Hemodynamic Control using Direct Model Reference Adaptive Control - Experimental Results

Regulation of hemodynamic parameters is a challenging problem, due to inter- and intra-patient variability, drug interactions, evolving patient condition, and in the case of simultaneous control of mean arterial pressure and cardiac output, tight coupling between the two variables. This requires a controller that can adapt to conditions as they evolve, and without the need to tune the controller for each individual patient. To address these problems, direct model reference adaptive control (DMRAC) is extended by incorporating mechanisms to handle rate and saturation constraints, allowing a range in the value of a regulated variable instead of fixed setpoint tracking, and a design method that directly considers time delay uncertainty. In this paper we present experimental results using extended DMRACon canines, for a variety of conditions. These results demonstrate the feasibility of using DMRAC, with the advantage of not requiring prior model identification for the individual patient.