Optimizing Template Models to Quantifiably Assess Center of Mass Kinematic Reconstruction

Area of Science:

• Biomechanics and Robotics
• Human Locomotion Analysis

Background:

• The center of mass (COM) is crucial for human walking, but body redundancy complicates dynamic modeling.
• Existing template models (B-SLIP, VPP) simplify dynamics but often yield qualitative results with overestimated COM vertical displacement.

Purpose of the Study:

• To develop a quantifiable, template-based analysis for human walking dynamics.
• To improve the accuracy of center of mass (COM) trajectory prediction using optimized template models.

Main Methods:

• An optimization framework was employed to determine parameter values for B-SLIP and VPP models.
• Model parameters were tuned to match explicit and implicit gait characteristics.
• Leg stiffness was allowed to vary dynamically throughout the gait cycle.

Main Results:

• Optimized template models demonstrated a 54%-63% reduction in error for vertical COM trajectories.
• The models showed potential for retaining ground reaction force (GRF) information despite not being explicit optimization targets.

Conclusions:

• Optimized template models offer a more accurate, quantifiable approach to analyzing human walking dynamics.
• These refined models provide a promising reference for developing control strategies for lower-limb prosthetics.