Design of a State Predictor

A state predictor design for deterministic discrete-time linear system is proposed. The structure of the predictor is determined from the discrete time state space model of the plant and consists of a feedforward gain and a feedback linear system. The boundedness conditions of the predicted states are analysed and do not necessarily require the plant stability. The state construction is based only on the knowledge of the plant outputs. The method is particularly suitable for the computation of the reference state trajectories required for the implementation of the State Space Model Predictive Control.