Scholarlink

<p>As the demand for Electric Vehicles (EVs) grows, ensuring the security and integrity of their communication networks becomes paramount. Traditional Intrusion Detection System (IDS) approaches often struggle with the dynamic and complex nature of EVNs. To address this challenge, the paper presents an enhanced IDS for Electric EVNs utilizing a hybrid architecture that combines the strengths of Res-Net and Alex-Net models. The hybrid model capitalizes on the residual learning of Res-Net and the deep feature extraction capability of Alex-Net, offering enhanced detection accuracy and reduced false positives. This proposed approach involves data preprocessing step, feature extraction with CNN-LSTM, model optimization in a timely manner for the detection of vehicular network. In this paper “Car hacking attack defense challenge dataset” is used for the detection of flooding, replay, fuzzing, and spoofing attacks in an intra-vehicular network. Through extensive experimentation and evaluation, the proposed IDS demonstrate superior performance in identifying malicious activities within EVNs, ensuring improved security and reliability. Res-Net for flooding, spoofing, replay, and fuzzing attacks, consistently outperforms these approaches, with accuracy values ranging from 98.64% to 99.74%, along with higher precision, recall, and F1 scores.</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

Hybrid net: enhanced DTL based intrusion detection system for electric vehicular network using hybrid architecture

  • 期刊论文

  • 发表日期：

    2025年11月10日

查看全文
  • Neetu Singh,
  • Ritu Agarwal

摘要

As the demand for Electric Vehicles (EVs) grows, ensuring the security and integrity of their communication networks becomes paramount. Traditional Intrusion Detection System (IDS) approaches often struggle with the dynamic and complex nature of EVNs. To address this challenge, the paper presents an enhanced IDS for Electric EVNs utilizing a hybrid architecture that combines the strengths of Res-Net and Alex-Net models. The hybrid model capitalizes on the residual learning of Res-Net and the deep feature extraction capability of Alex-Net, offering enhanced detection accuracy and reduced false positives. This proposed approach involves data preprocessing step, feature extraction with CNN-LSTM, model optimization in a timely manner for the detection of vehicular network. In this paper “Car hacking attack defense challenge dataset” is used for the detection of flooding, replay, fuzzing, and spoofing attacks in an intra-vehicular network. Through extensive experimentation and evaluation, the proposed IDS demonstrate superior performance in identifying malicious activities within EVNs, ensuring improved security and reliability. Res-Net for flooding, spoofing, replay, and fuzzing attacks, consistently outperforms these approaches, with accuracy values ranging from 98.64% to 99.74%, along with higher precision, recall, and F1 scores.