Center of mass kinematic reconstruction during steady-state walking using optimized template models

Area of Science:

• Biomechanics
• Robotics
• Human Locomotion

Background:

• Template models like Bipedal Spring-Loaded Inverted Pendulum (B-SLIP) and Virtual Pivot Point (VPP) simplify legged locomotion dynamics.
• These models qualitatively match human walking but overestimate center of mass (CoM) displacement and underestimate gait event timing.

Purpose of the Study:

• To investigate the hypothesis that constant leg stiffness in template models causes discrepancies with human walking data.
• To systematically assess the impact of variable leg stiffness on model fidelity and human walking characteristics.

Main Methods:

• An optimization framework was developed to identify optimal leg stiffness trajectories.
• The framework was applied to 24 subjects across a range of walking speeds (40%-145% preferred walking speed).
• Effects of stiffness variation on ground reaction forces (GRF) and gait event timing were quantified.

Main Results:

• Variable leg stiffness models reduced CoM error by over 80% in both B-SLIP and VPP models.
• Improved prediction of human GRF profiles was observed with variable stiffness.
• Accuracy of gait event timing did not consistently improve across all conditions.

Conclusions:

• Variable leg stiffness is crucial for accurately simulating human walking dynamics, particularly CoM displacement and GRF profiles.
• The optimized stiffness profiles reflect biomechanical strategies like ankle push-off and reduced CoM vaulting.
• Further research may be needed to improve gait event timing predictions in these models.