Summary

The thesis explores the use of unmanned aerial vehicles (UAVs), emphasizing their rising popularity due to their low cost and simplicity compared to manned aircraft, particularly in dangerous or remote areas. It focuses on ensuring UAV safety and reliability in sensitive airspace by detecting sudden and anomalous behaviors. The study investigates the use of Inertial Measurement Units (IMU) and vibration analysis for monitoring UAV health.

A Digital Twin is employed to simulate physical system operations, generating data to identify normal UAV operation modes. Real-time operational data is compared against this to detect anomalies, using TimeGAN, a GAN variant for time series data, and k-Means clustering algorithm.