An adaptive controller for noisy pressure controlled ventilation

Abstract

There is a growing interest in the use of variable ventilation and pressure controlled ventilation (PCV). However, the combination of these approaches as “noisy PCV” would require a mechanical ventilation system that adapts to the respiratory system mechanics. In this work, we evaluated a new control system based on an adaptive least mean squares approach, which automatically tunes the pattern of the driving pressure during PCV to achieve a desired variability pattern of tidal volume (VT). The controller was tested during numerical simulations, applying step changes in respiratory system mechanics and in mechanical ventilation settings. The time needed to converge (t_c) to the desired VT variability pattern after each change, and the difference in minute ventilation ($\rm \triangle$MV) between measured and target pattern of VT during t_c were determined. The numerical simulations of the new controller resulted in: 1) t_c< 30 respiratory cycles, except when underestimation of the apparent elastance of the respiratory system (E^*) was > 25 %; 2) t_c only minimally influenced by E^*; 3) larger t_c when E^* was not correctly estimated; 4) absolute value of $\rm \triangle$MV < 22.2 %. The new noisy PCV controller had a satisfactory performance and could prove interesting for mechanical ventilation practice.