[This article belongs to Volume - 56, Issue - 02, 2024]
Gongcheng Kexue Yu Jishu/Advanced Engineering Science
Journal ID : AES-11-11-2024-39

Title : ANOMALY DETECTION AND IN VEHICULAR AD HOC NETWORKS (VANETS) USING INTEGRATED ANOMALY DETECTION MODEL (IFAE-IADM)
Mrs.V.Saranya, Dr.P.Nirmaladevi

Abstract :

Vehicular Ad Hoc Networks (VANETs) are integral to modern intelligent transportation systems, enabling vehicles to communicate for enhanced road safety, traffic management, and efficient routing. However, the reliability and security of VANETs are often challenged by various anomalies, such as cyberattacks, rogue vehicles, or sensor failures, which can disrupt normal operations. To address this, an efficient and real-time anomaly detection system is crucial. This research work proposes a hybrid anomaly detection framework Integrated Anomaly Detection Model (IFAE-IADM) that combines Isolation Forest and Autoencoder models through a majority voting mechanism to identify anomalous patterns in VANET data. The Isolation Forest model detects anomalies by isolating data points with fewer partitions, while the Autoencoder neural network identifies irregularities through reconstruction error. Both models are trained on a synthetic dataset simulating vehicular movement, speed, communication parameters, and data traffic. The framework is evaluated using performance metrics such as accuracy, precision, recall, and F1-score, with the combined model achieving improved results over individual models. This work demonstrates the effectiveness of hybrid anomaly detection in VANETs and provides valuable tools for maintaining network security and reliability.