Noitom Unveils ITW-Ego Plus at CoRL 2026; VREgo Framework Released for VR-Based Robot Learning
Overview
At CoRL 2026, Noitom Robotics introduced ITW-Ego Plus, a system capturing in-the-wild egocentric video and full-body motion using Sol headsets and Jupiter wearables. Concurrently, the VREgo framework was released on arXiv, demonstrating robot manipulation training from consumer VR headset demonstrations like the PICO 4 Ultra.
Generated from attributed reports · Updated 11 hours ago
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Environment: simulation
“positions after release, and propagate segmentation masks between keyframes. We evaluate policies in simulation on two tabletop tasks recorded by one operator. For cube placement, we compare PICO and Project Aria recordings of the same 41 layouts. With 40 training demonstrations, the PICO policies placed the cube on the tray in 67, 74, and 60 of 100 evaluation layouts across three training seeds. The corresponding Ar”
Exact source · revision 1Source owner not reported
Reported quantity · training demonstrations: 40 other · Basis not reported
“positions after release, and propagate segmentation masks between keyframes. We evaluate policies in simulation on two tabletop tasks recorded by one operator. For cube placement, we compare PICO and Project Aria recordings of the same 41 layouts. With 40 training demonstrations, the PICO policies placed the cube on the tray in 67, 74, and 60 of 100 evaluation layouts across three training seeds. The corresponding Ar”
Exact source · revision 1Source owner not reported
Reported quantity · evaluation layouts: 100 other · Basis not reported
“positions after release, and propagate segmentation masks between keyframes. We evaluate policies in simulation on two tabletop tasks recorded by one operator. For cube placement, we compare PICO and Project Aria recordings of the same 41 layouts. With 40 training demonstrations, the PICO policies placed the cube on the tray in 67, 74, and 60 of 100 evaluation layouts across three training seeds. The corresponding Ar”
Exact source · revision 1Source owner not reported
Data origin: human egocentric
“agree to the Terms of Service and Privacy Policy . Egocentric video is now widely available. What remains scarce is in-the-wild data that is learnable: first-person video paired with body motion, viewpoints and contact signals. NR's ITW line has two products: ITW-EgoVision , egocentric vision captured on NR's Sol headset, and ITW-Ego Plus , which adds modalities à la carte: Body (full-body motion), Wrist (wrist-pers”
Exact source · revision 2Source owner not reported
Reported quantity · egocentric video hours: 11 hours · Basis not reported · Differing source assertions
“m Robotics. Three main-track papers - HiPHI, OmniContact and AdaPT - presented Nov. 10 by NEAR Lab, the Noitom Robotics Embodied-AI Research Lab introduced at CoRL under Dr. Han Lei, with the lab’s Unified Motion Retargeting (UMR) research validated on five humanoids. HiPHI-Hand released openly on Hugging Face at CoRL: 11 hours of egocentric video with marker-based 3D hand ground truth, produced for NVIDIA’s hand-rec”
Exact source · revision 2Source owner not reported
Reported quantity · egocentric video hours: 11 hours · Basis not reported · Differing source assertions
“same family of openly released data. NR produced the capture for NVIDIA's hand-reconstruction research: 11 hours of synchronized egocentric video from the Sol headset with marker-based 3D hand ground truth at sub-millimeter precision, captured with an optical motion-capture system using micro-markers on the fingers. NVIDIA processed this capture into its own training representation and used it as one of the real-wor”
Exact source · revision 2Source owner not reported
Reported quantity · whole-body motion hours: 617.5 hours · Basis not reported
“in: HiPHI: A Large-Scale Benchmark for High-Precision Human Motion and Object-Interaction (Jiahao Ji et al.) - Uses FrameNet to plan what to capture: 617.5 hours of whole-body motion and real-object interaction, evaluated from motion coverage through to deployment on real humanoids. OmniContact: Chaining Meta-Skills via Contact Flow (Runyi Yu et al.) - Contact flow links task-level reasoning to whole-body execution,”
