移动通信2025，Vol.49Issue(5)：49-56,8.DOI:10.3969/j.issn.1006-1010.20250320-0002
面向工业互联网的恶意流量智能检测模型
Intelligent Detection of Malicious Traffic for the Industrial Internet
摘要
Abstract
Malicious attack detection in Industrial Internet systems faces critical challenges including heterogeneous protocol diversity,complex traffic characteristics,and the evolving stealthiness of emerging attack patterns.This work conducts an in-depth analysis of multi-dimensional temporal and frequency-domain characteristics inherent to terminal-side traffic in Industrial Internet environments.To address these challenges,we develop an intelligent detection model tailored for terminal-side deployment,incorporating a novel traffic feature extraction module to achieve precise and efficient malicious traffic identification.The proposed model has been comprehensively evaluated through deployment in real-world Industrial Internet infrastructures.Experimental results demonstrate significant improvements in malicious traffic detection accuracy compared to conventional approaches.关键词
工业互联网/恶意攻击检测/流量特征提取/TransformerKey words
industrial internet/malicious attack detection/traffic feature extraction/Transformer分类
信息技术与安全科学引用本文复制引用
何承润,闫皓楠,侯志青,李超豪,周少鹏,王星,王滨..面向工业互联网的恶意流量智能检测模型[J].移动通信,2025,49(5):49-56,8.基金项目
国家重点研发计划"大规模异构物联网威胁可控捕获与分析技术"(2022YFB3104104) （2022YFB3104104）
国家自然科学基金资助项目"多模态智能模型安全保护关键技术研究"(62402373) （62402373）
中国博士后科学基金资助项目"智能物联中多模态数据安全关键技术"(2024M764269) （2024M764269）
浙江省博士后科研择优资助项目"面向图像处理场景中的智能模型安全与隐私保护关键技术研究"(ZJ2024009) （ZJ2024009）
