Scholarlink

The rapid growth of the Internet and the telecommunications industry has led to a huge explosion of network size and data consistency. Therefore, when VANET (vehicular ad hoc network) grows and grows more and more in terms of transmission speed, network connectivity, security, and safety with the deployment of cutting applications, there will be major changes in wireless communication. In order to prevent external communications from being hacked, this study provides an intrusion detection system (IDS) for VANET that makes use of obfuscation. In this paper we proposed intrusion detection system for VANET which used the Attentive Interpretable Tabular Learning (TabNet) architecture of deep learning. TabNet model has characteristics of good interpretability and fast training speed. we examined the existing machine and deep learning framework and application in cyber security and VANET and PyTorch-Fast.ai is selected due to it is designed for Central Processing Unit (CPU) usage but because of training speed is high we changed data block fetching method from Tabular panda to NumPy array. Finally, from the evaluated metrics, we have proposed the best Deep Neural Network (DNN) design suitable for the IDS. With an accuracy of 98.12% and a false alarm rate (FAR) of 0.78%.

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

Tabular Deep Learning-Based Light-Weight Intrusion Detection System for VANET System

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Javed Shaikh,
  • Vinay Yandrapalli,
  • Nilesh T. Gole,
  • Aditya Patil,
  • Nilesh P. Bodne,
  • Smitta Matte,
  • Syed Umar

摘要

The rapid growth of the Internet and the telecommunications industry has led to a huge explosion of network size and data consistency. Therefore, when VANET (vehicular ad hoc network) grows and grows more and more in terms of transmission speed, network connectivity, security, and safety with the deployment of cutting applications, there will be major changes in wireless communication. In order to prevent external communications from being hacked, this study provides an intrusion detection system (IDS) for VANET that makes use of obfuscation. In this paper we proposed intrusion detection system for VANET which used the Attentive Interpretable Tabular Learning (TabNet) architecture of deep learning. TabNet model has characteristics of good interpretability and fast training speed. we examined the existing machine and deep learning framework and application in cyber security and VANET and PyTorch-Fast.ai is selected due to it is designed for Central Processing Unit (CPU) usage but because of training speed is high we changed data block fetching method from Tabular panda to NumPy array. Finally, from the evaluated metrics, we have proposed the best Deep Neural Network (DNN) design suitable for the IDS. With an accuracy of 98.12% and a false alarm rate (FAR) of 0.78%.