Scholarlink

<p>Vehicular Ad Hoc Networks (VANETs) are increasingly vulnerable to black-hole attacks that disrupt routing, waste energy, and compromise data reliability. Existing detection methods lack adaptability, struggle with high-dimensional traffic features, and consume excessive energy in resource-constrained environments. This study proposes the Sooty Tern Optimization-Driven Recurrent Network (STO-DRN) for accurate and energy-efficient black-hole attack detection in VANETs. The VANET Routing Attack Detection Dataset (V-RADD) containing 3,000 network traffic records was used for evaluation. Z-score normalization was applied during preprocessing to ensure data consistency, while Principal Component Analysis (PCA) performed efficient feature extraction by reducing dimensionality while preserving essential information. The STO algorithm optimized the DRN weight parameters, improving convergence and classification performance. The proposed STO-DRN achieved 95% accuracy, 96.5% sensitivity, 98.15% specificity, and 96.25% F1-score, outperforming existing methods including PA-GPSR, DPBHA, and Gradient Boosting. These results confirm that STO-DRN delivers superior detection capability, reduced energy consumption, and reliable network performance, establishing it as a robust and practical solution for VANET cybersecurity.</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

Energy-Aware Detection of Black-Hole Attacks in VANETs Using Deep Learning

  • 期刊论文

  • 发表日期：

    2026年6月25日

查看全文
  • Ahmad Adel Abu-Shareha,
  • Mahendihasan S. Heera,
  • Bharat Bhushan,
  • Vemula Harini,
  • Ali Alkwzahy,
  • Vikas Wasson,
  • Prabhat Kumar Sahu,
  • Mary Posonia A.

摘要

Vehicular Ad Hoc Networks (VANETs) are increasingly vulnerable to black-hole attacks that disrupt routing, waste energy, and compromise data reliability. Existing detection methods lack adaptability, struggle with high-dimensional traffic features, and consume excessive energy in resource-constrained environments. This study proposes the Sooty Tern Optimization-Driven Recurrent Network (STO-DRN) for accurate and energy-efficient black-hole attack detection in VANETs. The VANET Routing Attack Detection Dataset (V-RADD) containing 3,000 network traffic records was used for evaluation. Z-score normalization was applied during preprocessing to ensure data consistency, while Principal Component Analysis (PCA) performed efficient feature extraction by reducing dimensionality while preserving essential information. The STO algorithm optimized the DRN weight parameters, improving convergence and classification performance. The proposed STO-DRN achieved 95% accuracy, 96.5% sensitivity, 98.15% specificity, and 96.25% F1-score, outperforming existing methods including PA-GPSR, DPBHA, and Gradient Boosting. These results confirm that STO-DRN delivers superior detection capability, reduced energy consumption, and reliable network performance, establishing it as a robust and practical solution for VANET cybersecurity.