Constraint fulfillment in control systems via predictive reference management

The problem of satisfying input and state-dependent inequality constraints in feedback control systems is addressed. The proposed solution is based on predicting the evolution of the constrained vector and, accordingly selecting on line the future reference based on both the current state and the desired set-point changes. An analysis is presented so as to establish stability and offset-free properties of the method when embodied in an LQ regulated system. Finally, simulations are used to evaluate the achievable performance.