Scholarlink

<p>Federated Learning (FL) has emerged as a crucial technique for collaborative model training, particularly in the domain of autonomous vehicles and Internet of Vehicles. The incorporation of FL for improved intrusion detection in the Internet of Vehicles presents a promising avenue to enhance security measures. However, successful deployment of FL in real-world scenarios necessitates addressing various technical challenges. This paper introduces SecNet-FLIDS, a pioneering model that leverages Blockchain-based FL for collaborative cyberattack detection in IoV. SecNet-FLIDS prioritizes robust privacy and security, establishing a collective defense mechanism. The proposed model architecture incorporates a novel TOP-K-based Node selection scheme to strategically enhance overall accuracy. Addressing imbalanced datasets during local training, the research implements the Synthetic Minority Over-sampling Technique (SMOTE) combined with Edited Nearest Neighbors (ENN). Context-aware Transformer networks enhance intrusion detection within the federated learning framework. Comparative analyses on the CAR-HACKING dataset demonstrate SecNet-FLIDS’s superiority over competing approaches, achieving remarkable results with an accuracy of 99.32%, precision of 99.32%, recall of 98.99%, and an F1 score of 98.65%. Furthermore, evaluations on the UNSW-NB15 dataset showcase SecNet-FLIDS’s exceptional performance with an accuracy of 99.69%, precision of 99.73%, recall of 99.70%, and an F1 score of 99.68%. These results affirm SecNet-FLIDS as a cutting-edge solution for collaborative cyberattack detection in IoV, emphasizing its prowess in ensuring privacy, security, and scalability.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Securing internet of vehicles: a blockchain-based federated learning approach for enhanced intrusion detection

  • 期刊论文

  • 发表日期：

    2025年2月25日

查看全文
  • Irshad Ullah,
  • Xiaoheng Deng,
  • Xinjun Pei,
  • Husnain Mushtaq,
  • Zia Khan

摘要

Federated Learning (FL) has emerged as a crucial technique for collaborative model training, particularly in the domain of autonomous vehicles and Internet of Vehicles. The incorporation of FL for improved intrusion detection in the Internet of Vehicles presents a promising avenue to enhance security measures. However, successful deployment of FL in real-world scenarios necessitates addressing various technical challenges. This paper introduces SecNet-FLIDS, a pioneering model that leverages Blockchain-based FL for collaborative cyberattack detection in IoV. SecNet-FLIDS prioritizes robust privacy and security, establishing a collective defense mechanism. The proposed model architecture incorporates a novel TOP-K-based Node selection scheme to strategically enhance overall accuracy. Addressing imbalanced datasets during local training, the research implements the Synthetic Minority Over-sampling Technique (SMOTE) combined with Edited Nearest Neighbors (ENN). Context-aware Transformer networks enhance intrusion detection within the federated learning framework. Comparative analyses on the CAR-HACKING dataset demonstrate SecNet-FLIDS’s superiority over competing approaches, achieving remarkable results with an accuracy of 99.32%, precision of 99.32%, recall of 98.99%, and an F1 score of 98.65%. Furthermore, evaluations on the UNSW-NB15 dataset showcase SecNet-FLIDS’s exceptional performance with an accuracy of 99.69%, precision of 99.73%, recall of 99.70%, and an F1 score of 99.68%. These results affirm SecNet-FLIDS as a cutting-edge solution for collaborative cyberattack detection in IoV, emphasizing its prowess in ensuring privacy, security, and scalability.