Generic and autonomous system for airborne networks cyber-threat detection

Cyber-security on airborne systems is becoming an industrial major concern that arises many challenges. In this paper, we introduce a generic security monitoring framework for autonomous detection of cyber-attacks on airborne networks based on unsupervised machine learning algorithm. The main challenge of anomaly detection with unsupervised techniques is to have an accurate detection since they tend to produce false alarms. After evaluating the suitability of the One Class SVM, we propose some hints to improve detection accuracy of the monitoring framework by collecting information from the airborne architecture.