Athlete framework teaches humanoid robots to play tennis reactively
ATHLETE: Learning Reactive Humanoid Tennis via Trajectory-Guided Motion Matching
Robotics
Summary
Playing tennis with a humanoid robot is hard because it needs to move its whole body quickly and accurately to hit a fast-moving ball. The authors created ATHLETE, a system that teaches a robot to play tennis by combining general walking and running movements with tennis hitting moves learned from ordinary videos. This system builds a virtual coach that guides the robot’s movements and then turns that into a real-time robot player that responds directly to the ball. When tested on an actual humanoid robot, ATHLETE could hit tennis balls with high success and even rally with humans.
What this means in practice
- •For robot developers: Enable robots to learn dynamic sports skills by combining generic locomotion data with video demonstrations for responsive control.
- •For robotics software engineers: Build reactive control policies for complex whole-body tasks without needing extensive task-specific motion capture data.
Authors
Zhihao Tian, Zifan Wang, Enzheng Zhu, Xun Yang, Jialong Qin, Mengbo Wang, Junwei Liang
Abstract
Humanoid tennis requires reactive whole-body coordination under rapidly changing ball dynamics, yet collecting task-specific motion-capture demonstrations that cover diverse approach states, contact locations, and timing conditions is costly. Inspired by how humans acquire new skills by combining reusable motor experience with task-specific demonstrations, we present ATHLETE, a framework for learning reactive humanoid tennis from generic locomotion data and ordinary tennis videos. ATHLETE combines LAFAN1 locomotion with video-derived striking motions and uses Trajectory-Guided Motion Matching (TGMM) to construct diverse task-conditioned whole-body references. These references are transformed into a reactive policy through physics-based teacher--student learning with online DAgger distillation, followed by PPO fine-tuning for the tennis task. The resulting policy responds directly to robot and ball observations without requiring reference-motion inputs at deployment. On a real humanoid robot, ATHLETE achieves forehand and backhand hit rates of $90.00\%$ and $86.67\%$, respectively, and enables multi-shot rallies with human players. These results demonstrate that ATHLETE can learn dynamic humanoid tennis skills from reusable locomotion data and video demonstrations. Result videos and code can be found at https://anonymous.4open.science/w/ATHLETE2027/.