Design and Implementation of an Exogenous Kalman Filter for UAVs

Abstract: State estimations for nonlinear systems is often done using the Extended Kalman Filter (EKF), which has good performance in terms of minimum variance for the estimation error. However, global stability is not guaranteed, and the filter has a risk of becoming unstable. A recent paper proposes the eXogenous Kalman Filter (XKF), which combines the stability properties of a nonlinear observer and the performance of the Kalman filter. The combination of the two observers leads to better stability while the accuracy of the estimations is maintained. The topic of this thesis is the design and implementation of an XKF for the control system of an Unmanned Aerial Vehicle (UAV). The filter was simulated, and the performance was compared to the existing EKF. Once the algorithm for the filter was constructed, it was implemented in a real avionics system for UAVs and tested using real sensors. It was shown through simulations that the XKF indeed provides better stability properties for the state estimations compared to the EKF, without loss of accuracy.  

Examiner: Bo Wahlberg