Fault detection methods for unmanned airvehicles

Various methods have been developed for the early detection of problems that may occur especially on the control surfaces with the rapid increase in the usage area of unmanned aerial vehicles. These fault detection methods play an important role in ensuring system security, durability and reliability for unmanned aerial vehicles (UAVs). This thesis demonstrates the application of Extended Kalman Filter and Unscented Kalman Filter methods for nonlinear systems for different fault scenarios such as locking of control surfaces such as rudder, elevator and aileron of unmanned aerial vehicles or random input values to the system and etc. It compares the sensor results with the estimation results and presents the estimation of the relevant situations. Firstly, the nonlinear state space system of the aircraft system is formulated with the help of dynamic and kinematic equations. Aerodynamic parameters of unmanned aircraft are determined by methods such as wind tunnel experiments, flight tests and parameter comparison methods, necessary assumptions were made and the system is taken from mathematical environment and modeled in Simulink environment. Then, different error scenarios are created for different control surfaces such as rudder, elevator and aileron in order to apply the filters and applied to the system model. Afterwards, the related motion cases are examined from the pitch, roll and yaw movements affected by the faulty control surface and the error determinations are observed. Finally, measurement and filter estimations are interpreted. The system is modeled in MATLAB and Simulink environment. The performances of the methods are interpreted using simulation results. Keywords: Kalman filter, Unmanned Air Vehicles, EKF, UKF, Nonlinear systems, Modeling, Simulation, Matlab, Simulink, Flight control systems, Actuator fault detection, State space modeling