Levels of activity identification & sleep duration detection with a wrist-worn accelerometer-based device

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Health Informatics

Background:

• Accurate physical activity (PA) monitoring is crucial for health assessment.
• Wrist-worn accelerometers offer a non-invasive method for objective PA measurement.
• Distinguishing between rest and sleep is important for sleep quality assessment.

Purpose of the Study:

• To develop and validate a model for classifying physical activity levels using wrist-worn accelerometer data.
• To implement an algorithm for distinguishing sleep duration from rest periods.
• To explore the potential of the proposed model for home-based healthcare applications.

Main Methods:

• Collected accelerometer data from 10 healthy subjects.
• Developed a model to categorize PA into rest/sleep, sedentary, light, moderate, and vigorous states.
• Implemented an activity-based algorithm to differentiate short rest periods from sleep.

Main Results:

• Successfully categorized physical activity levels based on accelerometer data.
• Demonstrated the ability to distinguish between different intensities of physical activity.
• Developed a functional algorithm for sleep duration detection.

Conclusions:

• The proposed accelerometer-based model effectively identifies various physical activity levels.
• The developed algorithm can differentiate sleep from rest, improving sleep monitoring.
• This technology holds promise for remote patient monitoring and personalized healthcare.