A HYBRID DEEP LEARNING APPROACH FOR ANOMALY DETECTION IN INTELLIGENT TRANSPORTATION SYSTEMS

Abstract

Vehicular Ad Hoc Networks (VANETs), which facilitate smooth and effective communication between automobiles and infrastructure, are an essential part of intelligent transportation systems. Threats like hostile nodes, malfunctioning sensors, and cyberattacks make it difficult to guarantee their security and dependability. Integrated Forest Autoencoder with Naïve Bayes Anomaly Detection Model (IFAE-NBADM) is a novel hybrid anomaly detection models that this research suggests as solutions to these problems. The suggested approach captures intricate patterns and isolates anomalous occurrences in high-dimensional vehicle data by combining an isolation forest ensemble with an autoencoder-based feature learning mechanism. A Naïve Bayes classifier is used for probabilistic anomaly classification to further improve detection reliability and facilitate effective decision-making in the face of uncertainty. A synthetic VANET dataset that replicates real-world vehicle features like speed, mobility patterns, communication characteristics, and traffic dynamics is used to assess the model. According to experimental data, the IFAE-NBADM model maintains low computing overhead while achieving improved performance in terms of accuracy, precision, recall, and F1-score. The results verify that the suggested IFAE-NBADM framework offers a reliable, scalable, and efficient anomaly detection approach, greatly enhancing VANET security and guaranteeing dependable vehicular communication.