Exact source · revision 2Source owner not reported
Reported quantity · egocentric video downloads: 15000 other · reported trials
“onstruction research and processed by NVIDIA as a real-world source for training NVIDIA EgoHand, demonstrated live at Booth 1105. HiPHI itself has passed 15,000 downloads. AUSTIN, Texas, Nov. 05, 2026 (GLOBE NEWSWIRE) -- Noitom Robotics (NR), a Physical AI data infrastructure company, will present three accepted papers at the Conference on Robot Learning ( CoRL 2026 ) in Austin, Nov. 9-12, and introduce ITW-Ego Plus”
Exact source · revision 2Source owner not reported
Dataset · license: ModalityNet Open Research License (MORL) v1.0
“data HiPHI , the open research subset of the HiPHI-MOV corpus released in August under the ModalityNet Open Research License (MORL) v1.0, free for research, has passed 15,000 downloads across Hugging Face and NR’s distribution partners. The full HiPHI-MOV corpus is licensed commercially through ModalityNet. HiPHI-Hand, which NR releases on Hugging Face at CoRL under the same license, adds hand-level precision to the”
Exact source · revision 2Source owner not reported
Dataset · modalities: egocentric video · Differing source assertions
“m Robotics. Three main-track papers - HiPHI, OmniContact and AdaPT - presented Nov. 10 by NEAR Lab, the Noitom Robotics Embodied-AI Research Lab introduced at CoRL under Dr. Han Lei, with the lab’s Unified Motion Retargeting (UMR) research validated on five humanoids. HiPHI-Hand released openly on Hugging Face at CoRL: 11 hours of egocentric video with marker-based 3D hand ground truth, produced for NVIDIA’s hand-rec”
Exact source · revision 2Source owner not reported
Dataset · modalities: full-body motion · Differing source assertions
“agree to the Terms of Service and Privacy Policy . Egocentric video is now widely available. What remains scarce is in-the-wild data that is learnable: first-person video paired with body motion, viewpoints and contact signals. NR's ITW line has two products: ITW-EgoVision , egocentric vision captured on NR's Sol headset, and ITW-Ego Plus , which adds modalities à la carte: Body (full-body motion), Wrist (wrist-pers”
Exact source · revision 2Source owner not reported
Dataset · modalities: third-person views · Differing source assertions
“pective video), Exo (fixed third-person views of task and user), Touch (hand pressure) and Sound (environmental audio). At CoRL, NR shows Body . Powered by Sol and Jupiter, both developed in-house by NR, the ITW-Ego Plus capture system connects first-person visual context with whole-body movement. Jupiter - a few small, discreet wearables - captures full-body motion without a motion-capture suit, so people perform ta”
Exact source · revision 2Source owner not reported
Dataset · modalities: hand pressure · Differing source assertions
“pective video), Exo (fixed third-person views of task and user), Touch (hand pressure) and Sound (environmental audio). At CoRL, NR shows Body . Powered by Sol and Jupiter, both developed in-house by NR, the ITW-Ego Plus capture system connects first-person visual context with whole-body movement. Jupiter - a few small, discreet wearables - captures full-body motion without a motion-capture suit, so people perform ta”
Exact source · revision 2Source owner not reported
Dataset · modalities: environmental audio · Differing source assertions
“pective video), Exo (fixed third-person views of task and user), Touch (hand pressure) and Sound (environmental audio). At CoRL, NR shows Body . Powered by Sol and Jupiter, both developed in-house by NR, the ITW-Ego Plus capture system connects first-person visual context with whole-body movement. Jupiter - a few small, discreet wearables - captures full-body motion without a motion-capture suit, so people perform ta”
Exact source · revision 2Source owner not reported
Dataset · tasks: tennis styles · Differing source assertions
“so humanoids chain skills into long-horizon tasks. Trained on the HiPHI-MOV (Motion with Object and Vision) corpus. AdaPT: Towards Professional Tennis Styles for Humanoid Robots with Adaptive Motion Planning and Tracking (Tao Huang, Ruofei Liu et al.) - Humanoid tennis in professional styles, learned from broadcast video. Pre-trained on part of the HiPHI-series data, including a relevant subset of the publicly releas”
