Scholarlink

<p>The significance of vehicular networks in enhancing quality of life, security, and protection is essential to the advancement of smart cities. As smart vehicles become more common, the confidentiality and security challenges in vehicular ad hoc networks (VANETs) have gained critical importance. VANETs are particularly susceptible to security threats due to their dynamic and decentralized communication networks. In this study, we propose a hybrid model that combines a 1-dimensional convolutional neural network (1D-CNN) for feature extraction and an XGBoost classifier for efficient VANET threat classification. Additionally, we employed the Robust Fox Optimization Algorithm (RFOA) for hyperparameter tuning of the XGBoost classifier, further refining the model’s performance. To ensure realistic evaluation, we utilized the Simulation of Urban MObility (SUMO) to simulate VANET scenarios and evaluated our approach using a comprehensive dataset of VANET communication patterns. Our model was trained on the CIC-IDS and TON_IoT datasets, known for capturing diverse network attack types, and was evaluated using performance metrics such as accuracy, precision, recall, F1-score, and ROC-AUC. The results demonstrated that our hybrid model outperforms recent VANET security methods, including rule-based approaches and support vector machines, in terms of accuracy and detection capability. The proposed model achieved high accuracy on the test set and successfully detected various threats in both the CIC-IDS and the TON_IoT datasets, such as DoS attacks and malicious node behaviors. This hybrid IDS model, combining 1D-CNN feature extraction with an RFOA-tuned XGBoost classifier, provides a powerful solution for enhancing VANET security and has the potential for further improvement by incorporating additional network insights.</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

A Novel and Robust Intrusion Detection Model for VANETs Using 1D-CNN and RFOA-Optimized XGBoost

  • 期刊论文

  • 发表日期：

    2026年4月5日

查看全文
  • Nabil Nissar,
  • Maha Karroum

摘要

The significance of vehicular networks in enhancing quality of life, security, and protection is essential to the advancement of smart cities. As smart vehicles become more common, the confidentiality and security challenges in vehicular ad hoc networks (VANETs) have gained critical importance. VANETs are particularly susceptible to security threats due to their dynamic and decentralized communication networks. In this study, we propose a hybrid model that combines a 1-dimensional convolutional neural network (1D-CNN) for feature extraction and an XGBoost classifier for efficient VANET threat classification. Additionally, we employed the Robust Fox Optimization Algorithm (RFOA) for hyperparameter tuning of the XGBoost classifier, further refining the model’s performance. To ensure realistic evaluation, we utilized the Simulation of Urban MObility (SUMO) to simulate VANET scenarios and evaluated our approach using a comprehensive dataset of VANET communication patterns. Our model was trained on the CIC-IDS and TON_IoT datasets, known for capturing diverse network attack types, and was evaluated using performance metrics such as accuracy, precision, recall, F1-score, and ROC-AUC. The results demonstrated that our hybrid model outperforms recent VANET security methods, including rule-based approaches and support vector machines, in terms of accuracy and detection capability. The proposed model achieved high accuracy on the test set and successfully detected various threats in both the CIC-IDS and the TON_IoT datasets, such as DoS attacks and malicious node behaviors. This hybrid IDS model, combining 1D-CNN feature extraction with an RFOA-tuned XGBoost classifier, provides a powerful solution for enhancing VANET security and has the potential for further improvement by incorporating additional network insights.