Patient-Adaptive Feedback Controllers for Heart Assist Devices

University Of Pittsburgh, Pittsburgh PA

Investigators

Abstract

Patient-Adaptive Feedback Controllers for Heart Assist Devices A. Project Summary Heart transplant candidates often wait long periods (300 days or more on the average) before a suitable donor heart becomes available, and many of these candidate die while awaiting heart transplantation. Consequently, the medical community has been placing increased emphasis on the use of mechanical devices that can substitute for, or enhance, the function of the natural heart. A Ventricular Assist Device (VAD) is such a device. This device if often used as a bridge to mechanically support a patient while awaiting heart transplantation, or while their natural heart recovers. But the ultimate use of this device is as a long-term iocurelr for heart failure. In either case, the goal of a VAD is to provide the patient with as close to a normal lifestyle as possible. The goal for all concerned is to allow patients to return home and return to the workforce. An important challenge facing the increased use of VADs is the development of an appropriate controller for these devices. The latest generation of VADs, based on turbo-dynamic pumps, requires a controller that can adjust the speed of the rotor (pump impeller) to meet the circulatory demand of the patient. Therefore, in addition to being robust and reliable, such a controller should be able to adapt to the daily activities and the physiological changes of the patient. In this project we propose to investigate the theoretical and fundamental issues associated with the development of such a controller. More specifically, (1) we will continue our ongoing work on the development of a generalized model of a rotary blood pump that, when coupled to existing model of the human cardiovascular system, will serve as a simulation model for further development; (2) we will develop a robust optimal feedback control algorithm that provides the flexibility to incorporate various sensor inputs as available, and (3) we will implement the control algorithm into a PC platform and test its responsiveness in a mock circulating flow loop. The principal functions of this feedback controller will be: (1) to determine and maintain appropriate cardiac output, subject to constraints imposed by the device and the patients condition including avoiding ventricular suction; (2) to identify system variables and parameters to monitor the patient's condition, indicating when the patient's cardiovascular status is improving or deteriorating and detecting component failures; (3) to recognize different operating conditions and data environments, responding to changes in the signals available for measurement from the patient and the assist device; (4) to monitor the control strategy (algorithm) itself to avoid uncontrollable situations and inappropriate or dangerous control actions; and (5) to provide a fail-safe mode of operation by entering a preset open-loop operating mode when a hardware or algorithmic failure occurs. The primary intellectual merit of the proposed research will be to increase our understanding of fundamental technical issues related to the control of non-conventional complex systems, such as the VADs, that involve both electromechanical as well as physiological components. A truly interdisciplinary approach, with expertise coming from the electrical, mechanical, and biomedical engineering fields as well as the medical profession, is needed to develop the mathematical framework needed to be able to effectively control such systems. The broader impact of the proposed research will primarily be in improving the quality of life for cardiovascular patients awaiting heart transplantation. Any improvement in the existing technology of the ventricular assist devices will have a tremendous effect on the physical condition and ultimately the recovery of patients suffering from congestive heart failure. It is hoped that this study, leading to the development of a patient adaptive feedback controller, will provide an opportunity for these patients to leave the hospital, return home, and re-enter the work force while awaiting for a donor heart. Another important impact of this project is in training engineering graduate (and a few undergraduate) students from the electrical, mechanical, and biomedical, departments to work together as a team in this highly interdisciplinary field. This project will provide an opportunity for students to apply what they have learned in the control and signal processing courses to a realistic (but not very well defined mathematically) problem that has a tremendous societal impact.