Phattaradanai KiratiwudhikulPornchai ChanyagornMahidol University2018-11-092018-11-092014-02-17Proceedings of 2014 International Conference on Smart Computing, SMARTCOMP 2014. (2014), 289-2942-s2.0-84946531367https://repository.li.mahidol.ac.th/handle/20.500.14594/33678© 2014 IEEE. Pre-term infants - less than 37 weeks gestational age - usually had immature lungs' development, which resulted of poor oxygen saturation in red blood cells. A blood oxygen saturation level was measured in percent of Peripheral capillary oxygen saturation (SpO2). Medical doctors needed to order an oxygen therapy to maintain SpO2 of the infants between 90-95% while SpO2 of normal infants is 99-100%. Oxygen therapy was a procedure to stimulate lung functions and to maintain life. A registered nurse (RN) was responsible for adjusting levels of a fractional of inspired oxygen (FiO2) from 21% to 100% which was a proportion of oxygen gas provided to the infants periodically. In real situation, the adjustment could only be made as often as every 20-30 minutes, which might not be adequate. This caused ineffectiveness of an oxygen therapy and result in a longer hospital stay. A critical error of this adjustment could also cause blindness due to oxygen toxicity or dead due to hypoxia. This research was to develop a reliable embedded system that allowed automatically control of FiO2 according to an order of SpO2 by medical doctors. As a result, risks of oxygen toxicity and hypoxia could be minimized. The system also allowed medical doctors to use recorded data for future care planning in oxygen therapy.Mahidol UniversityComputer ScienceConference PaperSCOPUS10.1109/SMARTCOMP.2014.7043870