Robust model based predictive controller design for NCSs with long time varying delays in feedback channels

An off-line optimization, on-line execution robust model predictive control algorithm (RMPC) is designed for networked control systems (NCSs) with long time varying delays in feedback channels. Invariant sets are predetermined via off-line optimization, and on-line controller gains are selected according to system states. Predicted values are used to compensate data dropouts. Simulations and comparisons show effectiveness of the designed controller.

1 Introduction

A closed-loop automatic control system implemented in the communication network, in which data are packed into small packets and transmitted via network, is named as Networked Control System (NCS) [1], see fig. 1. Figure 1 Structure of NCS