Quaternion-based robust fault-tolerant control of a quadrotor UAV

This paper deals with the active fault-tolerant control (FTC) of a quadrotor in the presence of actuator faults. Generally, it is assumed that the fault has been detected, isolated or identified, and then, one proceeds with a reconfigurable controller. This paper presents a complete active fault-tolerant control system with both fault detection and diagnosis (FDD), and FTC. The proposed FTC design is based on gain-scheduling control in the framework of structured H∞ synthesis. The scheduled gains are parameterized as polynomial functions of the loss of control effectiveness of the quadrotor actuators, estimated by a two-stage Kalman filter. Then, the MATLAB-based function systune tunes the polynomial coefficients to meet the robustness and performance requirement. In comparison with other FTC systems that use switching mechanisms, this smooth self-scheduled controller allows one to avoid undesirable transient phenomena during the controller reconfiguration process. Numerical simulations performed on an underactuated six degrees of freedom quadrotor nonlinear model show the effectiveness of the proposed FTC in accommodating different levels of actuator faults. To avoid singularities associated with Euler angles, unit quaternion representations are used to describe the rotational motion of the quadrotor.