Hybrid Deep Learning and Blockchain-Enabled Intrusion Detection System for IoT Networks using Enhanced Dataset Fusion

Abstract

The growth of the IoT ecosystem is matching an increase in sophisticated cyber-attacks. Traditional intrusion detection systems are unable to manage the rapid changes in threats within an IoT environment comprising diverse devices. This paper proposes a hybrid deep learning and blockchain-enabled intrusion detection system (HDB-IDS) designed for securing IoT networks. To boost the accuracy and generalization capability of our model, we merge the data from UNSW-NB15 and BoT-IoT and address class imbalance and enhance attack diversity. The deep learning component identifies attack patterns, while the blockchain ensures secure, decentralized alert sharing and auditability. The suggested model has been shown to achieve higher accuracy, precision, and F1-score than the leading IDS models. The integration of blockchain ensures data integrity and reliability with minimal latency. This innovative architecture enhances cybersecurity resilience in modern IoT environments.