EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Next-Gen IoT Security: Deep Learning-Based Detection of RPL Attacks in Mobile Converged Networks

Muhammad Nadeem Ghouri, Ghufran Ahmed, Ahmad Sami Al-Shamayleh, Shahbaz Siddiqui, Muhammad Maaz, Adnan Akhunzada

IEEE Open Journal of the Communications Society · 2026 · 7 · 902-924

DOI: 10.1109/ojcoms.2026.3650763 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Open Journal of the Communications Society, 2026. Verfasst von Muhammad Nadeem Ghouri und 5 weiteren Autor:innen. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: open access, Qatar National Library.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

IoT networks based on RPL are still susceptible to routing manipulation attacks like the Version Number Attack (VNA) and Worst Parent Attack (WPA) that may strongly compromise the stability and reliability of the network. In order to counter such threats, this paper presents three hybrid deep learning models CNN-GRU, CNN-LSTM, and GRU-GRU based on the joint mechanisms of spatial forwarding behavior and temporal routing dynamics of RPL traffic. Instead, a feature augmentation pipeline is used to increase the discriminative power and then grid search hyperparameter optimization is used and it increases detection accuracy. The models are tested on four actual RPL datasets of large and small network topology. It has been proven in experiments that all hybrid models have a good detection performance at different network scale with VNA reaching up to 98 percent F1-score and WPA reaching up to 92 percent F1-score. The models achieve a high accuracy of over 92% on VNA and over 83% F1-score on WPA even in complex large-network conditions, which is substantially high compared to traditional machine learning and single-architecture deep learning baselines. Comparative performances also show that CNN-LSTM offers the best accuracy and GRUGRU is the best in recall and CNN-GRU is light and high sensitivity, which can be deployed in a wide variety of IoT applications. These findings validate the fact that hybrid spatial-temporal deep learning architectures provide a scalable and potent solution to securing the RPL-based IoT networks against advanced routing attacks.

Förderung

  • open access
  • Qatar National Library

Zitieren

APA

Ghouri, M. N., Ahmed, G., Al-Shamayleh, A. S., Siddiqui, S., Maaz, M., & Akhunzada, A. (2026). Next-Gen IoT Security: Deep Learning-Based Detection of RPL Attacks in Mobile Converged Networks. IEEE Open Journal of the Communications Society, 7, 902-924. https://doi.org/10.1109/ojcoms.2026.3650763

Vancouver

Ghouri MN, Ahmed G, Al-Shamayleh AS, Siddiqui S, Maaz M, Akhunzada A. Next-Gen IoT Security: Deep Learning-Based Detection of RPL Attacks in Mobile Converged Networks. IEEE Open Journal of the Communications Society. 2026;7:902-924. doi:10.1109/ojcoms.2026.3650763

BibTeX
@article{Ghouri2026,
  author = {Ghouri, Muhammad Nadeem and Ahmed, Ghufran and Al-Shamayleh, Ahmad Sami and Siddiqui, Shahbaz and Maaz, Muhammad and Akhunzada, Adnan},
  title = {Next-Gen IoT Security: Deep Learning-Based Detection of RPL Attacks in Mobile Converged Networks},
  journal = {IEEE Open Journal of the Communications Society},
  year = {2026},
  volume = {7},
  pages = {902--924},
  doi = {10.1109/ojcoms.2026.3650763}
}
Kostenlos starten →