Model Predictive Control Using a Non-minimal State Space Form with an Integral-of-error State Variable

Abstract: This paper considers Model Predictive Control (MPC) using a Non-Minimal State Space (NMSS) form, in which the state vector consists only of directly measured system variables. Compared to previous research on NMSS/MPC, the new approach alters the state vector with the introduction of an explicit integral-of-error state variable to preserve steady state tracking of the reference input. The disturbance rejection properties of the controller are established and simulation examples presented. The development of this approach is motivated by the Proportional-Integral-Plus (PIP) control structure that has been successfully used in NMSS design for many years. The paper suggests that the tuning techniques developed for the latter case, can now be used for straightforward NMSS/MPC design.