Application of Extremum Seeking Control to Turbodynamic Blood Pumps

Abstract

Ventricular assist devices now clinically used for treatment of end-stage heart failure require responsive and reliable control to accommodate the continually changing demands of the body. However, due to the varying physiologic conditions and the limited use of the sensors to detect hemodynamic load and suction, it is difficult to control pump speed appropriately. The author introduces an adaptive pump speed controller to provide maximum cardiac perfusion while avoiding ventricular suction. The controller is based on an extremum seeking control (ESC) algorithm and a slope seeking control (SSC) algorithm, which find and track unknown and moving peak points of a prescribed cost function. The controller was validated with in vivo data using time-averaged diastolic pump flow as the cost function for ESC/SSC. Initial results demonstrate the successful application of ESC/SSC as a physiologic pump speed controller.