Realistic, fast robotics simulations could enable large-scale training of embodied AI models that transfer to the real world. However, despite decades of progress, available simulations cover only a narrow range of real-world tasks and physical phenomena. The core bottleneck is that creating better simulations requires significant expert effort and therefore does not scale. To address this, we introduce Gym-Anything for Robotics, a framework in which coding agents create realistic robotics simulations. Our core insight is that simulation creation is a coding agent task, so improvements in coding agents translate directly into richer and more faithful simulations. We select simulations based on U.S. GDP data to cover economically valuable physical work, define what makes a good simulation and a good task, and develop an adversarial creation-audit loop that checks each simulation and task against these definitions. Using this pipeline, we build RoboWorld, a set of 25 simulations with 25 test and 175 training RL environments of varying difficulty. Even the best frontier models fail on every RoboWorld test task, achieving an average reward of only 0.06. Building on RoboWorld, we introduce several new ways for frontier models to improve, including in-context learning across repeated attempts, a privileged teacher that reads the simulator state and teaches a student in context, self-assisted learning from the model's own privileged experience, and code policies that generate demonstrations across varied scenes to train vision-language-action models (VLAs) such as $π_{0.5}$. Each of these substantially improves performance. We release our pipeline, RL environments, and benchmark to facilitate future research. Code and data are available at https://gym-anything.github.io/robotics
Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning
NVIDIA(英伟达)
Isaac Gym提供了一个高性能的学习平台,可以直接在GPU上为各种机器人任务进行训练策略。物理模拟和神经网络策略训练都驻留在GPU上,并通过直接将数据从物理缓冲区传递到PyTorch张量进行通信,而无需通过任何CPU瓶颈。这导致在单个GPU上进行复杂机器人任务的快速训练时间,与使用基于CPU的模拟器和用于神经网络的GPU的传统RL训练相比,提高了2-3个数量级。我们在\url{https://sites.google.com/view/isaacgym-nvidia}和isaac健身房托管结果和视频可以在\url{https://developer.nvidia.com/isaac-gym}下载。
safe-control-gym: a Unified Benchmark Suite for Safe Learning-based Control and Reinforcement Learning in Robotics
University of Toronto(多伦多大学);Vector Institute for Artificial Intelligence(向量人工智能研究所)
近年来,强化学习和基于学习的控制-以及对它们的安全性的研究,这对于在现实世界的机器人中部署至关重要-都获得了巨大的吸引力。然而,为了充分衡量新结果的进展和适用性,我们需要工具来公平地比较控制和强化学习社区提出的方法。在这里,我们提出了一个新的开源基准套件,称为安全控制健身房,支持基于模型和基于数据的控制技术。我们提供了三个动态系统的实现-车杆,1D和2D四旋翼-和两个控制任务-稳定和轨迹跟踪。我们建议扩展OpenAI的Gym API -强化学习研究中的事实标准-(i)指定(和查询)符号动力学的能力和(ii)约束,以及(iii)(可重复)在控制输入,状态测量和惯性属性中注入模拟干扰。为了证明我们的建议,并试图使研究社区更紧密地联系在一起,我们展示了如何使用安全控制健身房定量比较控制性能,数据效率和安全性的多种方法从传统的控制,基于学习的控制,和强化学习的领域。
Robust Gymnasium: A Unified Modular Benchmark for Robust Reinforcement Learning
University of California, Berkeley(加州大学伯克利分校);California Institute of Technology(加州理工学院);Shanghai Jiao Tong University(上海交通大学);Virginia Tech(弗吉尼亚理工大学);Carnegie Mellon University(卡内基梅隆大学)
SafeOR-Gym: A Benchmark Suite for Safe Reinforcement Learning Algorithms on Practical Operations Research Problems
Davidson School of Chemical Engineering, Purdue University(普渡大学化学工程学院);Energy Sciences, ExxonMobil Technology and Engineering Company(埃克森美孚技术与工程公司能源科学部)
MarineGym: A High-Performance Reinforcement Learning Platform for Underwater Robotics
The State Key Laboratory of Fluid Power and Mechatronic Systems, Zhejiang University(浙江大学流体动力与机电系统国家重点实验室);School of Engineering and Physical Sciences, Heriot-Watt University(赫瑞瓦特大学工程与物理科学学院);School of Informatics, University of Edinburgh(爱丁堡大学信息学院);Department of Mechanical and Aerospace Engineering, the Hong Kong University of Science and Technology(香港科技大学机械与航空航天工程系)
RoboNurse-VLA: Robotic Scrub Nurse System based on Vision-Language-Action Model
Multi-Scale Medical Robotics Centre, Ltd., The Chinese University of Hong Kong(香港中文大学多尺度医疗机器人中心有限公司);Istituto Italiano di Tecnologia(意大利技术研究院);Department of Surgery, The Chinese University of Hong Kong(香港中文大学外科系)
Safety-Driven Deep Reinforcement Learning Framework for Cobots: A Sim2Real Approach
Data Science Software Competence Center Hagenberg(哈根伯格数据科学软件能力中心);Industrial Automation Pilz(皮尔茨工业自动化公司);Computer Science Trinity College Dublin(都柏林圣三一学院计算机科学系);Food Science and Environmental Health Technological University Dublin(都柏林理工大学食品科学与环境健康学院);Robotics and Automation Irish Manufacturing Research(爱尔兰制造研究机构机器人与自动化部门)
提出一种将 ISO 10218 安全约束嵌入 DRL 训练并经 Sim2Real 迁移到协作机械臂的框架,在仿真和实测中提升成功率且无安全违规。
Grasp-Anything: Large-scale Grasp Dataset from Foundation Models
FPT Software AI Center(FPT软件人工智能中心);TU Wien(维也纳工业大学);AIT Austrian Institute of Technology(奥地利技术研究所);Imperial College London(伦敦帝国学院);Hanoi University of Science and Technology(河内科技大学);Ton Duc Thang University(孙德胜大学);University of Liverpool(利物浦大学)
URDF-Anything: Constructing Articulated Objects with 3D Multimodal Language Model
State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University(多媒体信息处理国家重点实验室,计算机学院,北京大学);University of Washington(华盛顿大学)
X-DiffVLA: X-Embodied Diffusion Action Heads for Vision-Language-Action Models
SKL-MAIS, Institute of Automation, Chinese Academy of Sciences(SKL-MAIS,自动化研究所,中国科学院);School of Artificial Intelligence, University of Chinese Academy of Sciences(人工智能学院,中国科学院大学);Beijing Academy of Artificial Intelligence(北京人工智能研究院);BeingBeyond;School of Computer Science, Peking University(北京大学计算机学院)