State space MPC using state observers

Abstract

The article is focused on state observers and their usage in model predictive control (MPC). The observers are used to track and reconstruct states of a model of a controlled system. Linear time-invariant (LTI) state space models are used in the article because this type of models is often used in different MPC techniques If the states of the controlled system are immeasurable a state observer (filter) is used to calculate current states in each control step. The paper is especially focused to finite impulse filters (FIR) as these filters do not require knowledge of initial state - contrary to infinite impulse response (IIR) filters. Different observers are tested and compared with proposed filters based on quadratic and linear programming. Filters were used in a very noisy environment to evaluate filter robustness. Then MPC using promising filters was applied to a three tank model Amira DTS 200.