Development of Deep Learning Model to Detect Cyber-Attacks within Vehicular Networks
R. D. Sandakelum, V. H. Liyanage, Pranieth Chandrasekara, V. Logeeshan, Sisil Kumarawadu, Chathura Wanigasekara · 2024
As the prevalence of connected and autonomous vehicles (CAVs) continues to grow over the years, it brings forth the potential for various anomalies within the vehicular network. To ensure that communication systems for vehicles are safe and dependable on the roads, a security scheme with accurate anomaly detection is needed. The available approaches for anomaly detection detect only specific cyber-attack types or not much accurate and efficient. As effective solution for this issue, a modified study was accompanied in this research paper to evaluate the effectiveness of Deep neural network (DNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Bi Long Short-Term Memory (BiLSTM) and Bi Gated Recurrent Unit (BiGRU) based recurrent neural network (RNN) models for identifying and classifying anomalies in vehicular networks. In this research, detection and classification models gained an accuracy of over 95%, and a lower latency in misbehavior detection a significant enhancement compared to the available approaches.