Comparison of Joint Axis Estimation Methods Using Inertial Measurement Units

Area of Science:

• Biomechanics
• Human Motion Analysis
• Sensor Technology

Background:

• Accurate joint angle measurement is crucial for biomechanical analysis.
• Inertial Measurement Units (IMUs) are widely used for motion tracking.
• Existing methods for estimating joint axes using IMUs lack comparative analysis for different conditions.

Purpose of the Study:

• To compare the accuracy and stability of two representative joint axis estimation methods: Principal Component Analysis (PCA) and an optimization method.
• To evaluate method performance under ideal and non-ideal conditions with varying sensor orientations.
• To identify the most suitable method for robust joint axis estimation in human motion analysis.

Main Methods:

• Experimental comparison of PCA and optimization methods for joint axis estimation.
• Utilized a multi-degree-of-freedom apparatus for controlled motion generation.
• Tested various sensor attachment positions to simulate diverse orientations and motions.

Main Results:

• Both PCA and optimization methods showed high accuracy (approx. 4.6°) under ideal conditions.
• Accuracy significantly decreased (over 15°) under non-ideal conditions for both methods.
• PCA demonstrated superior performance with lower errors (3.7° ideal, 16.0° non-ideal) compared to the optimization method (5.5° ideal, 25.3° non-ideal).
• Optimization method convergence was sensitive to initial values and motion ratios.

Conclusions:

• Principal Component Analysis (PCA) is a more promising method for accurate and stable joint axis estimation using IMUs.
• Sensor attachment position and motion variability significantly impact the performance of joint axis estimation algorithms.
• PCA offers greater robustness and reliability for joint angle measurement in diverse biomechanical applications.