AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

RoboWheel: A Data Engine from Real-World Human Demonstrations for Cross-Embodiment Robotic Learning

Yuhong Zhang, Zihan Gao, Shengpeng Li, Ling-Hao Chen, Kaisheng Liu, Runqing Cheng, Xiao Lin, Junjia Liu, Zhuoheng Li, Jingyi Feng, Ziyan He, Jintian Lin and 3 more

CVPR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Memory and subgoal-based manipulation · 9 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add TSDF penetration optimization and RL refinement together (ablation)RoboWheel · HO-Cap HOI reconstructionreplay success rate: 78.7 -> 93.6Improved
add RL refinement only (no TSDF) (ablation)RoboWheel · HO-Cap HOI reconstructionreplay success rate: 78.7 -> 61.7Hurt
add TSDF only (ablation)RoboWheel · HO-Cap HOI reconstructionhand jitter: 3.34 -> 0.84Improved
add RL on top of TSDF (ablation)RoboWheel · HO-Cap HOI reconstructionobject F10: 83.2 -> 89.1Improved
train on 10 HORA trajectories instead of 10 teleop demosRDT · Real-world easy taskssuccess rate (%): 66.3 -> 47.5Hurt
train on 10 HORA trajectories instead of 10 teleop demosPi0 · Real-world hard taskssuccess rate (%): 40.0 -> 31.3Hurt
train on 10 HORA trajectories instead of teleopDiffusion Policy · Real-world hard taskssuccess rate (%): 1.3 -> 6.3Mixed
pretrain on 5k HORA trajectories before fine-tuningPi0 · Real-world hard taskssuccess rate (%): 31.3 -> 58.8Improved
pretrain on 5k HORA trajectories before fine-tuningRDT · Real-world hard taskssuccess rate (%): 25.0 -> 47.5Improved
add simulation data augmentation (HORA-aug) (ablation)RDT · Unseen backgroundsuccess (out of 10): 1.50 -> 4.00Improved
add simulation data augmentation (HORA-aug) (ablation)RDT · Unseen objectsuccess (out of 10): 3.75 -> 4.25No clear effect
use RoboWheel gripper retargeting vs prior methodsRoboWheel · Robot direct replaysuccess rate (%): 50.0 -> 91.7Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy