Ensemble Learning-based Intrusion Detection System for Autonomous Vehicle

Autonomous vehicles (AVs) are a potential technology for improving safety and driving efficiency in intelligent transportation systems (ITSs). However, AVs are subject to various cyber-attacks, comprising denial-of-service, spoofing, sniffing, and cross-site scripting. To overcome the security issues in AV, this paper proposed an intelligent framework that impersonates the intrusion detection system (IDS) that intelligently classifies malicious and non-malicious data requests of AVs. For that, we utilized ensemble-based machine learning classifiers, such as decision tree, random forest, extra tree, XGboost, K-nearest neighbor, and support vector machine (SVM), to train them on different attacks and simultaneously use their learning for classification. The proposed model is bifurcated into different phases of machine learning, like data collection, pre-processing, and prediction. Finally, we evaluate the ensemble models using different evaluation metrics, such as accuracy, precision, recall, and f1-score. XGBoost outperforms other classifiers in terms of accuracy, i.e., 98.57%, which benefits from attaining a high detection rate and low computational cost at the same time for the AV systems.