Scholarlink

<p>The rapid expansion of Industrial Internet of Things networks has amplified the need for robust Intrusion Detection Systems (IDS) to safeguard critical infrastructures. This study introduces a novel hybrid IDS combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) architectures with the Reptile Search Algorithm (RSA), an innovative meta-heuristic optimization approach inspired by reptilian adaptive behaviors. This integrated system automates hyperparameter optimization, addressing challenges associated with traditional manual tuning methods. The proposed approach leverages the strengths of CNN for spatial feature extraction and LSTM for temporal sequence analysis, ensuring comprehensive IIoT traffic anomaly detection. The RSA-enhanced IDS outperformed traditional optimization algorithms like Particle Swarm Optimization by achieving faster convergence and improved performance metrics. Despite its effectiveness, the system’s computational complexity during optimization necessitates further refinements. This research highlights the potential of combining deep learning and bio-inspired optimization for fortifying IIoT security against emerging cyber threats. Experimental evaluations were conducted using UNSW-NB15, NCTUKM-IIoT, and TON-IoT datasets, representing diverse real-world IIoT attack scenarios. Results demonstrate superior detection accuracy exceeding 98% across most attack types, coupled with reduced computational overhead and enhanced scalability.</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

An intrusion detection system for industrial IoT using CNN-LSTM and reptile search algorithm

  • 期刊论文

  • 发表日期：

    2025年7月31日

查看全文
  • Yang Chen,
  • Mahdi Mir

摘要

The rapid expansion of Industrial Internet of Things networks has amplified the need for robust Intrusion Detection Systems (IDS) to safeguard critical infrastructures. This study introduces a novel hybrid IDS combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) architectures with the Reptile Search Algorithm (RSA), an innovative meta-heuristic optimization approach inspired by reptilian adaptive behaviors. This integrated system automates hyperparameter optimization, addressing challenges associated with traditional manual tuning methods. The proposed approach leverages the strengths of CNN for spatial feature extraction and LSTM for temporal sequence analysis, ensuring comprehensive IIoT traffic anomaly detection. The RSA-enhanced IDS outperformed traditional optimization algorithms like Particle Swarm Optimization by achieving faster convergence and improved performance metrics. Despite its effectiveness, the system’s computational complexity during optimization necessitates further refinements. This research highlights the potential of combining deep learning and bio-inspired optimization for fortifying IIoT security against emerging cyber threats. Experimental evaluations were conducted using UNSW-NB15, NCTUKM-IIoT, and TON-IoT datasets, representing diverse real-world IIoT attack scenarios. Results demonstrate superior detection accuracy exceeding 98% across most attack types, coupled with reduced computational overhead and enhanced scalability.