Sleep Position Classification using a Single Abdominal Accelerometer

Citation Author(s):
Dieu
Vu
Phenikaa University
Submitted by:
Dieu Vu
Last updated:
Wed, 06/12/2024 - 11:47
DOI:
10.21227/evc4-vq42
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

Study on Sleep Positions using a Wearable Device

This device constantly collects data about acceleration in three directions (tri-axial) 50 times a second (50 Hz). It has several components: a microcontroller for gathering and sending data (ESP8266), a battery (lithium-ion), a sensor for measuring acceleration (ADLX345 accelerometer), and a protective case made of plastic. The device can also store data temporarily on a microSD card in case the wireless connection is lost.
We designed a special device worn on the belly to track a person's movements during sleep. This device collects data constantly (50 times per second) and can store information temporarily if the wireless connection drops out. It then sends the data wirelessly to a user-friendly smartphone app.

Fifteen volunteers (average age 20.5 years) participated in the study. They had a variety of body sizes and shapes, as shown by their height, weight, and body mass index (BMI) measurements.

The volunteers were trained on how to identify 12 different sleep positions. They wore the device while lying in each position for one minute. The data was carefully labeled with the participant's ID and the sleep position. This resulted in 180 data files, each containing one minute of information (3,000 data points).

Comments

for edge AI prototype usecase

Submitted by Fatima Rani on Tue, 09/17/2024 - 05:02