Estimating foot inertial parameters: a new regression approach.

Area of Science:

• Biomechanics
• Human Motion Analysis

Background:

• Accurate estimation of foot inertial parameters (mass, center of mass, inertia tensor) is vital for biomechanical applications like inverse dynamics and stiffness measurements.
• Current scaling equations based on foot length and body mass may oversimplify complex foot geometry, leading to inaccuracies.

Purpose of the Study:

• To evaluate existing methods for estimating foot inertial parameters.
• To propose and validate a new, more accurate method for determining these parameters.

Main Methods:

• Utilized 3D surface scanning to create geometrical references of 34 right feet.
• Calculated reference inertial parameters directly from 3D models assuming uniform density.
• Compared direct calculations with estimates from traditional scaling and multiple regression equations.
• Developed and tested an alternative method using multiple non-linear regressions incorporating foot length and ankle width.

Main Results:

• Reference inertial parameters significantly differed from scaling predictions (up to 33% for mass, 16% for moments of inertia).
• Traditional scaling methods showed high standard errors for moments of inertia (up to 26%).
• The proposed method incorporating ankle width reduced standard errors to 8.7%, demonstrating improved accuracy for both genders.

Conclusions:

• The new regression method using foot length and ankle width provides a more accurate estimation of foot inertial parameters compared to existing scaling equations.
• This approach offers a practical and relevant alternative for clinical use due to its simplicity and improved accuracy.