Constrained dynamic compensation with model predictive control for tracking

Abstract

This article extends the benefits of model predictive control for the linear dynamic compensators commonly designed for aerospace systems. The traditional control structures often used in output tracking are synthesized, and the proposed model predictive control scheme modifies the plant input and reference to be tracked, creating artificial demands and control corrections to guarantee the state and satisfaction of the output constraint. The novel predictive controller, with a related reference governor, is presented with assured closed-loop stability, convergence and recursive feasibility for attainable piecewise constant setpoints along the system operation. The modification of the feasible demands and control inputs enlarges the domain of operation of the constrained closed-loop system. The elimination of the quadratic programming solvers running during real-time operation is also presented to reduce the computational burden. The simulation results using a fighter trainer model are finally shown, illustrating the benefits of the proposed technique.