Model-Free Adaptive Control of the Failing Heart Managed by Mechanical Supporting Devices

Abstract

Left ventricular assist devices (LVADs)—mechanical circulatory pumps—have been used as a therapeutic option for patients who have progressed to the advanced stage of left-sided heart failure (HF), which is a chronic condition that the heart is unable to pump sufficient blood out of the left ventricle as it should. This surgically implanted device supports the native heart to pump blood from the heart to the remainder of the body to keep the patient alive as a temporary therapy until a donor's heart is available or as a destination therapy. However, the pump speed of the LVAD is typically set in a constant mode and cannot be freely changed. To promote the clinical use of LVADs as a long-term treatment option, a physiological control system is required to adaptively adjust the pump speed in response to time-varying blood demand. To this end, a model-free adaptive control (MFAC) algorithm is developed in this work, which estimates the control parameter over time using the data of manipulated and controlled variables (i.e., end-diastolic pressure of left ventricle and pump speed, respectively). In addition, since hemodynamic parameters in the cardiovascular system change over time due to short-term autonomic nerve regulation, we further considered baroreflex regulation to improve control performance; baroreflex regulation is one of the key factors of human's hemostatic mechanisms. The performance of the physiological controller is investigated and validated with computer simulations, which shows that the LVAD can respond to constant and time-varying blood demand.