Reliable recognition of lying, sitting, and standing with a hip-worn accelerometer

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Physical Activity Measurement

Background:

• Hip-worn accelerometers are common for physical activity (PA) assessment.
• Traditional methods struggle with accurate classification of sedentary behaviors and body postures (lying, sitting, standing).

Purpose of the Study:

• To develop a novel method for precise body posture classification using hip-worn triaxial accelerometer data.
• To evaluate the method's performance in free-living conditions against a thigh-worn accelerometer.

Main Methods:

• Developed an Angle for Posture Estimation (APE) method using Earth's gravity vector and upright walking as references.
• Tested on 30 healthy adults performing various posture tasks (lying, sitting, standing, walking).
• Validated against a thigh-worn accelerometer in free-living conditions.

Main Results:

• Optimal APE cut-points achieved high sensitivity and specificity for distinguishing between lying/sitting and sitting/standing.
• High agreement (89.2-90.4%) between hip- and thigh-worn accelerometers in identifying sedentary periods.
• Walking proved a valid reference for posture determination.

Conclusions:

• The proposed APE analysis offers accurate and specific classification of daily lying, sitting, and standing times.
• This novel method enhances the reliability of hip-worn accelerometers for detailed physical activity and posture monitoring.