Researchers have developed PingpongDuo, a robotic framework that enables humanoid robots to play table tennis doubles through coordinated movements and alternating strikes. The system addresses the challenges of sharing limited space, repositioning quickly, maintaining striking accuracy, and avoiding collisions between teammates.
It combines coordinated positioning with learned movement and racket-control skills to help robots respond to changing ball trajectories while working together.
In experiments, the framework improved ball-return performance in simulation. It demonstrated that two humanoid robots could sustain consecutive rallies while following the strict alternating-strike rule required in doubles play.
Researchers from the University of Hong Kong, Shanghai Innovation Institute, Archon Robotics, and NVIDIA have developed PingpongDuo, a control framework that enables two humanoid robots to play table tennis doubles while coordinating their movements and alternating shots.
The system combines AI, motion learning, and risk estimation to help robots strike the ball, clear space for teammates, and reposition for subsequent returns without colliding.
Unlike conventional humanoid table tennis systems designed for individual play, PingpongDuo coordinates two robots operating in the confined space on one side of the table. Each robot must strike in strict alternation, moving away from the central hitting area after its shot while its teammate prepares to return the next ball.
The framework uses a hierarchical control system that combines team-level positioning with individual movement and striking skills. At the team level, a planner selects positioning targets for both robots. A learned risk estimator evaluates potential movements for collision risks involving their hands, feet, and bodies.
Source link
