Real-World Manipulation Tasks
收藏arXiv2025-09-30 收录
下载链接：
https://sites.google.com/view/dexterous-avail/
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了现实世界中的操作任务，其中机器人手通过视觉里程碑学习执行刮擦、连接接插件和钩取物体等操作。评估的任务包括刮盘、连接接插件和钩取物体，其性能指标基于二元成功测量。该数据集规模涉及多项需要复杂协调和操作的任务，主要任务是使用多指机器人手进行灵巧的对象操作。
This dataset contains real-world manipulation tasks, where robotic hands perform operations including scraping, connector mating, and object hooking via visual milestone learning. The evaluated tasks include scraping plates, connector mating, and object hooking, with performance metrics based on binary success measurements. This dataset covers a variety of tasks requiring complex coordination and manipulation, with the primary task focusing on dexterous object manipulation using multi-fingered robotic hands.
