Ear-Worn Inertial Sensors Can Predict Gait Metrics and Reconstruct Vertical Ground Reaction Force Curves During...

Area of Science:

• Biomechanics
• Wearable Technology
• Sports Science

Background:

• Ear-worn wearables are popular among runners for entertainment.
• Modern earbuds contain inertial sensors for user interaction.
• Existing wearables for gait analysis (insoles, ankle/sacrum units) have limitations in adoption and usability.

Purpose of the Study:

• To leverage ear-worn wearable technology for running gait analysis.
• To develop a method for capturing running mechanics outside a laboratory setting.
• To assess the accuracy and usability of ear-worn devices for gait monitoring.

Main Methods:

• Thirty healthy participants ran on an instrumented treadmill and force plates.
• A gait event detection algorithm was developed using head-mounted sensor data (vibrations and motion).
• A regression model was created to predict vertical ground reaction force waveforms.

Main Results:

• The algorithm achieved a 4.8% mean absolute percentage error on temporal metrics.
• Scalar ground reaction force metrics showed a 9.0% mean absolute percentage error.
• The model achieved an 11.1% normalized root mean square error on unseen participants for the vertical ground reaction force curve.

Conclusions:

• Ear-worn devices offer a viable and user-friendly approach to running gait analysis.
• The accuracy of ear-worn devices is comparable to specialized equipment for temporal and kinetic gait parameters.
• This technology paves the way for widespread, accessible running gait monitoring.