Real-Time Stylized Humanoid Behavior Control through Interaction and Synchronization

Area of Science:

• Robotics
• Computer Animation
• Human-Computer Interaction

Background:

• Simulating human-level perception and motion control in characters is challenging due to behavioral diversity.
• Existing methods often require extensive data or complex preprocessing.

Purpose of the Study:

• To present a novel style-based teleoperation framework for character control.
• To achieve human-like motion synchronization and style mimicking.
• To ensure character balance and robustness in real-time.

Main Methods:

• A style-based teleoperation framework integrating human perception and analysis.
• Motion optimization synthesizing high-level style features with balance strategies.
• Center-of-mass and root virtual control (CR-VC) for motion rhythm synchronization.

Main Results:

• The framework successfully mimics human motion styles and synchronizes character movement.
• Character balance is maintained throughout teleoperation.
• The system demonstrates robustness to real-time human behavior changes without inverse dynamics knowledge.

Conclusions:

• The proposed framework offers a generalized and robust solution for character motion control.
• It facilitates human-robot interaction by enabling characters to understand and perform tasks.
• This research advances the field of character animation and virtual embodiment.