Effect of sensor number and location on accelerometry-based vertical ground reaction force estimation during walking

Area of Science:

• Biomechanics
• Biomedical Engineering
• Orthopedics

Background:

• Knee osteoarthritis (KOA) is a leading cause of disability, with lower extremity loading influencing its onset and progression.
• Current technology for optimizing lower extremity loading in rehabilitation is limited.
• The peak vertical component of the ground reaction force (vGRF) is crucial for KOA symptom management.

Purpose of the Study:

• To determine the optimal number and placement of accelerometers for accurate vGRF estimation during walking.
• To assess how different loading conditions affect vGRF estimation accuracy.
• To identify sensor configurations for monitoring and potentially intervening in KOA-related gait abnormalities.

Main Methods:

• Twenty participants walked on a force-instrumented treadmill under controlled and biofeedback-guided loading conditions.
• Five accelerometers were placed on the waist, shanks, and feet.
• Machine learning models were trained to predict vGRF from accelerometer data, evaluating various sensor combinations.

Main Results:

• A single waist-mounted accelerometer achieved the most accurate peak vGRF estimation (4.4% body weight error) using a neural network.
• The waist-only configuration successfully differentiated between normal and overloaded gait conditions.
• Incorporating foot or shank accelerometers decreased accuracy, especially during overloading.

Conclusions:

• A single waist accelerometer is sufficient for accurately estimating walking vGRF in healthy individuals.
• This finding supports the development of simplified wearable systems for monitoring gait loading and guiding KOA rehabilitation.
• Future systems for KOA management may require minimal sensor setups for effective biofeedback and monitoring.