Quadrotor Control and Actuator Fault Detection: LQG Versus Robust $H_{-}/H_{\infty}$ observer

This paper proposes a robust model based observer for actuators fault detection and diagnosis (FDD), applied to an unmanned aerial vehicle: quadrotor. The observer is designed aiming at maximizing the fault to residual sensitivity by using the H_- index properties, and minimizing the H_∞ norm for worst case exogenous signals attenuation. Fault detection is formulated as LMI feasibility problem for minimizing a cost function based on a trade-off between fault sensitivity and robustness against disturbances. The effectiveness of the proposed observer is demonstrated through its application to a quadrotor and comparing the results with that from a Kalman filter.