Automatic Control of Fraction of Inspired Oxygen in Neonatal Oxygen Therapy using Fuzzy Logic Control

Premature babies of less than 37 weeks gestation might require oxygen therapy as an integral part of treatment and respiratory support. Because of their under-developed lungs, these so-called "preemies" might contract respiratory distress syndrome (RDS). To treat RDS, neonatal oxygen therapy is administered, where controlled oxygen gas is measured as a fraction of inspired oxygen ($FiO_2$). However, exposure to high oxygen content during long treatment could cause oxygen intoxication, which might cause permanent blindness due to retinopathy of prematurity (ROP), whereas insufficient oxygen exposure could cause severe hypoxia. A doctor would use oxygen saturation ($SpO_2$) data and prescribe a dose of $FiO_2$ to maintain $SpO_2$ within a suitable range. One objective is to maintain $SpO_2$ within the acceptable range using $FiO_2$ that is as low as possible. Adjustment of $FiO_2$ would normally be done by nurses every 15 to 30 minutes, which might not be safe in many situations. An error in $FiO_2$ adjustment during a manual procedure could be as large as +/- 2.5%. This paper presents a system that can determine an $FiO_2$ value suitable to the current $SpO_2$ and that automatically adjusts $FiO_2$ with an error clearance of +/- 0.25%.