Combining model-free and model-based angle of attack estimation for small fixed-wing UAVs using a standard sensor suite

We propose to estimate steady and turbulent wind velocities and aerodynamic coefficients of a fixed-wing Unmanned Aerial Vehicle (UAV) by using frequency separation as well as kinematic, aerodynamic and wind models combined in an Extended Kalman Filter (EKF). With these estimates it is possible to calculate the angle of attack and the magnitude of the airspeed. Avoiding the need for prior knowledge of UAV parameters, the proposed method utilizes only sensor information that is part of a standard sensor suite, which consists of a Global Navigation Satellite System (GNSS), an Inertial Measurement Unit (IMU) and a pitot-static tube, and attitude information obtained from these sensors. An observability analysis shows that attitude changes are necessary during the initialization phase and from time to time during the flight. Simulation results indicate that, with typical sensor accuracy, the estimates are close to the reference values of the aerodynamic coefficients and wind velocities and is capable of estimating the Angle of Attack with an Root Mean Square Error (RMSE) of 0.33°, the Sideslip Angle with an RMSE of 3.21° and the airspeed with an RMSE of 0.23 m/s.