Classifying sitting, standing, and walking using plantar force data.

Area of Science:

• Biomechanics
• Occupational Health
• Wearable Technology

Background:

• Prolonged static weight-bearing at work is linked to increased plantar fasciitis (PF) risk.
• Objective measurement of workplace postures is needed to confirm weight-bearing's causal role in PF.

Purpose of the Study:

• To assess the accuracy and sensitivity of low-resolution plantar pressure measurements.
• To distinguish between different workplace postures (sitting, standing, walking).

Main Methods:

• In-shoe plantar pressure measurement system used in 8 healthy participants.
• Data resampled to simulate 24 on/off plantar force sensitive resistors.
• Machine learning algorithms (SVM, DT, DA, KNN) evaluated sensor classification accuracy.

Main Results:

• Support Vector Machines (SVM) and Decision Trees (DT) showed best classification accuracy.
• High accuracy (98.4-99.1%) achieved with as few as five or even a single sensor.
• Central forefoot and midfoot sensor locations were most critical for classification.

Conclusions:

• On/off plantar pressure measurements in the midfoot and central forefoot accurately classify workplace postures.
• This forms the basis for a low-cost tool to objectively measure sedentary workplace behaviors.
• Enables further research into the relationship between workplace posture and plantar fasciitis.