Closed-Loop Fluid Resuscitation: In-Silico Study

Abstract

This thesis validates a control-oriented model of human circulatory dynamics associated with hemorrhage and fluid resuscitation, uses this model to design a series of closed-loop controllers for fluid resuscitation, and demonstrates the viability of these controllers by examining the performance of these controllers in an established model of human cardiovascular physiology. First, a recently developed control-oriented model of hemorrhage and fluid resuscitation was validated across diverse physiological conditions using an established model of human cardiovascular physiology, by employing a system identification procedure. Second, a series of closed-loop controllers were designed based on the nominal control-oriented model, including proportional control and proportional-derivative control based on the root locus analysis, as well as observer-based optimal state feedback control based on modern control theory. Third, the performance and robustness of the designed closed-loop controllers were validated using the established model of human cardiovascular physiology. The results suggested that model-based closed-loop control may be a viable alternative to today’s manual fluid infusion practices.