Estimation of Unmanned Aerial Vehicle Dynamics in the Presence of Sensor/Actuator Faults

Abstract

The process noise covariance adaptation procedure (Q-adaptation) for unmanned aerial vehicle (UAV) state estimation is introduced and the robust adaptive Kalman filter (RAKF) algorithm with R- and Q-adaptations against sensor/actuator failures is proposed. Thus, the filter proves to be robust against faults and, even in case of sensor/actuator failure, keeps providing accurate estimation results. The performance of the proposed RAKF is investigated via simulations for the state estimation procedure of the UAV. The presented RAKF ensures that the parameter estimation system of the UAV is not influenced by sensor/actuator failures and, therefore, autonomous missions can be performed successfully.