Model Predictive Control Based on Multivariable Disturbance Observer

This paper proposes a multivariable disturbance observer based on the state space model. The relationship between the input sequence and the output sequence is analyzed through the state space model and described by a numerical matrix. Then the equivalent disturbance can be estimated by the numerical matrix which replaces the inversion of the transfer function matrix. Subsequently, a coefficient matrix and a filter matrix are adopted to make a good tradeoff between disturbance rejection performance and control effort. Simulation results verify the rationality and effectiveness of the proposed multivariable disturbance observer.