Gas exchange models have the potential to individualize therapy for patients requiring mechanical ventilation and aid in its automation. This is especially true in neonates, who face an increased risk for respiratory complications and ventilator-induced lung injury. Low-order, low-complexity models are commonly used for controller design due to their computational efficiency. Controller performance is affected by the model’s prediction accuracy, necessitating thorough evaluation. This work evaluates the prediction accuracy of a two-compartment gas exchange model with animal trial data. Individual parameter sets have been estimated within physiological bounds.
Michael et al. (Wed,) studied this question.