EUVIMEDEUVIMED entdecken
Open Access · CC BY-SA 4.0

Edge-Sage: GraphSAGE Framework for Decentralized Intrusion Detection in Edge IoT Networks

Farah Alaa A.Hassan, Asraa A. Abd Al-Ameer

Journal of Robotics and Control (JRC) · 25.12.2025 · 6 (6) · 2896-2915

DOI: 10.18196/jrc.v6i6.28114 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Robotics and Control (JRC), online erschienen am 25.12.2025. Verfasst von Farah Alaa A.Hassan und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

IoT networks that are edge based pose difficulties to intrusion detection because of unstable communication links, non-IID traffic distributions and adversarial threats. This pa per presents Edge-Sage, a federated learning architecture that combines both the GraphSAGE-based flow modeling and trust weighted and sparsified gradient aggregation in which clients only send top-k gradients, multiplied by trust scores, to detect intru sions in decentralized edge networks. Edge-Sage compares with 16 recent baselines on BoT-IoT and ToN-IoT datasets consisting of a variety of federated and centralized systems. It has 97.3% accuracy and 96.7% F1-score and decreases the cost of communi cation by 83.2%. The comprehensive testing of the label-flip and DDoS attacks, non-IID conditions has proven that the framework is healthy, supported by DropEdge filtering and gradient clipping. The experimental procedures will consist of 5-fold cross-validation, statistical significance testing, and adversarial recall as a resilience measure. These findings indicate that trust-based aggregation and sparse updates enhance robustness and scalability of intrusion detection of distributed IoT networks.

Zitieren

APA

A.Hassan, F. A., & Al-Ameer, A. A. A. (2025). Edge-Sage: GraphSAGE Framework for Decentralized Intrusion Detection in Edge IoT Networks. Journal of Robotics and Control (JRC), 6(6), 2896-2915. https://doi.org/10.18196/jrc.v6i6.28114

Vancouver

A.Hassan FA, Al-Ameer AAA. Edge-Sage: GraphSAGE Framework for Decentralized Intrusion Detection in Edge IoT Networks. Journal of Robotics and Control (JRC). 2025;6(6):2896-2915. doi:10.18196/jrc.v6i6.28114

BibTeX
@article{AHassan2025,
  author = {A.Hassan, Farah Alaa and Al-Ameer, Asraa A. Abd},
  title = {Edge-Sage: GraphSAGE Framework for Decentralized Intrusion Detection in Edge IoT Networks},
  journal = {Journal of Robotics and Control (JRC)},
  year = {2025},
  volume = {6},
  number = {6},
  pages = {2896--2915},
  doi = {10.18196/jrc.v6i6.28114}
}
Kostenlos starten →