Robust output-feedback control of fluid resuscitation with absolute stability guarantee against uncertain physiology and therapeutic responsiveness

This paper presents a robust closed-loop control approach to fluid resuscitation in patients with hemorrhagic blood loss. A unique strength of the proposed approach is its robustness against uncertain and time-varying patient physiology and therapeutic effectiveness. First, we adopted an observer-based control architecture that can fulfill set point tracking and disturbance rejection objectives. Second, we determined the control gains to achieve adequate transient response performance using the linear quadratic regulator design. Third, we determined the observer gains as a solution to a set of linear matrix inequalities so that the overall closed-loop fluid resuscitation control system is (i) robust against the variability in patient physiology and (ii) absolutely stable against unknown therapeutic effectiveness. We demonstrated the initial proof-of-concept of the proposed approach by conducting rigorous in silico testing using a large number of physiologically plausible virtual patients, while ascertaining the absolute stability via the circle criterion analysis. The results suggested that the proposed approach to closed-loop control of fluid resuscitation is a promising option to advance automation of fluid resuscitation armed with stability against a large variability in patient physiology and therapeutic effectiveness as well as adequate performance in set point tracking and disturbance rejection.