Sign in Subscribe
Data security

Adaptive Personalized Federated Learning for Cybersecurity in Internet of Vehicles

Strategic Insights

— 1 min read

Research on cybersecurity in Internet of Vehicles (IoV) has reached a critical juncture with the increasing adoption of Connected and Autonomous Vehicles (CAVs) within intelligent transportation systems. The University of Ha'il has proposed a novel approach to address the limitations of centralized Intrusion Detection Systems (IDS), leveraging Adaptive Personalized Federated Learning (APFed) and Lightweight Depthwise Convolutional Bottleneck Network (LDwCBN) to ensure privacy-preserving, resource-efficient, and accurate intrusion detection. The proposed method outperforms state-of-the-art federated IDS approaches, achieving accuracy improvements of up to 5% and significant gains in precision, recall, and F1-Score.

Key Takeaways:

  • The University of Ha'il has proposed a novel approach to address the limitations of centralized IDS in IoV, leveraging APFed and LDwCBN for privacy-preserving, resource-efficient, and accurate intrusion detection.
  • The proposed method, Adaptive Personalized Federated Learning with Lightweight Depthwise Convolutional Bottleneck Network, achieves accuracy improvements of up to 5% over FedAvg and FedProx models.
  • The system ensures precision improvements of up to 4%, recall improvements of up to 3%, and F1-Score improvements of up to 4%.
  • The research uses extensive evaluations on benchmark datasets, including CIC-IDS2017, CSE-CIC-IDS2018, Car-Hacking, and CAN-Train-Test.
  • The proposed method is designed to operate under heterogeneous and non-IID data conditions.
  • The authors, Shahad Almansour, Kusum Yadav, Lulwah M. Alkwai, Norah Saleh Alghamdi, Wattana Viriyasitavat, and Gaurav Dhiman, aim to enhance model personalization and generalization through fine-grained adaptive updates and dynamic weight fusion.

Statistics:

  • 5% accuracy improvement over FedAvg and FedProx models
  • Precision improvements of up to 4%
  • Recall improvements of up to 3%
  • F1-Score improvements of up to 4%

Sources:

  • Adaptive personalized federated learning with lightweight depthwise convolutional bottleneck network for novel intrusion detection system in internet of vehicles. Scientific Reports, 2025,15(1):1-14.
  • Journal of Engineering. November 3, 2025; p 2717.

Read more