Dimensions overall score 8.0

GitHub Code Pulse

No public code linked for this paper yet.

• The LATENT system learns athletic humanoid tennis skills from imperfect human motion data consisting only of motion fragments capturing primitive tennis skills.method 95%
• Using imperfect human motion data significantly reduces the difficulty of data collection compared to requiring precise and complete human-tennis motion sequences from real matches.technical 90%
• The learned humanoid policy can consistently strike incoming tennis balls under a wide range of conditions and return them to target locations.result 90%
• The method preserves natural human motion styles in the learned robot behaviors.result 85%
• The system achieves successful sim-to-real transfer and deployment on the Unitree G1 humanoid robot.result 90%
• The method achieves surprising results in the real world and can stably sustain multi-shot rallies with human players.result 90%
• The key insight is that imperfect, quasi-realistic human motion data still provide priors about human primitive skills in tennis scenarios.technical 85%
• The imperfect human motion data consist only of motion fragments capturing primitive tennis skills rather than complete motion sequences.method 90%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card