Funnel control for oxygenation during artificial ventilation therapy

Abstract

A patient with hypoxia requires immediate oxygen therapy to enhance oxygenation. Individual model including disease and parameter variation is needed, which limits all model-based approaches for the task of reference tracking control. In this article, a non-identifier based control algorithm, called funnel control, is proposed to solve this clinical problem in real practice. With this control technique, it requires less a priori knowledge of the complicated nonlinear and time-delay plant. The design of a funnel controller relies only upon an output feedback and the difficult control problem of such system is simplified by the design of a funnel boundary, which is a time-varying gain to force error dynamics inside this boundary. Funnel controllers with relative degree one and two are presented and simulated with the proposed cardiopulmonary system for low tidal volume, simulating the existing and abrupt change of atelectasis.