# DWMP：利用双世界模型实现人形机器人障碍穿越

> 原文链接：https://robotics.hellobike.com/papers/2026-09-11-2609-12347
> 频道：papers | 更新日期：2026-09-11 | 标签：arxiv, cs.RO
> 核心摘要：针对人形机器人依靠本体与视觉穿越杂乱障碍场的问题，DWMP融合双世界模型表征，并在仿真和Unitree G1真机验证。

# DWMP：利用双世界模型实现人形机器人障碍穿越

::: v-pre
> 原文链接：[DWMP: Leveraging Dual World Models for Humanoid Obstacle Traversal](https://arxiv.org/abs/2609.12347)（arXiv 2609.12347 · 发布日期：2026-09-11）

**DWMP: Leveraging Dual World Models for Humanoid Obstacle Traversal**

针对人形机器人依靠本体与视觉穿越杂乱障碍场的问题，DWMP融合双世界模型表征，并在仿真和Unitree G1真机验证。

## 核心要点

- 现有方法通常未显式考虑本体观测与自我中心视觉观测的不同特性。
- DWMP用Koopman动力学世界模型提取近似线性的本体时序特征。
- RSSM视觉世界模型将自我中心深度观测压缩为保留障碍几何的随机状态。
- 仿真和Unitree G1实验显示，DWMP较基线提升障碍穿越性能。

**验证范围与局限**：验证范围为仿真和一台Unitree G1人形机器人上的随机障碍布局穿越。

## 英文摘要 (Abstract)

Humanoid robots must traverse cluttered obstacle fields using onboard proprioceptive and visual observations, yet existing methods usually process multimodal observations without explicitly considering their different characteristics: proprioceptive observations are low-dimensional but governed by highly nonlinear robot dynamics, while egocentric visual observations are high-dimensional, noisy, and redundant. We propose DWMP (Dual World Model Policy), a framework that provides the actor with separate but complementary world-model representations for humanoid obstacle traversal. A Koopman-based dynamics world model lifts proprioceptive observations into a latent space where their temporal evolution is approximately linear, making the dynamics features easier for the actor to learn from. An RSSM-based visual world model compresses egocentric depth observations into compact stochastic states while preserving obstacle-related geometry. The student policy receives the fused latent representation for action generation, combining linearized proprioceptive dynamics with compressed visual perception. Experiments in simulation and on a Unitree G1 humanoid robot show that DWMP improves obstacle traversal performance over baselines and supports real-world deployment under randomized obstacle layouts.

作者：Rong Jin, Jianming Ma, Yue Gao
:::

## 原始参考来源 (Sources)

- [arXiv 2609.12347：DWMP: Leveraging Dual World Models for Humanoid Obstacle Traversal](https://arxiv.org/abs/2609.12347)