Adaptive Actuator Fault Detection and Identification for UAV Applications

In this paper a two step algorithm, based on a simplified linear model of the longitudinal aircraft dynamics is introduced, which allows for reliable detection and identification of actuator faults in small unmanned aerial vehicles (UAVs) with unknown aerodynamic parameters. In the first step, a Kalman filter is used to estimate the pitch rate from elevator deflections. The filter innovation is then used to adapt the elevator control power of the simplified linear model according to the current flight condition. Both, filter innovation and adaptive parameter are then used for detection and identification of different fault modes. The method has been successfully tested with real flight data from a small fixed wing UAV with unknown aerodynamic parameters.