RoboDream: Compositional World Models for Scalable Robot Data Synthesis

Published
Source
arXiv
Paper number
287
Field
Robotics
arXiv ID
2606.02577

Key points

  • This paper proposes a generalizable body-centric world model that achieves scalable data generation by synthesizing realistic demonstrations in new objects, new scenes, and new viewpoints.
  • The approach anchors generation to rendered robot motions while conditioning on explicit scene and object priors, effectively separating trajectory execution from environment synthesis.
  • Real-world experiments show that the generated data consistently improves downstream policy performance across diverse manipulation tasks and substantially reduces the need for real-world data.

Paper links

External research summaries. These are not HDATF publications or measured product results.

Read original (opens in a new tab)