Abstract (EN):
Several applications require humans to be in high-altitude environments, whether for recreational purposes, like mountaineering or light sport aviation, or for labour, as miners. Although in these conditions the monitoring of physiological variables is, per se, of interest, the direct correlation of these variables with altitude itself is not usually explored towards the development of decision-support systems and/or critical event alarms. This paper proposes two neural networks approaches to assess and explore this correlation. One, based on dynamic SISO models, estimates physiological variables using the aircraft pressure altitude as input. A second approach uses static MISO networks to classify the flight stage (and therefore the altitude variation) from physiological variables. Both models were developed and validated using real data acquired in hypobaric chamber tests simulating a real flight. The good results obtained indicate the viability of the approach. Copyright
Language:
English
Type (Professor's evaluation):
Scientific
No. of pages:
6