外观
论文arXiv 2610.2315612 天前
ORD-WM：面向机器人操作中3D世界建模的以物体为中心的刚体动力学学习
原文链接：ORD-WM: Learning Object-Centric Rigid Dynamics for 3D World Modeling in Robotic Manipulation（arXiv 2610.23156 · 发布日期：2026-10-16）
ORD-WM: Learning Object-Centric Rigid Dynamics for 3D World Modeling in Robotic Manipulation
ORD-WM通过学习以物体为中心的刚体动力学，在仿真和真实操作场景中验证3D世界预测与重规划。
核心要点
- 现有图像和稠密3D世界模型未显式施加物体级刚体运动约束。
- 方法用SDF监督的向量神经VAE编码点云，并预测物体级SE(3)运动增量。
- ORD-WM将预测刚体变换解析作用于当前潜变量，分离几何保持与位姿变化。
- 实验显示其支持稳定保形多步预测和闭环MPC的实时重规划。
验证范围与局限：验证范围为仿真和真实世界的机器人操作场景。
英文摘要 (Abstract)
World models provide robots with a predictive mechanism for reasoning about how actions change object states in manipulation tasks. However, existing image-based and dense 3D world models do not explicitly impose object-level rigid motion, leaving rigid-object geometry preservation to be learned implicitly and making multi-step rollouts susceptible to shape drift or object-level motion inconsistency. To address this, we propose ORD-WM, a 3D world model that learns object-centric rigid dynamics for robotic manipulation. ORD-WM learns a vector-neuron variational autoencoder supervised by signed distance fields, which encodes each object point cloud as a geometry-aware latent representation composed of 3D vector features, and uses an action-conditioned dynamics model to predict object-wise SE(3) motion increments. Rather than generating unconstrained next latent states, ORD-WM analytically applies predicted object-wise rigid transformations to the current latent states, separating persistent object geometry from pose changes. Experiments show that ORD-WM enables stable, geometry-preserving multi-step prediction and supports efficient real-time replanning for closed-loop model predictive control in simulated and real-world manipulation scenarios.
作者：Yueyang Weng, Xiaopeng Zhang, Yuxi Ou, Yongjin Mu, Chengzhi Zhao, Yanjie Li
来源
本页事实对应的公开页面，以原文为准。