DexYCB
收藏资源简介:
DexYCB数据集由华盛顿大学创建,专注于捕捉人类抓取物体的行为,用于模拟人机交互中的物体传递过程。该数据集包含1000个由10名参与者抓取20种不同物体的动作捕捉序列,每个序列记录了从抓取到传递的完整动作。创建过程中,利用了高精度的动作捕捉技术,确保数据的准确性和实用性。DexYCB数据集主要应用于机器人学习和人机交互领域,旨在通过模拟真实的人类行为,提高机器人在复杂环境中的交互能力和操作效率。
The DexYCB dataset, developed by the University of Washington, focuses on capturing human object-grasping behaviors for simulating object handover processes in human-robot interaction (HRI). It comprises 1000 motion capture sequences, where 10 participants perform grasping actions on 20 distinct objects, with each sequence recording the complete action spanning from grasping to handover. High-precision motion capture technology was employed during its creation to ensure the accuracy and practicality of the dataset. The DexYCB dataset is primarily applied in the fields of robotic learning and human-robot interaction, with the goal of enhancing robots' interactive capabilities and operational efficiency in complex environments by simulating realistic human behaviors.

- 1HandoverSim: A Simulation Framework and Benchmark for Human-to-Robot Object Handovers华盛顿大学 · 2022年



