A Blockchain-Enabled Security Framework for Cloud-Based Sensor Systems with Deep Learning-Driven Attack Classification
Sensors · 17.08.2026 · 26 (16) · 5198
DOI: 10.3390/s26165198 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 17.08.2026. Verfasst von Naveed Ahmad und 2 weiteren Autor:innen. Beteiligte Einrichtung: College of Computer & Information Sciences, Prince Sultan University, Riyadh 12435, Saudi Arabia. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Prince Sultan University.
Abstract (Originalsprache)
Cloud-integrated sensor and Internet of Things (IoT) systems enable scalable data storage, processing, and intelligent monitoring, but their distributed nature exposes network-flow and host-level data to unauthorized access, tampering, and cyberattacks. This study proposes a Weighted Symmetric Hashed Blockchain (WSHB) framework that integrates mutual-information-based feature weighting, deep-learning-based attack classification, AES-256-GCM authenticated encryption, cryptographic hashing, and permissioned-ledger logging. The framework was evaluated independently using HIKARI-2021 for network-based intrusion detection and ADFA-LD for host-based intrusion detection. A transparent comparative evaluation was conducted against LSTM and CNN–RNN baselines using identical data splits, preprocessing settings, input representations, hyperparameter-search budget, and repeated initialization seeds. The proposed WSHB-DNN classifier achieved macro-F1 scores of 0.6837 on HIKARI-2021 and 0.7914 on ADFA-LD, showing the strongest overall classification performance among the evaluated models. The cryptographic and permissioned-ledger components provide confidentiality protection, record-level authentication, integrity verification, and tamper-evident logging for confirmed attack-event records. These results demonstrate the potential of WSHB as a reproducible framework for attack classification and secure event logging in cloud-integrated sensor environments.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- College of Computer & Information Sciences, Prince Sultan University, Riyadh 12435, Saudi Arabia.
- School of Cyber Science and Engineering, Wuhan University, Wuhan 430072, China.
- School of Computing, Electrical and Applied Technologies, Unitec Institute of Technology, Auckland 1025, New Zealand.
ORCID-Profile: Naveed Ahmad, Yue Cao, William Liu
Förderung
- Prince Sultan University
Zitieren
APA
Ahmad, N., Cao, Y., & Liu, W. (2026). A Blockchain-Enabled Security Framework for Cloud-Based Sensor Systems with Deep Learning-Driven Attack Classification. Sensors, 26(16), 5198. https://doi.org/10.3390/s26165198
Vancouver
Ahmad N, Cao Y, Liu W. A Blockchain-Enabled Security Framework for Cloud-Based Sensor Systems with Deep Learning-Driven Attack Classification. Sensors. 2026;26(16):5198. doi:10.3390/s26165198
BibTeX
@article{Ahmad2026,
author = {Ahmad, Naveed and Cao, Yue and Liu, William},
title = {A Blockchain-Enabled Security Framework for Cloud-Based Sensor Systems with Deep Learning-Driven Attack Classification},
journal = {Sensors},
year = {2026},
volume = {26},
number = {16},
pages = {5198},
doi = {10.3390/s26165198}
}
EUVIMED