Saleem, Jibran ORCID: https://orcid.org/0009-0001-0546-6737, Raza, Umar
ORCID: https://orcid.org/0000-0002-9810-1285, Hammoudeh, Mohammad
ORCID: https://orcid.org/0000-0003-1058-0996 and Holderbaum, William
ORCID: https://orcid.org/0000-0002-1677-9624
(2025)
Machine Learning-Enhanced Attribute-Based Authentication for Secure IoT Access Control.
Sensors, 25 (9).
2779.
ISSN 1424-8220
|
Published Version
Available under License Creative Commons Attribution. Download (3MB) | Preview |
Abstract
The rapid growth of Internet of Things (IoT) devices across industrial and critical sectors requires robust and efficient authentication mechanisms. Traditional authentication systems struggle to balance security, privacy and computational efficiency, particularly in resource-constrained environments such as Industry 4.0. This research presents the SmartIoT Hybrid Machine Learning (ML) Model, a novel integration of Attribute-Based Authentication and a lightweight machine learning algorithm designed to enhance security while minimising computational overhead. The SmartIoT Hybrid ML Model utilises Random Forest classifiers for real-time anomaly detection, dynamically assessing access requests based on user attributes, login patterns and behavioural analysis. The model enhances identity protection while enabling secure authentication without exposing sensitive information by incorporating privacy-preserving Attribute-Based Credentials and Attribute-Based Signatures. Our experimental evaluation demonstrates 86% authentication accuracy, 88% precision and 96% recall, significantly outperforming existing solutions while maintaining an average response time of 112ms, making it suitable for low-power IoT devices. Comparative analysis with state-of-the-art authentication frameworks shows the model’s security resilience, computational efficiency and adaptability in real-world IoT applications.
Impact and Reach
Statistics
Additional statistics for this dataset are available via IRStats2.