Adaptive regulation of arterial gas pressures by using robust adaptive control scheme

A robust adaptive control scheme is investigated and implemented to regulate arterial oxygen pressure and carbon dioxide pressure independently at their desired levels by automatically adjusting two control inputs, the inspired oxygen concentration and the respiratory frequency, in accordance with continuously monitored transcutaneous oxygen and carbon dioxide pressures. A least-squares scheme using multiple models with different dead time and adjustable parameters can effectively determine the linearized first-order input-output model, including uncertain parameters and dead time. The proposed adaptive model-matching algorithm includes a supervisory controller and a robust controller with a Smith predictor to minimize the sensitivity to modeling errors, variations, and disturbances in the controlled subject. An animal experiment has shown that the proposed control algorithm can be easily implemented in an ordinary artificial ventilator.