Vision in Action: Learning Active Perception from Human Demonstrations
Active perception learned from human demonstrations: a 6-DoF robotic neck and a VR interface let bimanual robots search, track and focus.从人类示教学习主动感知：6 自由度机械颈加 VR 界面，让双臂机器人能搜索、跟踪和聚焦目标。
BibTeX
@misc{xiong2025vision,
title = {{Vision in Action: Learning Active Perception from Human Demonstrations}},
author = {Haoyu Xiong and Xiaomeng Xu and Jimmy Wu and Yifan Hou and Jeannette Bohg and Shuran Song},
year = {2025},
eprint = {2506.15666},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2506.15666},
}Manipulation & Dexterous Hands操作与灵巧手Hardware & Open-Source Robots硬件与开源机器人Robot arm机械臂active-perceptionbimanualteleoperation