Segment inertial parameter evaluation in two anthropometric models by application of a dynamic linked segment model

Area of Science:

• Biomechanics
• Human Movement Analysis
• Anthropometry

Background:

• Accurate estimation of segment inertial parameters (SIPs) is crucial for reducing errors in inverse dynamic analysis.
• Current methods for SIP estimation include regression-based proportional models and shape-approximating geometric models.

Purpose of the Study:

• To compare the accuracy of proportional and geometric anthropometric models in inverse dynamic analysis.
• To evaluate the reliability of these models across different types of human movements.

Main Methods:

• Five males and five females performed four distinct sagittal plane lifting movements.
• A full-body linked segment model was applied to the same dataset using both proportional and geometric anthropometric models.
• Systematic errors were assessed using an overdetermined system of equations derived from the linked segment model.

Main Results:

• The geometric model demonstrated superior overall performance (higher coefficients of correlation) compared to the proportional model.
• Systematic errors were observed in the proportional model during back lifting movements.
• Systematic errors were identified in the geometric model during leg lifting movements.

Conclusions:

• Neither the proportional nor the geometric anthropometric model is consistently reliable across all movement types.
• Evaluating model reliability requires analyzing a variety of movements, as conclusions drawn from a single movement type may be misleading.