Actuator fault detection method of quadrotor UAV based on dual channel inertial sensors

Abstract Aiming at the problem of actuator failure of quadrotor unmanned aerial vehicle (UAV), a novel fault detection method based on dual channel inertial sensors is proposed. Firstly, the actuator partial failure fault model is introduced into the UAV dynamic model. Then, a double channel observer is designed with the angular velocity of the rolling channel and the pitch channel as the state variable. The angular velocity estimation errors of the two channels are calculated and compared with the estimation error threshold. Based on this result, the fault channels are initially located while the actuator faults are identified. Finally, according to the angular velocity control characteristics of the two channels, an accurate fault location algorithm is designed to determine the location of the faulty rotor. Two groups of simulation flight experiments were carried out in MATLAB/SIMULINK environment. During the simulation process, the algorithm proposed by this paper was compared and analyzed with the H-/L∞ algorithm proposed in the recent literature. The results show that the algorithm in this paper can not only achieve precise fault location, but also detect faults quickly. Compared with the H-/L∞ algorithm, the detection speed of this algorithm has increased by more than 80%. The real flight experiment results of the quadrotor UAV prototype verify the effectiveness of the proposed algorithm for actuator fault detection in actual flight scenarios.