RynnWorld-4D: 4D Embodied World Models for Robotic Manipulation

2026-07-07

AI Summary

This work addresses the challenge of open-world robotic manipulation by jointly modeling 4D spatiotemporal dynamics under scene appearance and interaction, a task where existing methods struggle to maintain consistency across geometry, appearance, and motion. The paper introduces RynnWorld-4D, the first framework to represent physical-aware 4D scenes using RGB-DF (RGB, depth, and optical flow) as input, leveraging a unified diffusion model to jointly generate future multimodal frames conditioned on language instructions for embodied 4D world modeling. The architecture features a three-branch design incorporating cross-modal attention and per-frame 3D rotational positional encoding (RoPE) to enforce spatiotemporal coherence, along with a single-step inverse dynamics policy head enabling efficient closed-loop control. Evaluated on the large-scale pseudo-labeled Rynn4DDataset 1.0, RynnWorld-4D achieves state-of-the-art performance on real-world bimanual dexterous manipulation tasks, excelling particularly in scenarios demanding high spatial precision and temporal coordination.

Abstract

Robotic manipulation in the open world requires not only recognizing what a scene looks like, but also anticipating how its 3D structure moves under interaction. We argue that synchronized RGB, depth, and optical flow, namely RGB-DF, provide a physically grounded representation that captures the underlying 4D dynamics of a scene. Compared to 2D pixel videos, this multi-modal synergy aligns visual appearance with geometric structure and temporal motion, creating a representation space significantly closer to the low-level end-effector actions demanded by robotic systems, thereby narrowing the gap between world prediction and policy learning. Building on this insight, we introduce RynnWorld-4D, a generative model that co-produces future RGB frames, depth maps, and optical flow from a single RGB-D image and a language instruction within one unified diffusion process. This 4D world model features a tri-branch architecture that integrates cross-modal attention with frame-wise 3D RoPE, ensuring that appearance, geometry, and motion evolve consistently. To supply training data at scale, we curate Rynn4DDataset 1.0, a massive dataset of over 254.4 million frames across egocentric human and robotic manipulation videos with high-quality pseudo-labels for depth and optical flow. We further propose RynnWorld-4D-Policy, an inverse dynamics head that consumes the internal 4D representations of RynnWorld-4D in a single forward pass, bypassing expensive multi-step denoising, to output robot actions in a closed-loop manner. Experiments show that RynnWorld-4D produces temporally and spatially coherent 4D predictions, and that RynnWorld-4D-Policy achieves state-of-the-art performance on real-world dexterous bimanual manipulation tasks, particularly excelling in tasks demanding spatial precision and temporal coordination.

Problem

• 4D world models
• robotic manipulation
• open-world perception
• multimodal dynamics
• embodied AI

Innovation

• 4D world model
• RGB-DF representation
• unified diffusion process
• RynnWorld-4D-Policy
• Rynn4DDataset