Fault-Tolerant Predictive Control with Sensor Faults
Résumé
In this paper, a Fault-Tolerant Model Predictive Controller (FTMPC) is developed for linear Variable Parameter Systems (LPVs) subject to sensor faults and input constraints. First, an augmented state-space model contains both state variables and estimation error is used to synthesize a robust predictive controller while an observer is designed to estimate state variables and sensor faults. In addition, the optimization problem of the proposed design is formulated to reject disturbances by merging the disturbance estimates in the prediction model. For design purposes, the proposed optimization along with all constraints is expressed in terms of linear matrix inequalities (LMIs). Furthermore, sufficient stability conditions are derived using a Lyapunov approach to assure the convergence of the proposed method. Next, the control problem is obtained via solving these linear matrix inequalities (LMIs) constraints. Finally, the proposed approach performance is tested by controlling an electric circuit.