Estimating Vertical Ground Reaction Force during Walking Using a Single Inertial Sensor.

Area of Science:

• Biomechanics
• Machine Learning
• Wearable Technology

Background:

• Vertical ground reaction force (vGRF) is crucial for analyzing gait, musculoskeletal injuries, and prosthesis evaluation.
• Current vGRF estimation methods typically require force plates, limiting real-world applications.
• There is a need for portable, non-invasive methods for gait analysis in everyday environments.

Purpose of the Study:

• To develop and validate a machine learning model for estimating vGRF from single inertial measurement unit (IMU) data.
• To assess the accuracy of vGRF and its peak parameters estimation using IMU data.
• To enable real-time vGRF monitoring in non-laboratory settings.

Main Methods:

• Employed a random forest model to estimate vGRF from IMU data collected from lower limb locations (foot, shank, thigh).
• Utilized data from nine volunteers walking on a force plate-instrumented treadmill at various speeds.
• Validated model performance against gold-standard force plate measurements for intra- and inter-participant accuracy.

Main Results:

• The model achieved high accuracy for vGRF estimation, with correlation coefficients of 1.00 (intra-participant) and 0.97 (inter-participant) for the shank location.
• Accurate estimation of passive and active vGRF peak timing and magnitude was achieved with minimal error (e.g., 0.02 BW difference).
• The developed system demonstrates the feasibility of using a single IMU for real-time vGRF monitoring.

Conclusions:

• Vertical ground reaction force (vGRF) can be accurately estimated using a single inertial measurement unit (IMU) and machine learning algorithms.
• This approach facilitates the development of portable wearable systems for real-life gait monitoring.
• The findings support advancements in musculoskeletal injury analysis, gait abnormality detection, and prosthetic evaluation.