"AI-Driven Secure Vehicular Networks: Intelligent Threat Detection and Trust-Aware Defense for V2X Systems"
Abstract
"Vehicular Ad Hoc Networks (VANETs) and Vehicle-to-Everything (V2X) systems are fundamental to intelligent transportation ecosystems. However, the highly dynamic topology, decentralized communication, and heterogeneous vehicle-edge infrastructure expose these networks to cyber threats such as Sybil attacks, spoofing, replay attacks, distributed denial-of-service, and malicious route manipulation. Traditional cryptographic defenses alone are insufficient against adaptive and zero-day threats.This paper proposes an AI-driven secure vehicular network framework that integrates federated learning, graph neural networks, and deep reinforcement learning for privacy-preserving intrusion detection, trust-aware routing, and autonomous attack mitigation. The architecture enables vehicles, roadside units, and edge servers to collaboratively train distributed models without sharing raw data, preserving privacy while improving detection generalization. A graph-based trust engine models inter-vehicle communication behavior, while DRL dynamically optimizes defense actions under changing network conditions. Experimental evaluation demonstrates superior performance in detection accuracy, false alarm reduction, attack response latency, and secure routing efficiency compared with conventional ML and signature-based IDS baselines"
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.