PACE: Physics Augmentation for Coordinated End-to-end Reinforcement Learning toward Versatile Humanoid Table Tennis
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911532557795328 |
|---|---|
| author | Hu, Muqun Chen, Wenxi Li, Wenjing Mandali, Falak He, Zijian Zhang, Renhong Krisna, Praveen Christian, Katherine Benaharon, Leo Ma, Dizhi Ramani, Karthik Gu, Yan |
| author_facet | Hu, Muqun Chen, Wenxi Li, Wenjing Mandali, Falak He, Zijian Zhang, Renhong Krisna, Praveen Christian, Katherine Benaharon, Leo Ma, Dizhi Ramani, Karthik Gu, Yan |
| contents | Humanoid table tennis (TT) demands rapid perception, proactive whole-body motion, and agile footwork under strict timing--capabilities that remain difficult for end-to-end control policies. We propose a reinforcement learning (RL) framework that maps ball-position observations directly to whole-body joint commands for both arm striking and leg locomotion, strengthened by predictive signals and dense, physics-guided rewards. A lightweight learned predictor, fed with recent ball positions, estimates future ball states and augments the policy's observations for proactive decision-making. During training, a physics-based predictor supplies precise future states to construct dense, informative rewards that lead to effective exploration. The resulting policy attains strong performance across varied serve ranges (hit rate$\geq$96% and success rate$\geq$92%) in simulations. Ablation studies confirm that both the learned predictor and the predictive reward design are critical for end-to-end learning. Deployed zero-shot on a physical Booster T1 humanoid with 23 revolute joints, the policy produces coordinated lateral and forward-backward footwork with accurate, fast returns, suggesting a practical path toward versatile, competitive humanoid TT. We have open-sourced our RL training code at: https://github.com/purdue-tracelab/TTRL-ICRA2026 |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_21690 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | PACE: Physics Augmentation for Coordinated End-to-end Reinforcement Learning toward Versatile Humanoid Table Tennis Hu, Muqun Chen, Wenxi Li, Wenjing Mandali, Falak He, Zijian Zhang, Renhong Krisna, Praveen Christian, Katherine Benaharon, Leo Ma, Dizhi Ramani, Karthik Gu, Yan Robotics Humanoid table tennis (TT) demands rapid perception, proactive whole-body motion, and agile footwork under strict timing--capabilities that remain difficult for end-to-end control policies. We propose a reinforcement learning (RL) framework that maps ball-position observations directly to whole-body joint commands for both arm striking and leg locomotion, strengthened by predictive signals and dense, physics-guided rewards. A lightweight learned predictor, fed with recent ball positions, estimates future ball states and augments the policy's observations for proactive decision-making. During training, a physics-based predictor supplies precise future states to construct dense, informative rewards that lead to effective exploration. The resulting policy attains strong performance across varied serve ranges (hit rate$\geq$96% and success rate$\geq$92%) in simulations. Ablation studies confirm that both the learned predictor and the predictive reward design are critical for end-to-end learning. Deployed zero-shot on a physical Booster T1 humanoid with 23 revolute joints, the policy produces coordinated lateral and forward-backward footwork with accurate, fast returns, suggesting a practical path toward versatile, competitive humanoid TT. We have open-sourced our RL training code at: https://github.com/purdue-tracelab/TTRL-ICRA2026 |
| title | PACE: Physics Augmentation for Coordinated End-to-end Reinforcement Learning toward Versatile Humanoid Table Tennis |
| topic | Robotics |
| url | https://arxiv.org/abs/2509.21690 |