A Model Predictive Approach to Dynamic Control Law Design in Discrete-Time Uncertain Systems

摘要

A model predictive control (MPC) scheme is mainly developed in discrete-time uncertain systems. The control law contains a dynamic property in the proposed MPC. Hence, the MPC with a dynamic control policy is simply known as model predictive dynamic control (MPDC). To this end, a suitable matrix transformation is suggested to convert the MPDC problem into another optimization issue. Then, a systematic procedure based on linear matrix inequality (LMI) is addressed to the MDPC design. Hence, the MPDC synthesis is translated into an LMI minimization problem, which handles both constraints on the control inputs and plant outputs. The optimization problem can be numerically solved at each sample time through the well-known LMI solver. Then, the parameters of the dynamic controller would be automatically updated at each sample time. The method is applied in a discrete-time example to verify the effectiveness of the presented approach versus similar results.