A challenging problem on observer-based, integrated fault diagnosis and fault-tolerant control for linear systems subject to actuator faults and control input constraints is studied in this paper. An adaptive observer approach is used for the joint state-fault magnitude estimation, and a feedback controller is designed to stabilize the closed-loop system without violating the actuator limits in the presence of actuator faults. Matrix inequality conditions are provided for computation of design parameters of the observer and the feedback controller, and the admissible initial conditions and estimation errors are bounded by invariant ellipsoidal sets. The design results are closely related to the fault magnitude and variation rate, and a necessary condition on the admissible fault magnitudes dependent on the control limits is directly obtained from the design process. The proposed design framework allows a direct application of the pole placement method to obtain stabilization results. To improve the system performance, a nonlinear programming-based optimization algorithm is proposed to compute an optimized feedback gain, whereas the one obtained by pole placement can be taken as an initial feasible solution for nonlinear optimization. Numerical studies with two flight control systems demonstrate the effectiveness of proposed design techniques.
Aircraft Inertial Measurement Unit Fault Diagnosis Based on Optimal Two-Stage UKF
