Fault diagnosis system with statistical attribures for fixed wing unmanned aerial vehicle

Thanks to technological developments, the role of humans has gradually decreased in the aviation sector where risk factor are high as in many other sectors. For this reason, the demand and development studies for unmanned aerial vehicles (UAVs) that minimize the human factor have gained great importance and momentum in our country as well as all over the world. As with all electronic systems, a malfunction in these vehicles can cause serious consequences. Malfunctions that may occur in vehicles cause loss of effectiveness in the control surface that enable the basic movements of the aircraft or even the complete loss of the control surface. For this reason, detecting malfunctions in advance or immediately and subsequently controlling the system smoothly has become a major research and development subject. This thesis focuses on diagnosis of sudden malfunction that cause loss of effectiveness in actuators in fixed wing unmanned aerial vehicles. Using a model available in MATLAB Simulink partial fault were created in the actuator block occurring after the transient regime. To examine these fault outputs, numerical data received from the sensors on the vehicle were examined in terms of various statistical attributes. During the data analysis, a threshold value was determined by examining the healthy data in terms of statistical parameters and by comparing it with the faulty state data, a 'fault' was detected in cases where the threshold value was exceeded.