A Dynamic Quorum Ensemble Intrusion Detection System for Enhanced IoT Security

Abstract

In the rapidly evolving landscape of Internet of Things (IoT) networks, ensuring robust security against sophisticated and evolving cyber threats remains a critical challenge, as Traditional Intrusion Detection Systems (IDS) often suffer from limitations in adaptability and accuracy within such dynamic and resource-constrained environments. To overcome these challenges, this study proposes a novel cooperative IDS based on a Dynamic Quorum Ensemble (DQE) methodology. The proposed framework employs a hierarchical architecture in which Cluster Heads (CHs) are selected using predefined criteria, and IDS nodes are individually trained for effective anomaly detection. Upon anomaly detection, a temporary ensemble of IDS nodes is formed to validate the threat, culminating in a consensus-driven decision communicated across the network. The performance of the proposed DQE-IDS was evaluated using the WUSTL-EHMS-2020 dataset and benchmarked against several state-of-the-art IDS approaches, including Extreme Learning Machine (ELM)-based IDS, Feature Reduction (FR)-IDS, and Graph Neural Network (GNN)-based IDS. Experimental results demonstrate that the DQE-IDS consistently achieves superior performance in terms of accuracy, precision, recall, F1-score, and False Positive Rate (FPR) across various ensemble sizes ranging from 3 to 9. Overall, the findings indicate that the proposed system represents a significant advancement in intrusion detection for IoT ecosystems, offering enhanced reliability, scalability, and responsiveness over existing methodologies.