Robust Active Fault Tolerant Tracking Control for Constrained NLPV System Subject to Sensor and Actuator Faults
Résumé
In this work, Sensor-Actuator Fault (SAF) estimations and accommodation for discrete-time constrained Nonlinear Parameter Varying (NLPV) descriptor system subject to input and state constraint, with unmeasurable premise variables and depending on the acting faults are investigated. By using L2 Proportional Integral (PI) observer and non-quadratic Lyapunov function, efficient conditions to design the observer are obtained via a linear matrix inequalities formulation. Then, based on the estimation of the SAF, an efficient Fault Tolerant Tracking Controller (FTTC) can be realized via SAF compensation. The gains of both observer and FTTC are computed by solving a set of LMI constraints in single step, with less conservative results and improving trajectory tracking performances. Robustness against external disturbances is analyzed using the L2 performance index to attenuate its effect on the tracking error, estimation error and faulty system. Finally, a robot 2DoF model is used to highlight the performances of the developed method.