Noitom Robotics and research partners Shanghai AI Laboratory, Dobot Robotics and Shanghai Jiao Tong University unveiled AdaPT on August 21, 2026, a system that teaches humanoid robots to play tennis in the styles of professional players. The reveal came two days after Noitom released its 617.5-hour HiPHI motion dataset at the World Robot Conference in Beijing.
Rallying and serving on real hardware
AdaPT — Adaptive Motion Planning and Tracking — learns professional tennis rally and serving styles and executes them on physical humanoids, using adaptive planning and tracking to bridge the sim-to-real gap. The accompanying paper reports that AdaPT reproduces the distinctive playing styles of Rafael Nadal, Roger Federer and Novak Djokovic from publicly available broadcast footage, plus one more professional style captured with high-precision motion capture. It is validated on the Unitree G1 and the full-size Dobot Atom, and demonstrates in-the-wild serving using only a camera and a consumer tracker — no motion-capture studio required.
Foundation pre-trained on HiPHI-series motion
The motion foundation behind AdaPT was pre-trained on part of the HiPHI-series data — a bundle spanning the publicly released HiPHI dataset and the substantially larger, commercially licensable HiPHI-MOV corpus available through Noitom’s ModalityNet platform. “On Tuesday we made the foundation public. Today you can watch what gets built on ground like this,” said founder and CEO Dr. Tristan Ruoli Dai. “A robot can only learn these strokes when precise, physically grounded motion makes that video learnable. That is the World Compiler working.”
A step toward pro-athlete embodiment
Chief of R&D Dr. Lei Han called AdaPT step one of a broader program on professional athletic skills for humanoids, pre-trained for breadth on HiPHI-series data and post-trained on high-precision capture of professional athletes. Noitom Robotics said it will present AdaPT at RO-MAN 2026 in Fukuoka from August 24 to 28, alongside a public project page, arXiv paper, GitHub source code and demonstration video.
Reporting based on coverage from GlobeNewswire.