Model-Based Step Length Estimation Using a Pendant-Integrated Mobility Sensor.

Area of Science:

• Biomechanics
• Gait Analysis
• Wearable Sensors

Background:

• Step length is crucial for gait analysis, but difficult to measure with unobtrusive sensors like IMUs.
• Long-term gait monitoring requires accurate, low-cost methods for spatial parameter estimation.

Purpose of the Study:

• To develop and validate a model-based algorithm for estimating step length using pendant-integrated IMUs.
• To assess the algorithm's performance at slow walking speeds relevant to elderly or impaired individuals.

Main Methods:

• A two-step model-based approach was developed: first, estimating the center of mass (CoM) vertical displacement, then estimating step length.
• The algorithm was compared against a known accelerometry-based method and validated using force plate data.
• Focus was placed on slow walking speeds (1-4 km/h).

Main Results:

• The proposed algorithm achieved 9.3% coefficient of variation (CoV) for vertical CoM displacement with 1.5 mm RMSE.
• Step length estimation showed a relative prediction error below 10% for walking speeds of 2-4 km/h.
• The model-based approach outperformed the existing accelerometry-based prediction method.

Conclusions:

• The developed model-based algorithm provides accurate step length estimation from IMU data, especially at slow walking speeds.
• This method enhances the potential for unobtrusive, long-term gait analysis and stability assessment.
• The findings support the use of IMUs for clinical gait monitoring in specific populations.