基于联邦强化学习的面向边缘网络的入侵检测方法研究  

Research on intrusion detection method for edge networks based on federated reinforcement learning

作　　者：丁凯 黄宜都 陶铭 谢仁平 DING Kai;HUANG Yidu;TAO Ming;XIE Renping(School of Computer Science and Technology,Dongguan University of Technology,Dongguan 523808,China;Guangdong Laboratory of Artificial Intelligence and Digital Economy(Shenzhen),Shenzhen 518107,China)

机构地区：[1]东莞理工学院计算机科学与技术学院,广东东莞523808 [2]人工智能与数字经济广东省实验室(深圳),广东深圳518107

出　　处：《物联网学报》2024年第4期140-155,共16页Chinese Journal on Internet of Things

基　　金：国家自然科学基金资助项目(No.62001113);广东省基础与应用基础研究基金资助项目(No.2021A1515010656);广东省人工智能与数字经济实验室(深圳)开放研究基金资助项目(No.GML-KF-22-02);东莞市社会发展科技项目(No.20221800902472)。

摘　　要：随着物联网(IoT,Internet of things)设备的迅速普及,针对IoT设备的攻击频率和强度不断上升,因而持续更新安全机制以保障物联网设备的安全显得尤为重要。然而,随着公众隐私意识的增强,越来越多的数据集不再对外共享,形成数据“孤岛”现象,阻碍了物联网安全防护能力的提升。为了解决这一问题,提出了一种基于联邦强化学习的入侵检测方法,并通过医疗物联网(IoMT,Internet of medical things)和车联网(IoV,Internet of vehicles)场景下的两个数据集进行实验验证。为模拟真实环境,在每个边缘代理中设计了不平衡的流量样本分布,进而评估全局模型的检测精度和鲁棒性。采用双深度Q网络(DDQN,double deep Q-network)为边缘代理的强化学习框架,并通过准确率、精确率、召回率和F1分数对实验结果进行评估。实验结果表明,提出的方法具有良好的鲁棒性和检测精度。With the rapid proliferation of Internet of things(IoT)devices,the frequency and intensity of attacks targeting these devices are constantly increasing.Therefore,it’s quite important that security mechanisms are continuously updated to ensure the safety of IoT devices.However,as public awareness of privacy grows,many datasets are no longer shared,leading to the emergence of data silos,which hinders the improvement of IoT security.To address this issue,a federated reinforcement learning-based intrusion detection method was proposed,and experiments were conducted using two datasets from the Internet of medical things(IoMT)and Internet of vehicles(IoV)scenarios.Imbalanced traffic sample distributions were designed for each edge agent to simulate a real-world environment,allowing for the evaluation of the detection accuracy and robustness of the global model.Double deep Q-network(DDQN)was employed as the reinforcement learning framework for the edge agents,and the experimental results were evaluated using accuracy,precision,recall,and F1-score.The results demonstrate that the proposed method exhibits strong robustness and detection accuracy.

关 键 词：联邦强化学习 入侵检测 物联网安全 物联网 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]