Continuous monitoring of premature babies remains essential given that nearly 15 million children are born prematurely each year and are hospitalized for several weeks in incubators that act as thermoregulated environments in neonatal intensive care. It is precisely in this context that the latest technologies are increasingly being used to monitor more efficiently but also discreetly.
A new set of algorithms has been developed by Rohan Joshi, a researcher at the Eindhoven University of Technology. The algorithms serve to prioritize the most important alarms.
In addition, the same researcher has also developed a non-contact sensor that monitors the breathing and movement of premature babies. The latter type of monitoring promotes the prevention of infections, sometimes even severe ones.
Joshi used the machine learning technique to analyze different amounts of data related to child monitoring signals. In this way he managed to create an automated system that eliminates less serious alarms, for example some related to heart rate, sometimes a little less regular, or to the saturation of oxygen in the incubator, allowing alarms to be given more priority critically.
Normally, this type of monitoring requires the use of adhesive electrodes on the skin of children, a system that is sometimes too invasive and at least uncomfortable for children. To improve the comfort of newborns, the scientist proposes this alternative monitoring system practically without contact thanks to a pressure sensor, similar to a film, that can be inserted into the incubator mat. Several tests have already shown, as reported in the press release, that this sensor is as reliable as the standard electrodes.
- Method for thawing surfaces in seconds is developed by scientists - November 3, 2019
- New material created that can withstand very high pressures - October 18, 2019
- Carbon dioxide emitted by factories used to make soap - October 11, 2019