Automated Drone Fault Detection Approach in Thrust Mode State

In recent years, there has been considerable interest in drones due to their flexibility and decreasing price. Drones are made up of electrical and mechanical components, and failure in these components might deteriorate the drone’s performance and, worse, lead it to crash. In this study, fault detection approaches based on raw vibration and acoustic parameters were introduced. Motor failure and loose screws are simulated to represent electrical and mechanical faults. Based on the raw data obtained from the MEMS sensors, motor failure can be successfully detected by both approaches. The mechanical fault is difficult to be detected by the raw data of both approaches as the amplitude values recorded are close to the normal drone condition. These approaches are suitable to be applied in detecting motor failure compared to mechanical faults.