Machine Learning for Device Identification with Application in Aviation Cybersecurity

The rapid expansion of the Internet of Things (IoT) introduces security risks, particularly in aviation, where rogue devices can compromise flight data and disrupt air traffic control (ATC). This chapter surveys machine learning (ML)-based techniques for detecting and identifying rogue IoT devices, covering device-specific pattern recognition, deep learning, and abnormal behavior detection. We explore cryptographic and non-cryptographic countermeasures, emphasizing physical-layer identification using unique wireless transmitter signatures. Special attention is given to aviation security, focusing on Automatic Dependent Surveillance–Broadcast (ADS-B) vulnerabilities. We review radiometric fingerprinting, context-based verification, and flight dynamics-based authentication for mitigating spoofing and hijacking threats. The chapter highlights the need for multi-layered security solutions integrating ML, anomaly detection, and behavioral analysis to enhance aviation IoT resilience.