Scholarlink

<p>Power plants are highly sensitive and interconnected environments where security breaches or reliability issues can lead to severe consequences. The increasing integration of Internet of Things (IoT) devices and sophisticated Cyber-Physical Human Systems (ECPHS) within these facilities significantly expands their attack surface. While Physical Unclonable Functions (PUFs) offer a promising solution for secure device authentication, their long-term effectiveness is inherently challenged by environmental fluctuations, aging, and physical tampering. This paper proposes a novel, dual-layered framework to robustly enhance the security and enduring reliability of PUF-enabled systems in such critical applications. Our approach synergistically combines cryptographic techniques with machine learning capabilities. First, the security of PUF responses is fortified through encryption using the lightweight SPECK algorithm, with cryptographic keys securely derived via SHA-3 hashing. This not only protects PUF responses from direct attacks but also strengthens the system’s ability to detect abnormal PUF behavior. Second, we integrate machine learning for proactive integrity monitoring. By training various regression models on preprocessed PUF data, the system learns to predict subtle changes in PUF characteristics caused by aging, environmental stress, or tampering. This predictive capability enables early detection of potential vulnerabilities, with performance validated through metrics such as Mean Absolute Error and Mean Squared Error. The proposed methodology delivers a comprehensive and adaptive security solution for IoT and ECPHS components within power plants, safeguarding critical assets like smart meters, grid sensors, and control systems. By combining resilient cryptography with intelligent, predictive analytics, our framework ensures the long-term integrity and operational continuity of interconnected systems, proactively identifying and mitigating risks to PUF reliability before they escalate.</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

Securing IoT Devices with PUFs: Mitigating Aging and Tampering through Cryptography and Machine Learning

  • 期刊论文

  • 发表日期：

    2026年4月18日

查看全文
  • Alaa Abdelmonem,
  • Ahmed Salem,
  • Fahima A. Maghraby

摘要

Power plants are highly sensitive and interconnected environments where security breaches or reliability issues can lead to severe consequences. The increasing integration of Internet of Things (IoT) devices and sophisticated Cyber-Physical Human Systems (ECPHS) within these facilities significantly expands their attack surface. While Physical Unclonable Functions (PUFs) offer a promising solution for secure device authentication, their long-term effectiveness is inherently challenged by environmental fluctuations, aging, and physical tampering. This paper proposes a novel, dual-layered framework to robustly enhance the security and enduring reliability of PUF-enabled systems in such critical applications. Our approach synergistically combines cryptographic techniques with machine learning capabilities. First, the security of PUF responses is fortified through encryption using the lightweight SPECK algorithm, with cryptographic keys securely derived via SHA-3 hashing. This not only protects PUF responses from direct attacks but also strengthens the system’s ability to detect abnormal PUF behavior. Second, we integrate machine learning for proactive integrity monitoring. By training various regression models on preprocessed PUF data, the system learns to predict subtle changes in PUF characteristics caused by aging, environmental stress, or tampering. This predictive capability enables early detection of potential vulnerabilities, with performance validated through metrics such as Mean Absolute Error and Mean Squared Error. The proposed methodology delivers a comprehensive and adaptive security solution for IoT and ECPHS components within power plants, safeguarding critical assets like smart meters, grid sensors, and control systems. By combining resilient cryptography with intelligent, predictive analytics, our framework ensures the long-term integrity and operational continuity of interconnected systems, proactively identifying and mitigating risks to PUF reliability before they escalate.