Scholarlink

<p>Zero-dynamics assaults (ZDAs) are a stealthy class of cyber-physical threats that leverage the inherent zero dynamics of industrial networked control systems (NCSs) to avoid detection. This paper offers a comparative analysis of two ZDA detection strategies-passive and active-applied to two benchmark industrial processes: the Tennessee Eastman Process (TEP) and the Sextuple Tank Process (STP). The passive technique involves changing system dynamics through matrix perturbation, actuator gain modification, and sensor enhancement, resulting in a detection accuracy of up to 83% with a minimum detection latency of 2.003 ms. An Intermittent Unknown Input Kalman Filter (IIKF) is used in the active technique, which allows for ongoing monitoring even in the event of induced data loss but results in higher detection delays, especially in noisy environments. In contrast to earlier theoretical studies, our research prioritizes real-world applicability by examining practical implementation issues, scalability, and the trade-offs among accuracy, detection speed, and system redesign costs. The results connect theoretical concepts with practical implementation, providing recommendations to improve the cybersecurity of industrial NCSs against advanced stealth assaults.</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

Detection of zero-dynamics attacks in industrial networked control systems: comparative analysis of passive and active approaches

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年5月23日

查看全文
  • Ayesha Ayub,
  • Muhammad Asim khan,
  • Zuhairiah Zainal Abidin,
  • Naglaa F. Soliman,
  • Hamza Ahmad

摘要

Zero-dynamics assaults (ZDAs) are a stealthy class of cyber-physical threats that leverage the inherent zero dynamics of industrial networked control systems (NCSs) to avoid detection. This paper offers a comparative analysis of two ZDA detection strategies-passive and active-applied to two benchmark industrial processes: the Tennessee Eastman Process (TEP) and the Sextuple Tank Process (STP). The passive technique involves changing system dynamics through matrix perturbation, actuator gain modification, and sensor enhancement, resulting in a detection accuracy of up to 83% with a minimum detection latency of 2.003 ms. An Intermittent Unknown Input Kalman Filter (IIKF) is used in the active technique, which allows for ongoing monitoring even in the event of induced data loss but results in higher detection delays, especially in noisy environments. In contrast to earlier theoretical studies, our research prioritizes real-world applicability by examining practical implementation issues, scalability, and the trade-offs among accuracy, detection speed, and system redesign costs. The results connect theoretical concepts with practical implementation, providing recommendations to improve the cybersecurity of industrial NCSs against advanced stealth assaults.