RL-100: Performant Robotic Manipulation with Real-World Reinforcement Learning

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Lei, Kun, Li, Huanyu, Yu, Dongjie, Wei, Zhenyu, Guo, Lingxiao, Jiang, Zhennan, Wang, Ziyu, Liang, Shiyu, Xu, Huazhe
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866918380470009856
author Lei, Kun
Li, Huanyu
Yu, Dongjie
Wei, Zhenyu
Guo, Lingxiao
Jiang, Zhennan
Wang, Ziyu
Liang, Shiyu
Xu, Huazhe
author_facet Lei, Kun
Li, Huanyu
Yu, Dongjie
Wei, Zhenyu
Guo, Lingxiao
Jiang, Zhennan
Wang, Ziyu
Liang, Shiyu
Xu, Huazhe
contents Real-world robotic manipulation in homes and factories demands reliability, efficiency, and robustness that approach or surpass those of skilled human operators. We present RL-100, a real-world reinforcement learning framework built on diffusion visuomotor policies. RL-100 unifies imitation and reinforcement learning under a single clipped PPO surrogate objective applied within the denoising process, yielding conservative and stable improvements across offline and online stages. To meet deployment latency requirements, a lightweight consistency distillation method compresses multi-step diffusion into a one-step controller for high-frequency control. The framework is task-, embodiment-, and representation-agnostic, and supports both single-action and action-chunking control. We evaluate RL-100 on eight diverse real-robot tasks, from dynamic pushing and agile bowling to pouring, cloth folding, unscrewing, multi-stage juicing, and long-horizon box folding. RL-100 attains 100 percent success across evaluated trials, for a total of 1000 out of 1000 episodes, including up to 250 out of 250 consecutive trials on one task. It matches or surpasses expert teleoperators in time to completion. Without retraining, a single policy attains approximately 90 percent zero-shot success under environmental and dynamics shifts, adapts in a few-shot regime to significant task variations (86.7 percent), and remains robust to aggressive human perturbations (about 96 percent). Notably, our juicing robot served random customers continuously for about seven hours without failure when deployed zero-shot in a shopping mall. These results suggest a practical path to deployment-ready robot learning by starting from human priors, aligning training objectives with human-grounded metrics, and reliably extending performance beyond human demonstrations.
format Preprint
id arxiv_https___arxiv_org_abs_2510_14830
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RL-100: Performant Robotic Manipulation with Real-World Reinforcement Learning
Lei, Kun
Li, Huanyu
Yu, Dongjie
Wei, Zhenyu
Guo, Lingxiao
Jiang, Zhennan
Wang, Ziyu
Liang, Shiyu
Xu, Huazhe
Robotics
Artificial Intelligence
Machine Learning
Real-world robotic manipulation in homes and factories demands reliability, efficiency, and robustness that approach or surpass those of skilled human operators. We present RL-100, a real-world reinforcement learning framework built on diffusion visuomotor policies. RL-100 unifies imitation and reinforcement learning under a single clipped PPO surrogate objective applied within the denoising process, yielding conservative and stable improvements across offline and online stages. To meet deployment latency requirements, a lightweight consistency distillation method compresses multi-step diffusion into a one-step controller for high-frequency control. The framework is task-, embodiment-, and representation-agnostic, and supports both single-action and action-chunking control. We evaluate RL-100 on eight diverse real-robot tasks, from dynamic pushing and agile bowling to pouring, cloth folding, unscrewing, multi-stage juicing, and long-horizon box folding. RL-100 attains 100 percent success across evaluated trials, for a total of 1000 out of 1000 episodes, including up to 250 out of 250 consecutive trials on one task. It matches or surpasses expert teleoperators in time to completion. Without retraining, a single policy attains approximately 90 percent zero-shot success under environmental and dynamics shifts, adapts in a few-shot regime to significant task variations (86.7 percent), and remains robust to aggressive human perturbations (about 96 percent). Notably, our juicing robot served random customers continuously for about seven hours without failure when deployed zero-shot in a shopping mall. These results suggest a practical path to deployment-ready robot learning by starting from human priors, aligning training objectives with human-grounded metrics, and reliably extending performance beyond human demonstrations.
title RL-100: Performant Robotic Manipulation with Real-World Reinforcement Learning
topic Robotics
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2510.14830