Scholarlink

<p>Nowadays, the Industrial Internet of Things (IIoT) is a quickly evolving advanced technology with the potential to digitalize and connect several sectors for enormous business prospects and the emergence of global networks. IIoT has been applied in various fields like transportation, logistics, energy utilities and aviation, oil and gas, manufacturing, mining, and metals. Though IIoT offers good opportunities for the growth of various industrial applications, they are expected to face cyberattacks and requires stronger security mechanisms. The intrusion detection system (IDS), which examines network traffic and recognizes network behaviour, serves as a primary security mechanism to secure IIoT applications from attacks. In recent times, the application of machine learning (ML) as well as deep learning (DL) have proven to reduce numerous security risks and improve the intrusion detection performance. This paper develops an Intrusion Detection System for Industrial Internet of Things Using Ensemble Learning and Bio-Inspired Optimization (IDSIIoT-ELBIO) Model. The paper aims to develop an efficient IDS for IIoT environments to enhance security, reliability, and real-time threat mitigation. Initially, the standard scaling method is leveraged in the data standardization step to transform the input data into beneficial formats. For the feature subset selection, the proposed IDSIIoT-ELBIO model designs a grasshopper optimization algorithm (GOA) to select the optimal features. Furthermore, the ensemble of three classifiers, such as variational autoencoder (VAE), temporal convolutional network (TCN), and deep Q-network (DQN), has been deployed for intrusion detection operation. Eventually, the parameter fine-tuning method is mainly implemented by a harris hawk optimizer (HHO) algorithm. The effectiveness of the IDSIIoT-ELBIO model has been validated through comprehensive experimental analysis utilizing the benchmark dataset. The numerical outcome displays that the IDSIIoT-ELBIO model has improved performance and scalability under different measures over the existing approaches.</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 Lightweight Intrusion Detection System for Real-Time Industrial Internet of Things using Ensemble Learning with Heuristic Approach

  • 期刊论文

  • 发表日期：

    2026年5月11日

查看全文
  • G. Anitha,
  • Hariprasath Manoharan

摘要

Nowadays, the Industrial Internet of Things (IIoT) is a quickly evolving advanced technology with the potential to digitalize and connect several sectors for enormous business prospects and the emergence of global networks. IIoT has been applied in various fields like transportation, logistics, energy utilities and aviation, oil and gas, manufacturing, mining, and metals. Though IIoT offers good opportunities for the growth of various industrial applications, they are expected to face cyberattacks and requires stronger security mechanisms. The intrusion detection system (IDS), which examines network traffic and recognizes network behaviour, serves as a primary security mechanism to secure IIoT applications from attacks. In recent times, the application of machine learning (ML) as well as deep learning (DL) have proven to reduce numerous security risks and improve the intrusion detection performance. This paper develops an Intrusion Detection System for Industrial Internet of Things Using Ensemble Learning and Bio-Inspired Optimization (IDSIIoT-ELBIO) Model. The paper aims to develop an efficient IDS for IIoT environments to enhance security, reliability, and real-time threat mitigation. Initially, the standard scaling method is leveraged in the data standardization step to transform the input data into beneficial formats. For the feature subset selection, the proposed IDSIIoT-ELBIO model designs a grasshopper optimization algorithm (GOA) to select the optimal features. Furthermore, the ensemble of three classifiers, such as variational autoencoder (VAE), temporal convolutional network (TCN), and deep Q-network (DQN), has been deployed for intrusion detection operation. Eventually, the parameter fine-tuning method is mainly implemented by a harris hawk optimizer (HHO) algorithm. The effectiveness of the IDSIIoT-ELBIO model has been validated through comprehensive experimental analysis utilizing the benchmark dataset. The numerical outcome displays that the IDSIIoT-ELBIO model has improved performance and scalability under different measures over the existing approaches.