Monitoring Obstructive Sleep Apnoea with Wearable Sensors | SURE Research Project
148Obstructive Sleep Apnoea (OSA) occurs when narrowing of the throat during sleep
causes partial or total blockages that reduce airflow, preventing sufficient oxygen from
reaching the brain. OSA causes tiredness and exhaustion, which can be severe enough to
make activities such as driving unsafe. Existing methods of diagnosis are relatively
impractical for the large number of undiagnosed OSA sufferers, as they involve overnight
stays in a sleep clinic.
This project sought to produce and evaluate an accessible, low cost method for
self-recording and remotely diagnosing OSA using a smartphone microphone, wireless
accelerometer, and wireless pulse oximeter. Over the course of five weeks, a smartphone
application for data collection using these methods was developed and a small group of
healthy and apnoeic subjects sampled. Using this data and the output of a snoring classifier
developed previously within the University of Sheffield Department of Computer Science,
experiments in training decision tree classifiers to predict whether a subject had OSA were
conducted. Candidates for data collection were limited, resulting in insufficient evidence
from which to thoroughly evaluate the effectiveness of combined audio and pulse oximetry
against these methods as used individually. However, the development of these techniques
and surrounding classification framework provide a valuable system through which more
data can be collected in the future.
Future work could continue research on incorporating motion sensors for more
robust classification of apnoeas, as well as experimenting with using heart rate data to
determine sleep stage in order to provide more accurate classification.
Full Report: https://drive.google.com/open?id=1dRV_pfBBQBV5iBno7K4XImgB_rnAVu8c
Support Me: https://www.patreon.com/HOCgaming
Hot Potato Podcast: https://www.youtube.com/channel/UC8Pj8DVyfePlud23DyxnPIQ
MUSIC:
"Inspired" Kevin MacLeod (incompetech.com)
Licensed under Creative Commons: By Attribution 3.0 License
http://creativecommons.org/licenses/by/3.0/