Scholarlink

Purpose <p>Vehicular Ad-Hoc Networks are very vulnerable to DDoS attacks. Existing methods are less efficient in terms of false alarms, computational overhead, and adaptability in DDoS attacks. This paper proposes an efficient framework for DDoS detection, mitigation, and secure communication in VANETs.</p> Methods <p>In this manuscript, Enhanced Distributed Denial of Service Detection and Mitigation in Vehicular Ad Hoc Networks utilizing Binarized Simplicial Convolutional Neural Network optimized with Educational Competition Optimizer (DDOS-VANET-BSCNN-ECO) is proposed. Input data is gathered from CIC-DDoS-2019 is pre-processed using Reverse Lognormal Kalman Filtering (RLKF), and extracting the statistical features through the Lifted Euler Characteristic Transform (LECT). Then the Soccer Match Algorithm (SMA) is used for selecting the most relevant features. The selected features are supplied to the Binarized Simplicial Convolutional Neural Network (BSCNN) to classify the DDoS as DrDoS_SSDP, TFTP, DrDoS_UDP, Syn and DrDoS_MSSQL. For effective mitigation, a Duelling Double Deep Q-Network (DDDQN) module dynamically regulates data transmission based on attack severity. To enhance security, Non-Interactive Practical Proof-of Storage (NPPOS) mechanism ensures encryption, safeguarding communication integrity.</p> Results <p>The proposed framework achieved 99.45% accuracy, 98.63% f1-score, and 98.83% recall, outperforming existing techniques like DDoS-VANET-LSTM, DDoS-VANET-KNN and DDoS-VANET-RTACNN respectively.</p> Conclusion <p>The proposed DDOS-VANET-BSCNN-ECO framework provides an efficient, scalable, and secure solution for real-time DDoS detection and mitigation in VANET environments, significantly improving the reliability and robustness of the overall network.</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

Enhanced Distributed Denial of Service Detection and Mitigation in Vehicular Ad Hoc Networks utilizing Binarized Simplicial Convolutional Neural Network Optimized with Educational Competition Optimizer

  • 期刊论文

  • 发表日期：

    2026年5月22日

查看全文
  • A. Umamaheswari,
  • Badarla Anil,
  • L. Sherly Puspha Annabel,
  • Moulieswaran Elavarasu

摘要

Purpose

Vehicular Ad-Hoc Networks are very vulnerable to DDoS attacks. Existing methods are less efficient in terms of false alarms, computational overhead, and adaptability in DDoS attacks. This paper proposes an efficient framework for DDoS detection, mitigation, and secure communication in VANETs.

Methods

In this manuscript, Enhanced Distributed Denial of Service Detection and Mitigation in Vehicular Ad Hoc Networks utilizing Binarized Simplicial Convolutional Neural Network optimized with Educational Competition Optimizer (DDOS-VANET-BSCNN-ECO) is proposed. Input data is gathered from CIC-DDoS-2019 is pre-processed using Reverse Lognormal Kalman Filtering (RLKF), and extracting the statistical features through the Lifted Euler Characteristic Transform (LECT). Then the Soccer Match Algorithm (SMA) is used for selecting the most relevant features. The selected features are supplied to the Binarized Simplicial Convolutional Neural Network (BSCNN) to classify the DDoS as DrDoS_SSDP, TFTP, DrDoS_UDP, Syn and DrDoS_MSSQL. For effective mitigation, a Duelling Double Deep Q-Network (DDDQN) module dynamically regulates data transmission based on attack severity. To enhance security, Non-Interactive Practical Proof-of Storage (NPPOS) mechanism ensures encryption, safeguarding communication integrity.

Results

The proposed framework achieved 99.45% accuracy, 98.63% f1-score, and 98.83% recall, outperforming existing techniques like DDoS-VANET-LSTM, DDoS-VANET-KNN and DDoS-VANET-RTACNN respectively.

Conclusion

The proposed DDOS-VANET-BSCNN-ECO framework provides an efficient, scalable, and secure solution for real-time DDoS detection and mitigation in VANET environments, significantly improving the reliability and robustness of the overall network.