Incorporating Human-like Walking Variability in an HZD-Based Bipedal Model.

Area of Science:

• Robotics
• Biomechanics
• Human Motion Simulation

Background:

• Predictive simulations of human walking are valuable for research.
• Current models using hybrid zero dynamics (HZD) lack variability, limiting fall risk analysis.
• Human gait variability is essential for realistic simulations and understanding fall risk.

Purpose of the Study:

• To incorporate human-like variability into HZD-based walking models.
• To develop a more realistic human walking simulation capable of analyzing fall risk.
• To ensure the bipedal model maintains stable walking while incorporating variability.

Main Methods:

• Augmenting the commanded motion in an HZD model with a sinusoidal variability function.
• Implementing a polynomial correction function to control the growth of variability.
• Conducting simulations to validate the necessity of the correction function and assess variability reduction.

Main Results:

• Successfully incorporated human-like variability into an HZD walking model.
• Demonstrated that a correction function is necessary to maintain stable walking.
• Showed that reducing stance ankle variability improves gait realism.
• Observed that simulated temporal variability aligns with experimental data.

Conclusions:

• The developed HZD model with incorporated variability provides a more realistic simulation of human walking.
• This approach enables the investigation of fall risk by accounting for gait variability.
• The findings highlight the importance of variability control in dynamic bipedal simulations.