Skip to content
Conference Open access

IndustrialShield-X: A Digital Twin-Enabled Explainable Federated Deep Learning Framework for Autonomous Cyber Attack Detection in Industry 5.0

Sep 2026 · 2026 International Conference on Secure IoT and Cybersecurity (ICOSICS) · pp. 747-754 · 0 citations · 24 references

TL;DR

A federated deep learning and explainable artificial intelligence (XAI) integrated cyber-physical intrusion detection framework, called IndustrialShield-X, designed to efficiently process the input images to determine the temperature of the object, proving the effectiveness of the framework with realistic non-IID data distributions.

Abstract

The more the Industrial Internet of Things (IIoT) devices are integrated into industrial systems and cyber-physical systems and edge infrastructure proliferate, the greater the risk of them being exposed to advanced cyber-physical attacks. Traditional IDSs are based on the information contained in the state of the network when it is detected and are dependent on the operational information unique to that particular IDS, but cannot correlate the cyber attack to the state of the physical machines. This paper introduces a federated deep learning and explainable artificial intelligence (XAI) integrated cyber-physical intrusion detection framework, called IndustrialShield-X, for these. The system is designed to efficiently process the input images to determine the temperature of the object, while at the same time each node has its own CNN-BiLSTM model to operate to keep the data localized and private, so the operational processes could be monitored using Digital Twin residual tracking and decide on security measures for an efficient and transparent response, which would be part of the features attributed with the SHAP approach. The framework is benchmarked using a wide variety of experiments performed on the ToN-IoT Network and benchmark datasets of Edge-IIoT, proving the effectiveness of the framework with realistic non-IID data distributions. IndustrialShield-X has an overall accuracy of 98.92%, an accuracy per class of 98.75%, an accuracy per class of 98.60%, an accuracy per class of 98.67% and a false false positive rate of 0.98%. The results are in line with the vision: Federated Intelligence and a physical (process) context provide a privacy preserving, transparent and resilient defence approach for Industry 5.0 ecosystems.

Read PDF

Similar papers

Open access Oct 2026

Autoencoder-based anomaly detection in water cyber-physical systems

Cyber-physical systems (CPS) are at the core of enabling services in the critical infrastructures, such as a smart water network, hence the need for effective anomaly detection to ensure operational safety and security. This paper proposes deep water anomaly detection (DeepWAD), a big data-based multiple purpose anomal...

Sara Tahiri, Mouad Choukhairi, Youssef Fakhiri et al. · 0 citations
Open access Oct 2026

A HYBRID TRANSFORMER-BASED DEEP LEARNING FRAMEWORK FOR REAL-TIME CYBERATTACK DETECTION IN SOFTWARE-DEFINED NETWORKS

A hybrid approach to intrusion detection known as TransSDN-Hybrid, that combines the self-attention mechanisms of transformers with one-dimensional Convolutional Neural Networks to boost the accuracy and effectiveness of detecting cyberattacks in Software-Defined Networking environments is introduced.

A. Vijayan, Pradeep Sankar D. V, C. E. Rajaprabha et al. · 0 citations
Conference Open access Sep 2026

Artificial Intelligence-Powered Cybersecurity in Big Data Environments: Scalable AI-Driven Threat Detection in Distributed Systems: Architecture, Adversarial Risks, and Real-Time Defense Strategies

As connected devices, cloud systems, and digital networks keep growing, cyber threats are becoming more complex and harder to detect. Organizations working with bigdata environments process huge amounts of network traffic, endpoint logs, and cloud activity records every day, so they cannot depend only on manual analysi...

Tripti Singh · 0 citations
Conference Open access Sep 2026

A Hybrid Deep Learning Framework for Intelligent Cyber Threat Detection in IoT Networks

Internet of Things (IoT) is a expertise that allows to interconnect devices to utilized in smart healthcare, intelligent transportation, industrial automation and smart cities. But, due to the heterogeneous and resource constrained nature of IoT devices, they are prone to any kind of cyberattack, such as Distributed De...

Nalini Nara, Raja Sekhar C, R. R. Kumar et al. · 0 citations
Open access Sep 2026

A unified unsupervised deep learning and stride threat modeling framework for multi-domain cyber-physical security in nuclear digital twins

The increasing adoption of Digital Twin (DT) architectures in nuclear power systems has accelerated the convergence of Operational Technology (OT) and Information Technology (IT), exposing critical cyber-physical infrastructure to sophisticated signature-independent attacks that cannot be reliably detected using conven...

Mehtab Alam, Akshay Chamoli, Ashraf Ali et al. · 0 citations
Open access Sep 2026

DUAL-HYBRID TRANSFORMER–CONVOLUTIONAL NEURAL NETWORK WITH EXPLAINABLE ARTIFICIAL INTELLIGENCE FOR ZERO-DAY INTRUSION DETECTION IN HETEROGENEOUS INTERNET OF THINGS NETWORKS

This paper proposes DB-TC-XAI, a 2-in-1 detector to merge 1D spatial features obtained by the 1D convolutional branch with 1D sequential features from traffic by the Transformer-encoder branch using the 2D bidirectional cross-attention module, and achieves competitive detection accuracy with a low false-positive rate.

Rabins Porwal, Anil Kumar Yadav, Shilpi Dubey et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.