Model-based Healthcare Applications at Óbuda University Levente Kovács * , Johanna Sápi * , György Eigner * , Tamás Ferenci * , Péter Szalay ** , József Klespitz * , Balázs Kurtán * , Miklós Kozlovszky*, Dániel A. Drexler ** , Péter Pausits * , István Harmati ** , Zoltán Sápi *** , Imre J. Rudas * * Óbuda University, Budapest, Hungary {kovacs.levente, ferenci.tamas, kozlovszky.miklos}@nik.uni-obuda.hu, {sapi.johanna, eigner.gyorgy, klespitz.jozsef}@phd.uni-obuda.hu, pausits.peter@hok.uni-obuda.hu, balazs.kurtan@gmail.com, rudas@uni-obuda.hu ** Budapest University of Technology and Economics, Budapest, Hungary {szalaip, drexler, harmati}@iit.bme.hu *** Semmelweis University, Budapest, Hungary sapi.zoltan.dr@gmail.com Abstract— Following the tradition established at previous IEEE INES conferences, this paper presents the results achieved by the Physiological Controls Group of Óbuda University in the field of physiological control last year (2013). The application target was focused on diabetes (artificial pancreas), tumor (antiangiogenic therapy), and hemodialysis (control of peristaltic pumps). I. INTRODUCTION Physiological control studies and applies identification and control strategies in order to understand and help automated treatment of various diseases or injuries of the human body. As a particular field of biomedical engineering [1], nowadays the key target of physiological control is to efficiently create and support personalized solutions for disease management in a model-based way. The aim of the recently established Physiological Controls Group of the Óbuda University is to develop solutions in the mentioned area working together in control applications with the Department of Control Engineering and Information Technology of the Budapest University of Technology and Economics and the Semmelweis University from Budapest. The article briefly summarizes the newest (last year, 2013) results obtained in the field, following the tradition established at the previous IEEE INES 2011-2013 conferences [2-4]. Control results in three different fields are presented: • cancer, where the uncontrolled growth of abnormal cells [5] are proposed to be handled by control engineering methods. As a result, not only the tumor volume can be minimized, but also the dosage of the inhibitor creating a cost- efficient therapy. For this our idea follows robust control methodology. • diabetes mellitus, where the problem of artificial pancreas [6] is investigated. Newest results in robust control are presented in order to deal with high uncertainties caused by imprecise descriptions of the glucose metabolism. • hemodialysis, where direct blood particle separa- tion is done with external devices, i.e. peristaltic pumps maintain the fluid flows [7]. However, the transfer depends on their loaded tube segment; hence, the fluid balance should be controlled. II. ROBUST CONTROL DEVELOPMENT FOR CANCER TREATMENT P. Hahnfeldt et al. created a dynamic model for tumor growth for antiangiogenic therapy [8]. The aim of antiangiogenic therapy is to prevent tumors from forming new blood vessels, because without angiogenesis tumor growth is inhibited [9]. The original model was modified; the simplified model takes into account continuous infusion therapy [10]. Fig. 1 presents the simulated tumor growth with and without antiangiogenic therapy. Without treatment the growing tumor size reaches lethal state (17340 mm 3 ), while with antiangiogenic therapy the tumor size can be reduced. Hence, robust control strategy has been investigated. The proposed control structure is presented in Fig. 2. K is the two-degrees of freedom controller which consists of two parts: K r is the feedforward branch and K y is the feedback branch. Multiplicative uncertainty (G unc ) has been taken into account in order to handle differences between the nominal model and the real system. – 183 – 9th IEEE International Symposium on Applied Computational Intelligence and Informatics • May 15-17, 2014 • Timişoara, Romania 978-1-4799-4694-5/14/$31.00 ©2014 IEEE