Scholarlink

<p>Modern technology’s increasing convergence has given rise to a wide variety of cyber threats. In a similar vein, the Industrial Internet of Things (IIoT) presents new opportunities for improving service quality for new applications due to the growth in data generated by linked devices. On the converse side, such remarkable opportunities may accelerate vulnerable circumstances for IoT. The expanded connectivity, more openness, and widespread use of low-power communication devices of IIoT make its security a major concern. Motivated by this, we propose an Artificial Intelligence (AI)- enabled, edge-based intrusion detection module for IIoT networks. The proposed work used the Sparse Auto Encoder-Bidirectional Long-short Term Memory (SAE-BLSTM) model for such a purpose. The SAE automatically extracts the patterns of the IIoT network through its sparse coding technique, while BLSTM then uses the output of the SAE for the identification of threats and intrusions. The proposed scheme is intended to pinpoint specific attack categories, aid security analysts in providing early warning, and help security analysts in adopting proactive defensive methods. The proposed framework delivers great efficiency and enhanced security. More precisely, the proposed framework achieved 99.61% and 99.22% accuracy under the CICIDDoS2019 and CICIDS2018 datasets and outclassed some traditional and contemporary state-of-the-art methods.</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 hybrid deep learning based intrusion detection framework to identify cyber attacks in edge-based IIoT

  • 期刊论文

  • 发表日期：

    2026年3月18日

查看全文
  • Mahmood Al-Bahri,
  • Mohammed Saleh Ali Muthanna,
  • Muhammad Zakarya,
  • Reem Ibrahim Alkanhel,
  • Ayaz Ali Khan,
  • Abdullah Alshahrani

摘要

Modern technology’s increasing convergence has given rise to a wide variety of cyber threats. In a similar vein, the Industrial Internet of Things (IIoT) presents new opportunities for improving service quality for new applications due to the growth in data generated by linked devices. On the converse side, such remarkable opportunities may accelerate vulnerable circumstances for IoT. The expanded connectivity, more openness, and widespread use of low-power communication devices of IIoT make its security a major concern. Motivated by this, we propose an Artificial Intelligence (AI)- enabled, edge-based intrusion detection module for IIoT networks. The proposed work used the Sparse Auto Encoder-Bidirectional Long-short Term Memory (SAE-BLSTM) model for such a purpose. The SAE automatically extracts the patterns of the IIoT network through its sparse coding technique, while BLSTM then uses the output of the SAE for the identification of threats and intrusions. The proposed scheme is intended to pinpoint specific attack categories, aid security analysts in providing early warning, and help security analysts in adopting proactive defensive methods. The proposed framework delivers great efficiency and enhanced security. More precisely, the proposed framework achieved 99.61% and 99.22% accuracy under the CICIDDoS2019 and CICIDS2018 datasets and outclassed some traditional and contemporary state-of-the-art methods.