Robotics PapersA programme for reading clubs

Topic 05 of 17

Egocentric & human data in robotics

Egocentric human data is first-person video and sensor capture of people doing everyday tasks, used as training signal for robots. It matters because teleoperated robot demonstrations are the main cost bottleneck in robot learning, and humans wearing cameras or smart glasses generate task demonstrations orders of magnitude faster. The open problem is the embodiment gap: human hands and robot grippers act differently.

54papers in 2026-07
4.62%of all cs.RO that month
+0.86ppshare change, last 6 vs prior 12 months
24-month volume

arXiv query: cat:cs.RO AND (abs:"egocentric" OR abs:"human video" OR abs:"human demonstrations")

Presented at reading clubs

These already have a slide kit, so they are the fastest ones to pick up and present.

2026-02-06
Shenyuan Gao, William Liang, Kaiyuan Zheng +27 · 96 citations
2025-12-27
Simar Kareer, Karl Pertsch, James Darpinian +5 · 39 citations
2025-09-23
Ryan Punamiya, Dhruv Patel, Patcharapong Aphiwetsa +5 · 30 citations
2025-09-04
Lawrence Y. Zhu, Pranav Kuppili, Ryan Punamiya +5 · 35 citations
2025-05-26
Vincent Liu, Ademi Adeniji, Haotian Zhan +4 · 41 citations
2025-05-16
Ryan Hoque, Peide Huang, David J. Yoon +2 · 193 citations

Recent work

2026-09-04
Masaki Kuribayashi, Zhongkai Shangguan, Eshed Ohn-Bar
2026-08-29
George Xi Wang, Henghao Li, Shan Lin +3
2026-08-27
Canzhi Chen, Zan Wang, Siqi Zhu +3

329 papers in this corpus carry the egocentric & human data tag. How tagging works.