Application of data fusion to neonate oxygenation control

Premature and newborn infants often have underdeveloped lungs and, as a result, benefit from additional oxygen. Unfortunately, too much oxygen risks inducing blindness as well as other toxicity phenomena. Since the infant's oxygen uptake and delivery vary widely and are constantly changing, automatic control of the supplied oxygen is desirable. Many of the issues and technologies of tactical control are applicable to this problem: real-time control, expert system knowledge, multi-sensor corroboration and data fusion, failure identification and artifact rejection, etc. This paper explores these issues as well as the results of current PSU research in neonate oxygen control.<
>