Modeling, fault detection, and stabilization of quadrotor unmanned aerial vehicle with rotor thrust deviation fault

This study deals with the thrust deviation of quadrotors during a flight accident. This is a novel structural fault wholly ignored in previous studies, and its kinematic and dynamic equations are developed using the Newton law and differentiating the angular momentum equation. The thrust deviation fault has been detected using the governing model of the faulty system, Thau observer, and calculation of residual value. For evaluation of the performance of the proposed fault detection method, it is assumed that limited disturbances are applied to the system before the fault occurrence. Besides, to land the quadrotor affected by the deviation of one of its rotors, the ordinary sliding mode controller is replaced by a modified one. This controlling procedure consists of two steps. First, the deviation angles are identified using the Extended Kalman Filter. Then, these angles are used to calculate the appropriate inputs that are the functions of rotational speed of rotors. The combination of this modified control unit and model-based fault detection method has improved the robustness of the system in the presence of the structural fault and has maintained the system's controlling performance in fault-free situations.