human-robot gym

human-robot gym是由慕尼黑工业大学信息学系创建的一个用于人机协作的强化学习基准数据集。该数据集包含8个具有挑战性的真实世界人机协作任务，如物体检测、传递和协作操作等。数据集通过集成SaRA shield确保机器人行为的安全性，支持多种机器人模型和高度可配置的场景。human-robot gym旨在解决人机协作中的安全性和任务多样性问题，为理论强化学习研究与实际应用之间搭建桥梁。

Human-Robot Gym is a reinforcement learning benchmark dataset for human-robot collaboration, developed by the Department of Informatics at the Technical University of Munich. It contains 8 challenging real-world human-robot collaboration tasks, including object detection, object handover, collaborative manipulation, and more. By integrating the SaRA shield, the dataset ensures safe robot behaviors, while supporting multiple robot models and highly configurable scenarios. The core goal of Human-Robot Gym is to address the issues of safety and task diversity in human-robot collaboration, and build a bridge between theoretical reinforcement learning research and real-world applications.