Unmanned Aircraft Vehicle Pilot Identification with Machine Learning

This PhD thesis from Khalifa University investigates the application of machine learning for unmanned aircraft vehicle (UAV) pilot identification. The research focuses on capturing behavioral data from pilots during UAV flights, utilizing machine learning algorithms to classify and identify individual pilots. The study employs 20 datasets with various flight modes, testing 20 features against 22 classifiers, with Random Forest identified as the most effective. The results demonstrate high accuracy in pilot identification, even with single features. The thesis also explores Random Forest tree reduction to optimize processing power and memory usage. The work covers UAV security, drone biometrics, machine learning techniques, and the architecture of UAVs, providing a comprehensive analysis of pilot identification methods to prevent unauthorized access and enhance drone security.