Explainable AI-Driven Security Framework for Cyber-Physical Production Systems in Industry 4.0: Leveraging Immersive Embedded CIoT

Farhan, Muhammad ORCID: https://orcid.org/0000-0002-3649-5717, Naeem, Muhammad Rehan, Almadhor, Ahma ORCID: https://orcid.org/0000-0002-8665-1669, Bashir, Ali Kashif ORCID: https://orcid.org/0000-0003-2601-9327, Zhu, Zhu ORCID: https://orcid.org/0009-0000-2911-1773 and Gadekallu, Thippa Reddy ORCID: https://orcid.org/0000-0003-0097-801X (2025) Explainable AI-Driven Security Framework for Cyber-Physical Production Systems in Industry 4.0: Leveraging Immersive Embedded CIoT. IEEE Transactions on Consumer Electronics. pp. 1-8. ISSN 0098-3063

Preview

Accepted Version Available under License Creative Commons Attribution. Download (6MB) | Preview

Abstract

This paper proposes a new security framework of Explainable Artificial Intelligence (XAI) for Cyber-Physical Production Systems (CPPSs) in the Industry 4.0 paradigm. An integral part of the framework is the incorporation of XAI into immersive embedded Consumer Internet of Things (CIoT) systems to promote transparency and interpretability, thereby improving real-time decision-making in AI-driven security mechanisms. The framework leverages SHAP and LIME techniques to provide human operators with clear insights into the AI-based security decision-making process, fostering trust and facilitating effective teaming between humans and AI assets. The proposed solution was validated and tested in an innovative manufacturing environment, where it could detect, interpret, and mitigate security threats in real-time, enhancing the security posture of the CPPS. Experimental results demonstrate that the framework achieves high accuracy in threat detection and significantly reduces false positives, as operators can fine-tune security policies based on the explainable AI insights. The importance of explainability in AI-driven security systems is emphasized, and it is demonstrated that immersive CIoT technologies can tackle the emerging security issues in CPPS.