Fault detection of a VTOL UAV using Acceleration Measurements

This paper proposes an actuator fault detection algorithm for vertical take-off and landing (VTOL) unmanned aerial vehicle (UAV), based on acceleration signals provided by a high-rate inertial measurement unit. The proposed algorithm is based on time and frequency-domain analysis of acceleration signals and features selection techniques in order to identify the most relevant frequencies of the spectrum and time-domain features involved in the occurrence of the fault. The testing of the algorithm is focused on the damage of a propeller blade and the effects of such damage during vertical take-off and landing, where all the lift is generated by the propellers. The algorithm is tested on real data acquired on the Songbird VTOL UAV [1]. The experimental results prove the possibility of using acceleration signals for diagnostic purposes on VTOL UAVs.