Video World Models · 2026-10-20
Ropedia W0: Scaling Loco-Manipulation World Models across Embodiments with Human Experience
Tianqi Liu, Hanfei Zhao, Qifeng Liang, Hui En Pang, Jiarui Xu, Sayang Mu, Penghao Qiu, YingSheng Zhang, Jun Wang, Ziang Cao, Mingze Sun, Yukang Cao, Fangzhou Hong, Zhaoxi Chen, Ziwei Liu
AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.
TL;DR
Ropedia W0 is a unified egocentric loco-manipulation world foundation model spanning humans and robots. It trains on Xp-Data (∼10k hours human + ∼3k hours robot) with unified geometric/kinematic/semantic annotations, using an explicit kinematic interface: metric camera poses via Plücker-ray maps and projected skeletons for manipulation. The authors evaluate on five cross-embodiment benchmarks plus held-out human egocentric video, and report real-time minute-long rollouts after distillation for live VR interaction.
Why it matters
The paper addresses cross-embodiment embodied world modeling by unifying locomotion (viewpoint changes) and manipulation (hand/robot actions) under a shared conditioning interface, and by scaling training data from both humans and robots. The authors also introduce Xp-Teleop, a long-horizon benchmark for evaluating locomotion + dexterous manipulation prediction.
Method
- Curate Xp-Data with ∼10k hours egocentric human data and ∼3k hours robot data, using unified multimodal annotations and systematic quality control; include extensive manual robot keypoint annotation and correction.
- Use a shared explicit kinematic control interface: encode locomotion with camera poses as dense Plücker-ray maps, and manipulation with image-aligned projected skeletons from human pose annotations and robot forward kinematics plus corrected keypoints.
- Train a latent video diffusion transformer conditioned on initial observation, text prompt, and kinematic controls; distill to a causal few-step student and extend to streaming for real-time, minute-long rollouts.
Limitation
The authors state that their kinematic interface captures camera and manipulator motion but not physical contact or object dynamics; coupling the world model with simulation-ready, physically grounded 3D assets and scenes is a promising direction toward modeling underlying dynamics.
Abstract (from arXiv)
World models enable embodied agents to anticipate action consequences, yet jointly modeling locomotion and manipulation across diverse embodiments remains challenging due to heterogeneous action spaces and limited multimodal supervision. We present Ropedia W0, the first egocentric loco-manipulation world foundation model spanning humans and robots with diverse embodiments. We curate Xp-Data, comprising 10k hours of egocentric human data and 3k hours of robot data with unified multimodal annotations and systematic quality control. A unified kinematic representation encodes locomotion through camera poses and manipulation through projected skeletons, with robot skeletons combining embodiment-specific forward kinematics with extensive manual keypoint annotation and correction. We also introduce Xp-Teleop, a long-horizon loco-manipulation benchmark based on high-quality dexterous humanoid teleoperation data. Evaluations demonstrate state-of-the-art performance across five cross-embodiment benchmarks (DreamDojo, $\Psi_0$, $\omega$-HOME, Xp-Teleop, and AgiBot) and on held-out human egocentric video, from zero-shot inference to few-shot and full-data post-training. Scaling human pre-training data improves performance on both human and robot evaluations. After distillation, our model supports real-time, minute-long rollouts, enabling live VR interaction. Applications in policy evaluation, model-based planning, and world-action modeling further demonstrate the model's utility for embodied decision-making.
Related papers
- Being-M0.7: A Latent World-Action Model for Humanoid Robots
- UNITAS: A 3D-Native World Action Model for Embodied Manipulation
- Cross-Embodiment Robot Foundation World Models with Latent Actions
- WorldLine: Action-Driven Visual Simulation for Robotic Manipulation
- Language-Forcing: Advancing Humanoid Loco-Manipulation via Semantic Alignment of Whole-Body World Action Models
Publication record
Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.
First public release: 2026-10-19 · Indexed by Pub-AI: 2026-10-20 UTC
Ropedia W0: Scaling Loco-Manipulation World Models across Embodiments with Human Experience
Preprint posted: 2026-10-19 · Original record ↗
- arxiv
- 2610.26855
Explicit source versions (1)
- v1 · 2026-10-19
- Preprint available
Source date: 2026-10-19 · First detected: 2026-10-25 UTC · Historical baseline
Source evidence (1)
- Official evidence ↗ · atom:published (official metadata)
Abstract · arxiv
World models enable embodied agents to anticipate action consequences, yet jointly modeling locomotion and manipulation across diverse embodiments remains challenging due to heterogeneous action spaces and limited multimodal supervision. We present Ropedia W0, the first egocentric loco-manipulation world foundation model spanning humans and robots with diverse embodiments. We curate Xp-Data, comprising 10k hours of egocentric human data and 3k hours of robot data with unified multimodal annotations and systematic quality control. A unified kinematic representation encodes locomotion through camera poses and manipulation through projected skeletons, with robot skeletons combining embodiment-specific forward kinematics with extensive manual keypoint annotation and correction. We also introduce Xp-Teleop, a long-horizon loco-manipulation benchmark based on high-quality dexterous humanoid teleoperation data. Evaluations demonstrate state-of-the-art performance across five cross-embodiment benchmarks (DreamDojo, $Ψ_0$, $ω$-HOME, Xp-Teleop, and AgiBot) and on held-out human egocentric video, from zero-shot inference to few-shot and full-data post-training. Scaling human pre-training data improves performance on both human and robot evaluations. After distillation, our model supports real-time, minute-long rollouts, enabling live VR interaction. Applications in policy evaluation, model-based planning, and world-action modeling further demonstrate the model's utility for embodied decision-making.