Networked Predictive Control for MIMO process control with model uncertainty corrections

In this paper, a multi-variable Network Predictive Control (NPC) method which is able to actively compensate for the effects of the network uncertainties using model based predictions is introduced. The proposed method is design for MIMO systems which are widely used in large-scale industrial applications. In practical systems, the accurate mathematical models are very difficult get, which could negatively affect the NPC control performance. To address this problem, a prediction correction method is proposed. It uses the previous prediction errors to correct the future ones, which increases the accuracy of the prediction and eventually improves the control performance. To verify the effectiveness of the proposed methods, a simulation example is also given at the end of the paper.