Scholarlink

<p>Integrating innovative technologies with future networks enables significant advances in real-time distributed systems. The Industrial Internet of Things (IIoT) and edge computing enable communication devices to handle time-constrained services in critical and real-time interactive systems. In recent decades, numerous interactive network applications have been proposed that leverage high-performance computing systems to fulfill industrial requirements and enhance production without human intervention. Recently, a Tiny Machine Learning (TinyML) approach has emerged for resource-constrained IIoT environments to achieve lightweight workflows, low latency, and efficient resource utilization. However, most approaches still introduce computational complexity, particularly as networks grow and resources are limited. In emerging applications, another major research issue is to attain reliable communication even in the presence of malicious devices, as a result, most of the approaches lack intelligence to address security attacks and enhance the probability of data breaches in a trustless environment. This study proposes a predictive model using regression model for real-time data processing and autonomous decision-making to improve energy efficiency in the IIoT network. Furthermore, secure edge-level communication is designed to operate under unpredictable conditions, enabling a chain of trust between IIoT devices and maintaining network integrity in the industrial environment. Based on simulation, the proposed model is tested and validated to assess its performance relative to existing solutions using various metrics across the IoT-edge infrastructure. Its statistical results significantly improved the packet reception rate by an average of 45% across different IoT devices and 52% across different attack levels, while reducing data latency by an average of 37% and 44%, respectively, even in the presence of malicious devices. Further performance results indicate improved energy consumption, reduced complexity overhead, and a lower false-positive rate compared to baseline schemes.</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

Distributed trust-driven intelligence for edge-level prediction in mobile industrial internet of things

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月27日

查看全文
  • Zahoor Jan,
  • Mohammad Siraj,
  • Khalid Haseeb,
  • Saif A. Alsaif,
  • Sadia Din

摘要

Integrating innovative technologies with future networks enables significant advances in real-time distributed systems. The Industrial Internet of Things (IIoT) and edge computing enable communication devices to handle time-constrained services in critical and real-time interactive systems. In recent decades, numerous interactive network applications have been proposed that leverage high-performance computing systems to fulfill industrial requirements and enhance production without human intervention. Recently, a Tiny Machine Learning (TinyML) approach has emerged for resource-constrained IIoT environments to achieve lightweight workflows, low latency, and efficient resource utilization. However, most approaches still introduce computational complexity, particularly as networks grow and resources are limited. In emerging applications, another major research issue is to attain reliable communication even in the presence of malicious devices, as a result, most of the approaches lack intelligence to address security attacks and enhance the probability of data breaches in a trustless environment. This study proposes a predictive model using regression model for real-time data processing and autonomous decision-making to improve energy efficiency in the IIoT network. Furthermore, secure edge-level communication is designed to operate under unpredictable conditions, enabling a chain of trust between IIoT devices and maintaining network integrity in the industrial environment. Based on simulation, the proposed model is tested and validated to assess its performance relative to existing solutions using various metrics across the IoT-edge infrastructure. Its statistical results significantly improved the packet reception rate by an average of 45% across different IoT devices and 52% across different attack levels, while reducing data latency by an average of 37% and 44%, respectively, even in the presence of malicious devices. Further performance results indicate improved energy consumption, reduced complexity overhead, and a lower false-positive rate compared to baseline schemes.