Secure Federated Learning Framework for Anomaly Detection in IoT Networks using Lightweight Cryptographic Hashing

The increasing deployment of Internet of Things (IoT) devices in critical infrastructures has made anomaly detection a cornerstone for maintaining system integrity and security. However, centralized data processing introduces privacy risks and communication overheads. This paper proposes a secure federated learning (FL) framework designed for distributed IoT environments, enabling collaborative anomaly detection without exposing raw sensor data. To address privacy and integrity concerns, this work proposes to integrate a lightweight cryptographic hashing mechanism that ensures data authenticity and resistance against adversarial attacks during model aggregation. The framework is tailored for low-power IoT devices, optimizing both communication efficiency and model accuracy. Through extensive simulations on benchmark IoT anomaly datasets, the proposed approach demonstrates superior detection performance while maintaining low computational and communication costs. The results confirm that our framework strikes a practical balance between privacy preservation, scalability, and real-time responsiveness for secure IoT network monitoring.