Xiang Y , Ali L , et al . A bidirectional LSTM deep learning approach for intrusion detection [J ] . Expert Systems with Applications , 2021 , 185 : 115524 .

Sharma B , Sharma L , Lal C , et al . Explainable artificial intelligence for intrusion detection in IoT networks: A deep learning based approach [J ] . Expert Systems with Applications , 2024 , 238 : 121751 .

Ihekoronye V U , Ajakwe S O , Lee J M , et al . Droneguard: An explainable and efficient machine learning framework for intrusion detection in drone networks [J ] . IEEE Internet of Things Journal , 2025 , 12 ( 7 ): 7708 - 7722 .

Nadiah A L N , Alamri A , Aljuhani A , et al . Transformer-based knowledge distillation for explainable intrusion detection system [J ] . Computers & Security , 2025 , 154 : 104417 .

Alzahrani A . Novel approach for intrusion detection attacks on small drones using ConvLSTM model [J ] . IEEE Access , 2024 , 12 : 149238 - 149253 .

庞衍硕 , 魏亿萍 , 单宝明 , 等 . 基于1DCNN和BiSRU的工控网络入侵检测方法 [J ] . 计算机应用与软件 , 2025 , 42 ( 12 ): 385 - 392 .

PANG Y S , WEI Y P , SHAN B M , et al . Intrusion detection method for industrial control networks based on 1DCNN and BiSRU [J ] . Computer Applications and Software , 2025 , 42 ( 12 ): 385 - 392 .

Ramadan R A , Emara A H , Al-Sarem M , et al . Internet of drones intrusion detection using deep learning [J ] . Electronics , 2021 , 10 ( 21 ): 2633 .

向思羽 , 刘才铭 . 结合混合特征选择和Transformer的网络数据流异常检测 [J ] . 电子科技大学学报 , 2025 , 54 ( 03 ): 442 - 454 .

XIANG S Y , LIU C M . Network data flow anomaly detection combining hybrid feature selection and Transformer [J ] . Journal of University of Electronic Science and Technology of China , 2025 , 54 ( 03 ): 442 - 454 .

DONG S , ZHANG T , HOU R . TTM: An intrusion detection system based on transformer and LSTM [J ] . IET Communications , 2026 , 20 ( 1 ): e70181 .

廖金菊 , 丁嘉伟 , 冯光辉 . 基于时间卷积网络的无监督入侵检测模型 [J ] . 电信科学 , 2025 , 41 ( 01 ): 164 - 173 .

LIAO J J , DING J W , FENG G H . Unsupervised intrusion detection model based on temporal convolutional network [J ] . Telecommunications Science , 2025 , 41 ( 01 ): 164 - 173 .

Mohammed A B , Fourati L C , Fakhrudeen A M . Comprehensive systematic review of intelligent approaches in UAV-based intrusion detection, blockchain, and network security [J ] . Computer Networks , 2024 , 239 : 110140 .

张锐 . 面向车联网的基于卷积神经网络的入侵检测模型 [J ] . 电信科学 , 2024 , 40 ( 12 ): 51 - 62 .

ZHANG R . Intrusion detection model based on convolutional neural network for Internet of Vehicles [J ] . Telecommunications Science , 2024 , 40 ( 12 ): 51 - 62 .

Gou J , Yu B , Maybank S J , et al . Knowledge distillation: A survey [J ] . International Journal of Computer Vision , 2021 , 129 ( 6 ): 1789 - 1819 .

Sun L , Liu Z , Wan L , et al . Cooperative knowledge-distillation-based tiny DNN for UAV-assisted mobile-edge network [J ] . IEEE Internet of Things Journal , 2024 , 11 ( 18 ): 30204 - 30216 .

Hadi H J , Cao Y , Khan M K , et al . UAV-NIDD: A dynamic dataset for cybersecurity and intrusion detection in UAV networks [J ] . IEEE Transactions on Network Science and Engineering , 2025 , 12 ( 4 ): 2739 - 2757 .

Yang S , Zheng X , Xu Z , et al . A lightweight approach for network intrusion detection based on self-knowledge distillation [C ] // ICC 2023-IEEE International Conference on Communications . Rome, Italy : IEEE , 2023 : 3000 - 3005 .

Wang X , Wang Z , Wang E , et al . Spatial-temporal knowledge distillation for lightweight network traffic anomaly detection [J ] . Computers & Security , 2024 , 137 : 103636 .

Wisanwanichthan T , Thammawichai M . A lightweight intrusion detection system for IoT and UAV using deep neural networks with knowledge distillation [J ] . Computers , 2025 , 14 ( 7 ): 291 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

A few-shot network intrusion detection method based on cross-attention fusion for IIoT

Cross-domain federated intrusion detection model under heterogeneous data distribution

Dynamic integrated intrusion detection model based on powershap and hybrid sampling

A modulation recognition model driven by lightweight knowledge distillation in the complex number domain

Attack detection model based on stacking ensemble learning for Internet of vehicles

Related Author

Yang Jun

Zhou Yiyang

Li Jinlong

Liang Xuecai

Shi Bo

Xu Congyuan

Yang Yun

Zhang Shangzhe

Related Institution

College of Artificial Intelligence, Jiaxing University

China United Network Communications Co., Ltd. Harbin Branch

College of Mathematics and Physics, Shanghai University of Electric Power

School of Information, Shanghai Ocean University

College of Electronic and Information Engineering, Shanghai University of Electric Power

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