FreeWorld Dataset
收藏资源简介:
FreeWorld Dataset是由清华大学人工智能产业研究院推出的首个针对非结构化环境中的端到端机器人导航任务的综合数据集。该数据集通过实际机器人收集和Isaac Sim模拟器生成的合成数据两种方式构建,包含了静态对象类(道路分隔线)的注释和动态对象的3D边界框。数据集旨在促进物流和服务机器人端到端导航技术的发展,特别是在非结构化环境下的应用。
The FreeWorld Dataset, launched by the Institute for Artificial Intelligence Industry of Tsinghua University, is the first comprehensive dataset tailored for end-to-end robotic navigation tasks in unstructured environments. It is constructed using two main data sources: real-world data collected by physical robots and synthetic data generated via the Isaac Sim simulator. The dataset includes annotations for static object classes (such as road dividers) and 3D bounding boxes for dynamic objects. Its goal is to promote the development of end-to-end navigation technologies for logistics and service robots, especially for applications in unstructured environments.




