Blockchain-Secured Federated Learning Approach for Adaptive Intrusion Detection in IoT Networks

Abstract

The widespread adoption of Internet of Things (IoT) devices has expanded the attack surface, exposing networks to an increasing range of cyber threats. Conventional intrusion detection systems (IDS) struggle to cope with the dynamic, heterogeneous nature of IoT environments and often rely on centralized data processing, which raises privacy risks. This study presents an adaptive intrusion detection system (A-IDS) that integrates federated learning (FL) with blockchain-based trust management to strengthen IoT network security. Using FL, multiple IoT devices collaboratively train a global detection model without sharing raw data, ensuring privacy preservation. The blockchain component establishes a decentralized trust framework that safeguards the FL process and prevents malicious contributions. Performance evaluation through realistic IoT network simulations demonstrates high detection accuracy, strong privacy protection, and resilience to adversarial behavior. Comparisons with centralized and decentralized IDS approaches highlight the proposed system’s advantages in accuracy, robustness, and data confidentiality.