Scholarlink

<p>The Internet of Things (IoT) is rapidly emerging across various domains. The Industrial Internet of Things (IIoT) refers to an extension of IoT to various industries. IIoT enhances efficiency in industries by connecting diverse smart devices. The heterogeneity of IIoT devices and their resource constraints demand adaptable, lightweight, and robust security solutions. This paper presents a two-stage authentication and anomaly detection framework tailored for IIoT systems. The proposed model integrates lightweight hash-based authentication for device authentication and an adaptive ML-based anomaly detection phase. The anomaly detection process analyses a device’s contextual data and network communication patterns.Both datasets are utilised to learn device behavioural patterns during training using time-stamped data. The model incorporates device-specific contextual features to account for the heterogeneity of different industrial device types. Evaluations using ToN-IoT and BoT-IoT datasets demonstrate the model’s high accuracy and low computational overhead. This efficient and adaptable solution effectively meets the security demands of the IIoT environments.</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 Two-Stage Authentication and Anomaly Detection Model for Securing IIoT Systems

  • 期刊论文

  • 发表日期：

    2026年5月20日

查看全文
  • Tithi Ghosh,
  • Chandan Giri,
  • Surajit Kumar Roy

摘要

The Internet of Things (IoT) is rapidly emerging across various domains. The Industrial Internet of Things (IIoT) refers to an extension of IoT to various industries. IIoT enhances efficiency in industries by connecting diverse smart devices. The heterogeneity of IIoT devices and their resource constraints demand adaptable, lightweight, and robust security solutions. This paper presents a two-stage authentication and anomaly detection framework tailored for IIoT systems. The proposed model integrates lightweight hash-based authentication for device authentication and an adaptive ML-based anomaly detection phase. The anomaly detection process analyses a device’s contextual data and network communication patterns.Both datasets are utilised to learn device behavioural patterns during training using time-stamped data. The model incorporates device-specific contextual features to account for the heterogeneity of different industrial device types. Evaluations using ToN-IoT and BoT-IoT datasets demonstrate the model’s high accuracy and low computational overhead. This efficient and adaptable solution effectively meets the security demands of the IIoT environments.