Newest papers

Ropedia W0: Scaling Loco-Manipulation World Models across Embodiments with Human Experience

Tianqi Liu, Hanfei Zhao, Qifeng Liang, Hui En Pang, Jiarui Xu, S Mu, 邱鹏昊, YingSheng Zhang, Jun Wang, Ziang Cao, Mingze Sun, Yukang Cao, Fangzhou Hong, Zhaoxi Chen, Ziwei Liu · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26855 · cited by 0 works (OpenAlex) · open-access PDF

French

Les modèles mondiaux d'interaction permettent aux agents incarnés d'anticiper les conséquences de l'action, mais la modélisation conjointe de la locomotion et de la manipulation à travers divers modèles d'exécution reste difficile en raison des espaces d'action hétérogènes et de la surveillance multimodale limitée. Nous présentons Ropedia W0, le premier modèle de fondation cinématique egocentrique de locomotion qui s'étend sur les humains et les robots avec divers modèles d'exécution. Nous curons Xp-Data, comprenant 10k heures de données humaines egocentriques et 3k heures de données d'évaluation des robots avec des annotations multimodales unifiées et un contrôle de qualité systématique. Une représentation

Machine translation (Meta M2M100); the English original follows.

Abstract

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.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Artificial intelligence, Robot, Humanoid robot, Kinematics, Action (physics)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange