Federated Learning-Driven Intrusion Detection with Blockchain-Backed Trust Frameworks to Strengthen IoT Network Security

Abstract

The swift expansion of Internet of Things (IoT) ecosystems greatly enlarged the attack surface potential, rendering traditional centralized intrusion detection systems (IDS) progressively inadequate for effective network defense. This paper presents an adaptive IDS framework that combines federated learning (FL) with blockchain trust management to enhance IoT network security. In the envisioned design, IoT devices cooperatively build a shared intrusion detection model without sharing raw data directly, ensuring user privacy and bandwidth saving. A blockchain layer is leveraged to establish a distributed trust protocol for transparent verification and logging of contributions from each device to the FL process. Experimental tests, incorporating simulation and field deployment, demonstrate that the system is capable of successfully identifying varied IoT-targeted attacks at low false-positive rates. The results highlight the efficiency of the introduced approach in improving IoT security, while at the same time overcoming privacy, scalability, and trust challenges.