Scholarlink

<p>The Industrial Internet of Things (IIoT) network has significantly advanced industrial automation by enabling continuous data exchange across smart devices; however, it also introduces critical challenges in ensuring data privacy, integrity, and secure communication. Traditional machine learning models often fall short in capturing the complex interconnectivity of IIoT networks and are prone to data breaches in decentralized settings. To address these issues, this paper presents a privacy-preserving blockchain learning model that leverages a custom Graph Neural Network (GNN) architecture to learn the spatial and temporal relationships among IIoT devices for secure and context-aware data classification. The model is evaluated on the X-IIoTID dataset, which includes diverse cyber-physical attack scenarios and normal operational data. Feature extraction is performed using Term Frequency-Inverse Document Frequency (TF-IDF) and statistical encodings, followed by optimal feature subset selection using a novel Adaptive Quantum-Inspired Firefly Optimization (AQFO) algorithm, which intelligently balances global search and local refinement. All model updates and data transfer logs are validated and stored using smart contracts on a private Ethereum blockchain, ensuring immutable record-keeping and decentralized trust. The experimental results demonstrate a classification accuracy of 99.03%, with strong resilience against spoofing, replay, and injection attacks, and low latency in deployment. This framework presents a robust, scalable, and privacy-aware approach for enhancing the security and reliability of IIoT data transmission across industrial systems.</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 Privacy-Preserving Blockchain Learning Model for Reliable Industrial Internet of Things Data Transmission

  • 期刊论文

  • 发表日期：

    2025年6月6日

查看全文
  • Abdulrahman Mathkar Alotaibi

摘要

The Industrial Internet of Things (IIoT) network has significantly advanced industrial automation by enabling continuous data exchange across smart devices; however, it also introduces critical challenges in ensuring data privacy, integrity, and secure communication. Traditional machine learning models often fall short in capturing the complex interconnectivity of IIoT networks and are prone to data breaches in decentralized settings. To address these issues, this paper presents a privacy-preserving blockchain learning model that leverages a custom Graph Neural Network (GNN) architecture to learn the spatial and temporal relationships among IIoT devices for secure and context-aware data classification. The model is evaluated on the X-IIoTID dataset, which includes diverse cyber-physical attack scenarios and normal operational data. Feature extraction is performed using Term Frequency-Inverse Document Frequency (TF-IDF) and statistical encodings, followed by optimal feature subset selection using a novel Adaptive Quantum-Inspired Firefly Optimization (AQFO) algorithm, which intelligently balances global search and local refinement. All model updates and data transfer logs are validated and stored using smart contracts on a private Ethereum blockchain, ensuring immutable record-keeping and decentralized trust. The experimental results demonstrate a classification accuracy of 99.03%, with strong resilience against spoofing, replay, and injection attacks, and low latency in deployment. This framework presents a robust, scalable, and privacy-aware approach for enhancing the security and reliability of IIoT data transmission across industrial systems.