An Autonomous Intrusion Detection System for Ethernet-Based Avionics Communication Bus

Cyber security has become a foremost concern for the latest generation of aircraft. While the improved connectivity of airborne elements has enabled new efficiencies and capabilities, it has also exposed the avionics communication network to targeted cyber-attacks. These attacks pose a major challenge to the security of aircraft network and may compromise flight-critical operations. Furthermore, the probability of new improvised attacks is constantly growing with the introduction of modern state-of-the-art systems and services. In order to address this threat, we propose an anomaly-based network intrusion detection system (NIDS). This paper presents an IDS framework which includes the monitoring and acquisition of network traffic data from an Ethernet-based avionics communication bus in real-time. Furthermore, the attack traffic was simulated offline to imitate the real threat scenario. This manuscript evaluates the effectiveness of different machine learning (ML) algorithms for the detection of cyber-attacks.