发表机构
Imperial College London; Robotics and AI Institute（帝国理工学院; 机器人与人工智能研究所）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
ZeroBot提出了一种real2sim框架，利用图像到3D生成模型和并行强化学习，在无人类示范、无预训练、无已知模型的情况下，仅用单视图和目标位姿，在数分钟内学会操作任务，实现87%成功率，平均训练119秒。
AI 中文摘要
我们提出了ZeroBot，一个real2sim框架，用于在具有挑战性的条件下从零开始学习机器人操作任务，这些条件包括：零人类示范、零策略预训练和零已知物体模型。仅给定物体的单一视图和该物体的目标位姿，ZeroBot使用图像到3D的生成模型来获得完整的物体网格，该网格用于模拟中的大规模并行强化学习。为了加速训练，我们引入了一个动作空间，该动作空间利用生成的几何形状和学习到的价值函数来采样涉及机器人-物体接触的状态。在包括抓取、推挤、铰接物体交互和多阶段操作在内的真实世界任务上评估时，ZeroBot实现了87%的成功率，平均训练时间为119秒。这些结果展示了在real2sim框架中使用图像到3D模型进行快速、自主机器人学习的价值。
英文摘要
We present ZeroBot, a real2sim framework for learning a robot manipulation task from scratch in minutes under challenging conditions: zero human demonstrations, zero policy pre-training, and zero known object models. Given only a single view of an object and a goal pose for that object, ZeroBot uses image-to-3D generative models to obtain a complete object mesh, which is used in simulation for large-scale parallel reinforcement learning. To accelerate training, we introduce an action space which leverages the generated geometry and learned value function to sample states involving robot-object contact. When evaluated on real-world tasks including grasping, pushing, articulated object interaction, and multi-stage manipulation, ZeroBot achieves an 87% success rate with an average training time of 119 seconds. These results show the value of using image-to-3D models in a real2sim framework for rapid, autonomous robot learning.
CommentsIEEE RA-Letters 2026. Project page: https://zerobot-rl.github.io
