Intelligent Anomaly Detection for Secure Vehicular Networks Using a Hybrid Learning Framework

Abstract

Vehicular Ad Hoc Networks (VANETs) play a significant role in Intelligent Transportation Systems (ITS) by enabling real-time communication among vehicles and between vehicles and roadside infrastructure. However, the dynamic and decentralized nature of VANETs makes them vulnerable to various security and reliability threats, including malicious nodes, sensor failures, abnormal communication behavior, and cyberattacks. These challenges highlight the need for efficient and intelligent anomaly detection mechanisms capable of identifying abnormal activities with minimal computational overhead. To address these challenges, this study proposes an Integrated Forest Autoencoder with Naïve Bayes Anomaly Detection Model (IFAE-NBADM) for reliable anomaly detection in VANET environments. The proposed framework integrates an Isolation Forest-based ensemble mechanism with an Autoencoder-based feature learning approach to identify complex patterns and detect anomalous behavior in high-dimensional vehicular data. The learned representations are subsequently processed using a Naïve Bayes classifier, which provides probabilistic anomaly classification and improves decision-making under uncertain network conditions. The effectiveness of the proposed model is evaluated using a synthetic VANET dataset incorporating representative vehicular characteristics, including vehicle speed, mobility behavior, communication parameters, and traffic conditions. Experimental results demonstrate that the proposed IFAE-NBADM framework achieves improved anomaly detection performance in terms of accuracy, precision, recall, and F1-score, while maintaining relatively low computational overhead. The integration of unsupervised anomaly detection, deep feature learning, and probabilistic classification enables the framework to effectively identify diverse abnormal patterns in VANET communication. The findings indicate that IFAE-NBADM provides a robust, scalable, and computationally efficient solution for enhancing VANET security and reliability, thereby supporting dependable communication and safer intelligent transportation environments