Unitree G1 Robots Achieve Ping-Pong Doubles Coordination via Joint Planning

1 reports archived1 independent sourcesupdated 10/17/2026, 12:00:00

Synthesis & Latest Updates

1 Sources Cross-Validated

Researchers from Shanghai Institute of Intelligent Science and Technology and others introduced PingpongDuo, enabling two Unitree G1 humanoid robots to play ping-pong doubles. The framework uses joint position planning and whole-body control to handle alternating hits, movement-induced swing errors, and collision risks.

LATEST/Researchers from Shanghai Institute of Intelligent Science and Technology and others introduced PingpongDuo, enabling two Unitree G1 humanoid robots to play ping-pong doubles. The framework uses joint position planning and whole-body control to handle alternating hits, movement-induced swing errors, and collision risks.

HEAT TRENDHourly heat curve

2 fully observed hours

Fewer than 3 fully observed hours; trend not plotted yet.

TIMELINECoverage timeline

Total 1 reports · Latest first

量子位 (微信公众号)T2·Oct 17, 12:0068 pts
Unitree G1 Robots Achieve Ping-Pong Doubles Coordination via Joint Planning
- Implemented a strict 'hit-yield' state machine to ensure alternating ball contact per ping-pong doubles rules.
- Introduced 'role weighting' to prioritize the stability of the robot about to hit the ball during collision avoidance adjustments.
- Adopted a hierarchical architecture: upper-level team positioning via PPO reinforcement learning, lower-level execution by fixed motion controllers.
View Related Reports & Analysis Original Source