Enhancing Security in Intelligent Transportation Systems: A Sequential Prediction Approach for Misbehavior Detection in V2V Using Machine Learning

In Intelligent Transportation Systems (ITS), vehicle-to-vehicle (V2V) communication improves traffic efficiency and road safety but is vulnerable to cyber threats like Denial-of-Service (DoS) and spoofing attacks. Traditional intrusion detection struggles with data imbalance and computational complexity. This study evaluates Machine Learning (ML) models for misbehavior detection using the VeReMi extension dataset, implementing a sequential classification approach to detect attackers and classify attack types. XGBoost and LightGBM are perfect for real-time ITS deployment according to the research since they reach 99.99% accuracy with minimal computational cost. Deep Learning (DL) models like CNN, LSTM, and GRU also do rather well—above 99.4% accuracy. These results highlight how well ML protects V2V communication.