Full-Body Animation of Human Locomotion in Reduced Gravity Using Physics-Based Control

Area of Science:

• Computer Graphics
• Physics Simulation
• Robotics

Background:

• Generating realistic character animation across varying gravitational conditions presents significant challenges.
• Existing methods often struggle with stability and robustness when simulating non-Earth gravity.

Purpose of the Study:

• To develop a physics-based approach for generating stable and robust full-body character animation under diverse gravitational conditions.
• To adapt motion-captured human data from Earth's gravity to simulate movement in hypogravity environments.

Main Methods:

• Utilized motion-captured human data under Earth's gravity as input.
• Developed an inverted-pendulum on cart (IPC) control model analyzed with motion data.
• Employed a pre-estimation model based on the Froude number to predict velocity and stride frequency for hypogravity.
• Generated full-body animation using a pendulum trajectory generator, motion planner, and tracking.

Main Results:

• Successfully generated stable and robust full-body animations for various gaits.
• Demonstrated the capability to adapt animations to different gravitational conditions, including hypogravity.
• The physics-based approach proved effective in creating realistic character movements.

Conclusions:

• The proposed physics-based method offers a robust solution for generating realistic character animations in simulated low-gravity environments.
• This approach enhances the fidelity of character motion in virtual simulations, particularly for space exploration or scientific visualization.
• The IPC control model and Froude number-based prediction are key components for successful gravity adaptation.