New Principles and Adequate Robust Control Methods for Artificial Pancreas

摘要

The current work is a short review of the research results obtained by the Biomedical Engineering laboratory of the Control Engineering and Information Technology Department, Budapest University of Technology and Economics (BME) in the field of automatic control of Type I diabetes mellitus, and mainly is focused on the PhD dissertation written by the first author [1]. The topic focuses on modeling formalisms and implementation of robust control algorithms for optimal insulin dosage in case of Type I diabetes patients. Regarding modelling concepts, an extension of the modified Bergman minimal model, the analytical investigation of the high complexity Sorensen-model and model synthesis of a novel molecular-based model are described. From robust control methods implementation firstly, the minimax method is applied and is described that its limitations can be spanned using Grobner basis. Secondly, the graphical interpretation of the Hinf method under Mathematica program is extended with disturbance rejection criteria. Thirdly, an LPV (Linear Parameter Varying) type robust control method is described for the high complexity Sorensen-model; hence it is possible to deal directly with the non-linear model itself. Finally, actual research tasks are summarized related to the previously mentioned topics and further research directions are formulated.