Robust model-based predictive control of a ventilation machine using LMI-techniques

In a co-operation between the Institute of Automatic Control at the Aachen University of Technology and the Dräger Medizintechnik GmbH, Lübeck, a well-known developer and manufacturer of respiration devices, a digital feedback control strategy for the pressure-based ventilation was developed. However, the achieved controller is not robust enough against parameter variations (e.g. different patients) of the controlled process. Because of known courses of the desired value a new technique has been attempted. This new approach was developed by Kothare et al. (see [Kot96]) and allows the synthesis of a robust model-based predictive control (MPC) law, using linear matrix inequalities (LMIs). This technique incorporates explicitly the plant uncertainty descriptions in the problem formulation and is robustly stabilising. In this paper the controlled process is presented. After giving an overview of the robust constrained MPC approach some simulation results are shown. To sum up some future work is outlined.