Extended model predictive control: robust controller with matching condition

Abstract

This paper describes a new robust controller based on Model Predictive Control (MPC). MPC is one of the useful control methods because it enables to design time response intuitively and to consider the constraint of state and input. However, MPC does not have robustness against modeling errors and disturbances due to the use of a predictive model without considering these uncertainties. Robust Model Predictive Control (RMPC) by using a Linear Matrix Inequality (LMI) or H-infinity control theory has been proposed to improve the robust performance. The theoretical complexity loses the characteristics of the MPC, such as the intuitive design of a controller. On the other hand, Sliding Mode Control (SMC), which is a simple theory, has high robustness compared to H-infinity control theory and RMPC when a system satisfies a matching condition. In addition, the Integral Sliding Mode Control (ISMC) that developed from SMC can constrain the system to the reference model and the response of the target can be made approximate to the response of the ideal system even under the influence of modeling errors and disturbances. It is noted that a controller based on SMC should restrain chattering of control input when applying the controller to a real system. We propose the extended MPC (EMPC) that utilize a matching condition to have robustness against uncertainties. The feature of the proposed control method is not only robustness but only easy tuning of the controller. The effectiveness of the proposed control system is verified by numerical simulation in comparison with the conventional controllers that are PID, ISMC, and MPC.

Keywords

• robust control
• sliding mode control
• integral sliding mode control
• model predictive control