The degrading performance of actuators in small unmanned aerial Vehicle (UAV) is often left unnoticed because it is masked by autopilot control. The faulty actuator will only be detected when the actuator has been severely damaged. If it occurs during flight, the UAV will be lost, same goes with the valuable data on the board. Usually, a pre-flight check is performed before each flight to ensure the overall condition of the UAV including the actuators. The actuator health deterioration is difficult to be recognized by visual inspection. This paper presents a method to detect the health of the actuator through the integration of Built-in Test System (BITE) using offline model estimation method. Least square regression estimation was performed on the healthy actuator for training data using fixed and increasing input signal. The fault is then simulated to test the training data accuracy in detecting actuator fault. An analysis is performed to show the advantage and disadvantage of each technique used. The estimation technique described was able to detect faulty actuator which could then be integrated with on-board health detection system in order to increase the reliability of the UAV.
All Science Journal Classification (ASJC) codes
- Computer Science(all)
- Health(social science)
- Environmental Science(all)