An Artificial Intelligence and Federated Learning-Based Cybersecurity Framework for Critical Industrial Infrastructure Connected to the Internet of Things (IoT)

The convergence of the Internet of Things (IoT) with critical industrial infrastructure has fundamentally transformed operational paradigms, enabling unprecedented levels of automation, real-time monitoring, and data-driven decision-making. However, this digital transformation has exponentially expanded the cyberattack surface, rendering traditional perimeter-based security models inadequate against sophisticated, evolving threats targeting sectors such as energy, water treatment, transportation, and healthcare. Recent reports indicate a ۳۸۷% increase in IoT and operational technology attacks on the energy sector compared to the previous year, while hacktivist sightings surged by ۵۱% globally in ۲۰۲۵. This paper presents a comprehensive cybersecurity framework that integrates Artificial Intelligence (AI) and Federated Learning (FL) to address the unique challenges of securing IoT-enabled critical infrastructure. The proposed framework leverages distributed intelligence through FL to enable collaborative threat detection across geographically dispersed industrial sites without compromising sensitive operational data privacy. Advanced deep learning architectures, including hybrid CNN-LSTM and CNN-GRU models, are deployed at edge nodes for real-time anomaly detection, while blockchain technology ensures the integrity and immutability of model updates. The framework incorporates zero-trust security principles, quantum-resistant cryptographic mechanisms, and explainable AI to enhance transparency and trust in decision-making processes. Experimental validation on benchmark datasets including Edge-IoTset, CIC-IDS۲۰۱۷, and UNSW-NB۱۵ demonstrates detection accuracy exceeding ۹۷.۸% with significant reductions in communication overhead and latency compared to centralized approaches. This paper also examines implementation challenges, including data heterogeneity, adversarial robustness, and resource constraints, while proposing future research directions for resilient, scalable, and privacy-preserving cybersecurity in industrial IoT ecosystems.

کلیدواژه ها:

Artificial Intelligence ، Cybersecurity ، Critical Infrastructure ، Federated Learning