A Privacy-Preserving Artificial Intelligence Framework for Intrusion Detection in Cognitive Radio Wireless Body Area Networks
Journal of Computer Science Innovations and Research · 10.11.2026 · 2 (2) · 10-27
DOI: 10.69055/jcsir.v2i2.21 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Computer Science Innovations and Research, online erschienen am 10.11.2026. Verfasst von Zeki Saeed Taufeeq und 2 weiteren Autor:innen. Die Arbeit nennt 17 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Cognitive Radio Wireless Body Area Networks (CR-WBANs) are emerging as a foundational technology for next-generation healthcare, uniting the dynamic spectrum access of cognitive radio with the continuous physiological monitoring of body area networks. The same openness that makes CR-WBANs efficient also widens their attack surface. Beyond the data-plane threats common to the Internet of Medical Things (IoMT) — data injection, spoofing, and denial of service — the cognitive radio layer introduces physical-layer threats unique to opportunistic spectrum access: Primary User Emulation Attacks (PUEA), Spectrum Sensing Data Falsification (SSDF), and jamming. Because CR-WBAN traffic carries life-critical and privacy-sensitive medical data, a successful intrusion can directly endanger patient safety. This paper proposes a privacy-preserving, artificial-intelligence-based Intrusion Detection System (IDS) tailored to the three-tier CR-WBAN architecture. The framework couples two complementary detection stages — a physical-layer module that identifies cognitive-radio-specific attacks from spectrum-sensing features, and a network-layer deep neural network that detects IoMT data-plane attacks from fused network-flow and biometric features — and trains them collaboratively across distributed WBANs using federated learning, so that raw patient data never leaves the edge. A feature-attribution layer makes alerts interpretable for clinical staff. The pipeline is implemented and evaluated end-to-end on data following the WUSTL-EHMS-2020 specification for the network layer and a controlled cognitive-radio simulation for the physical layer. Fusing biometric and network features improved detection for every model; the proposed network-layer detector achieved an accuracy of 0.9496, the highest precision (0.8367) and the lowest false-alarm rate (0.0208) among the compared methods, and its federated version stayed within roughly one percentage point of the centralized model while sharing no raw data. The physical-layer detector separated benign, PUEA, SSDF, and jamming traffic with an overall accuracy of 0.9945.
Zitieren
APA
Taufeeq, Z. S., Al-Hamami, A. H., & Abass, G. Y. (2026). A Privacy-Preserving Artificial Intelligence Framework for Intrusion Detection in Cognitive Radio Wireless Body Area Networks. Journal of Computer Science Innovations and Research, 2(2), 10-27. https://doi.org/10.69055/jcsir.v2i2.21
Vancouver
Taufeeq ZS, Al-Hamami AH, Abass GY. A Privacy-Preserving Artificial Intelligence Framework for Intrusion Detection in Cognitive Radio Wireless Body Area Networks. Journal of Computer Science Innovations and Research. 2026;2(2):10-27. doi:10.69055/jcsir.v2i2.21
BibTeX
@article{Taufeeq2026,
author = {Taufeeq, Zeki Saeed and Al-Hamami, Alaa Hussein and Abass, Gufran Yassin},
title = {A Privacy-Preserving Artificial Intelligence Framework for Intrusion Detection in Cognitive Radio Wireless Body Area Networks},
journal = {Journal of Computer Science Innovations and Research},
year = {2026},
volume = {2},
number = {2},
pages = {10--27},
doi = {10.69055/jcsir.v2i2.21}
}
EUVIMED