Galbot and Tsinghua's LATENT wins IROS 2026 award for humanoid tennis forehand
Overview
A paper from Galbot, Tsinghua University and collaborators, describing LATENT, a method for humanoid tennis returns trained on imperfect amateur motion-capture clips, won the Best Entertainment and Amusement Paper Award at IROS 2026.
In simulation, the full forehand policy succeeded on 96.52 percent of returns, compared with 71.85 percent for the PULSE baseline. The paper reports 90.90 percent forehand success across 20 consecutive rallies on a real Unitree G1 robot.
The results are for a lab forehand return, not a full tennis match, and the real-robot tests still rely on motion capture rather than the robot's own cameras.
Written by AI from the articles below · updated Oct 8, 9:46 PM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- PandailyGalbot and Tsinghua's LATENT Wins IROS Award for Humanoid Tennis Forehand
AIA Galbot, Tsinghua University and collaborators paper won IROS 2026's Best Entertainment and Amusement Paper Award for LATENT, a humanoid tennis-return method trained on imperfect amateur motion-capture clips. In simulation, the full forehand policy succeeded on 96.52 percent of returns, versus 71.85 percent for PULSE. On a real Unitree G1, the paper reports 90.90 percent forehand success across 20 consecutive rallies, with motion capture still used rather than the robot's own cameras.
Heat trend
Not enough continuous observations to show a trend yet.