Constrained MIMO nonlinear predictive control based derivate-free state estimators

This paper presents a constrained multivariable model predictive control for discrete nonlinear systems based on Divided Difference filters (DDFs) which are derivate free state estimators. The estimators are used simultaneously for both nonlinear model predictive control (NMPC) formulation by determining the state variation at each time step, and for a state feedback closed loop implementation. The proposed combination presents a high performance in set-points tracking and control law stability when implemented in presence of low and high noisy measurements compared to the NMPC based Extended Kalman Filter (EKF).