Scholarlink

<p>Intrusion Detection Systems (IDSs) have become a critical security concern due to the increasing number of connected vehicles and the sensitive data exchanged within Vehicular Ad-hoc Networks (VANETs). This paper presents a comprehensive and up-to-date survey of IDSs designed for VANETs, focusing on advances in machine and deep learning based techniques published between 2020 and 2025. The review analyzes hybrid IDS approaches and emerging datasets used for VANET security evaluation. A critical analysis of evaluation metrics (precision, F1-score, AUCROC) and simulation tools (SUMO, OMNeT++) underscores the gap between theoretical models and real-world validation. Furthermore, the survey identifies open issues, including energy efficiency, time complexity, and scalability in dynamic VANET environments. It explores the architecture, characteristics, security, and threats in VANETs, offering an in-depth analysis of the latest developments in intrusion detection methods. The paper highlights the challenges, emphasizing the importance of real-world validation. Additionally, suggests potential directions for future directions emphasize the integration of Software-Defined Networking (SDN), Federated Learning (FL), edge computing, and quantum machine learning (QML) to enhance resilience against adversarial attacks and ensure privacy-preserving, low-latency communication. By synthesizing advances, difficulties, and emerging paradigms, this work aims to guide researchers toward robust and adaptive IDS frameworks, ultimately contributing to safer and more intelligent transportation ecosystems.</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

Advances in machine and deep learning for intrusion detection in VANETs: a comprehensive survey

  • 期刊论文

  • 发表日期：

    2026年2月12日

查看全文
  • Islem Chouchene,
  • Hamdi Eltaief,
  • Habib Youssef

摘要

Intrusion Detection Systems (IDSs) have become a critical security concern due to the increasing number of connected vehicles and the sensitive data exchanged within Vehicular Ad-hoc Networks (VANETs). This paper presents a comprehensive and up-to-date survey of IDSs designed for VANETs, focusing on advances in machine and deep learning based techniques published between 2020 and 2025. The review analyzes hybrid IDS approaches and emerging datasets used for VANET security evaluation. A critical analysis of evaluation metrics (precision, F1-score, AUCROC) and simulation tools (SUMO, OMNeT++) underscores the gap between theoretical models and real-world validation. Furthermore, the survey identifies open issues, including energy efficiency, time complexity, and scalability in dynamic VANET environments. It explores the architecture, characteristics, security, and threats in VANETs, offering an in-depth analysis of the latest developments in intrusion detection methods. The paper highlights the challenges, emphasizing the importance of real-world validation. Additionally, suggests potential directions for future directions emphasize the integration of Software-Defined Networking (SDN), Federated Learning (FL), edge computing, and quantum machine learning (QML) to enhance resilience against adversarial attacks and ensure privacy-preserving, low-latency communication. By synthesizing advances, difficulties, and emerging paradigms, this work aims to guide researchers toward robust and adaptive IDS frameworks, ultimately contributing to safer and more intelligent transportation ecosystems.