Exact source · revision 2Source owner not reported
Dataset · tasks: object interaction · Differing source assertions
“in: HiPHI: A Large-Scale Benchmark for High-Precision Human Motion and Object-Interaction (Jiahao Ji et al.) - Uses FrameNet to plan what to capture: 617.5 hours of whole-body motion and real-object interaction, evaluated from motion coverage through to deployment on real humanoids. OmniContact: Chaining Meta-Skills via Contact Flow (Runyi Yu et al.) - Contact flow links task-level reasoning to whole-body execution,”
Exact source · revision 2Source owner not reported
Dataset · embodiments: humanoids
“m Robotics. Three main-track papers - HiPHI, OmniContact and AdaPT - presented Nov. 10 by NEAR Lab, the Noitom Robotics Embodied-AI Research Lab introduced at CoRL under Dr. Han Lei, with the lab’s Unified Motion Retargeting (UMR) research validated on five humanoids. HiPHI-Hand released openly on Hugging Face at CoRL: 11 hours of egocentric video with marker-based 3D hand ground truth, produced for NVIDIA’s hand-rec”
Exact source · revision 2Source owner not reported
Dataset · version: v1.0
“data HiPHI , the open research subset of the HiPHI-MOV corpus released in August under the ModalityNet Open Research License (MORL) v1.0, free for research, has passed 15,000 downloads across Hugging Face and NR’s distribution partners. The full HiPHI-MOV corpus is licensed commercially through ModalityNet. HiPHI-Hand, which NR releases on Hugging Face at CoRL under the same license, adds hand-level precision to the”
Exact source · revision 2Source owner not reported
Evidence dependency changes (1)
- 2026-11-09T15:14:32.485Z · evidence updated · source revision 2. Evidence extraction was updated; current source attributions are shown above.
Developments
- 2026-11-06 05:00 UTC · 1 reportsVREgo: Consumer VR Headsets for Learning Robot Manipulation from EgocentricDemonarXiv Robotics — research abstracts：VREgo: Learning Robot Manipulation from Egocentric Demonstrations Using Consumer VR Headsets
- 2026-11-05 05:50 UTC · 1 reportsNoitom Robotics introduces ITW-Ego Plus and HiPHI-Hand at CoRL 2026Robotics — Web news discovery：Noitom Robotics Introduces ITW-Ego Plus at CoRL 2026: In-the-Wild Egocentric Data With Full-Body Motion
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsSignalVREgo: Learning Robot Manipulation from Egocentric Demonstrations Using Consumer VR Headsets
This research presents VREgo, a framework for training robot manipulation policies using demonstrations recorded on consumer VR headsets like the PICO 4 Ultra. The system utilizes on-device tracking and stereo localization to capture head and hand movements, adapting HumanEgo preprocessing techniques to estimate grasp thresholds and update object positions. Evaluated in simulation on two tabletop tasks, VREgo trained with 40 demonstrations from a single operator achieved success rates of 67%, 74%, and 60% across three seeds for cube placement, compared to Project Aria's 73%, 78%, and 0%. For a stacking task, the baseline achieved 94% success. The authors address issues like premature gripper opening by adapting trajectories and adding execution rules, ultimately reaching 93.8% success across noise seeds. The code is available at https://vrego.chuange.org/.
- Robotics — Web news discoverySignalNoitom Robotics Introduces ITW-Ego Plus at CoRL 2026: In-the-Wild Egocentric Data With Full-Body Motion
At the Conference on Robot Learning (CoRL) 2026 in Austin, Noitom Robotics introduced ITW-Ego Plus, a new category of in-the-wild human data that combines egocentric video with full-body motion and other modalities. The system utilizes Sol headsets for first-person visual context and Jupiter wearables for capturing whole-body movement without bulky suits.
Event coverage history
Current attention 3·Peak within the comparable range 15（2026-11-06 06:00 UTC）·Change within the comparable range over 24 hours -67%
The trend compares participants observed continuously; its scope may be narrower than current attention totals. Hover or click to explore hourly observations, or use the left and right arrow keys.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).