Scholarlink

The Internet of Vehicles (IoV) is transforming transportation by improving safety, efficiency, and connectivity. However, this digital integration also introduces challenges like misbehaviour and new types of cyber-attacks. Many current detection methods depend on machine learning models that require prior knowledge of attack patterns to work effectively. To overcome this limitation, we have developed a powerful Misbehaviour Detection System and Novel Attack Detection System specifically for the IoV using Random Forest, Autoencoder and Graph Attention Network. This system can detect new types of attacks without needing prior exposure to their patterns. Our analysis shows its strong ability to identify these unknown threats, achieving a detection rate of 96.79% for novel attacks, even without prior information.

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

Misbehaviour Detection System and Novel Attack Detection Using Deep Learning in Internet of Vehicles

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Lakshana,
  • G. Abirami,
  • S. Nithiyashri,
  • V. Vetriselvi,
  • R. Aishwarya,
  • A. Aruna Gladys

摘要

The Internet of Vehicles (IoV) is transforming transportation by improving safety, efficiency, and connectivity. However, this digital integration also introduces challenges like misbehaviour and new types of cyber-attacks. Many current detection methods depend on machine learning models that require prior knowledge of attack patterns to work effectively. To overcome this limitation, we have developed a powerful Misbehaviour Detection System and Novel Attack Detection System specifically for the IoV using Random Forest, Autoencoder and Graph Attention Network. This system can detect new types of attacks without needing prior exposure to their patterns. Our analysis shows its strong ability to identify these unknown threats, achieving a detection rate of 96.79% for novel attacks, even without prior information.