SoMA: A Real-to-Sim Neural Simulator for Robotic Soft-body Manipulation

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Huang, Mu, Wang, Hui, Ren, Kerui, Xu, Linning, Zhou, Yunsong, Yu, Mulin, Dai, Bo, Pang, Jiangmiao
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866912868869341184
author Huang, Mu
Wang, Hui
Ren, Kerui
Xu, Linning
Zhou, Yunsong
Yu, Mulin
Dai, Bo
Pang, Jiangmiao
author_facet Huang, Mu
Wang, Hui
Ren, Kerui
Xu, Linning
Zhou, Yunsong
Yu, Mulin
Dai, Bo
Pang, Jiangmiao
contents Simulating deformable objects under rich interactions remains a fundamental challenge for real-to-sim robot manipulation, with dynamics jointly driven by environmental effects and robot actions. Existing simulators rely on predefined physics or data-driven dynamics without robot-conditioned control, limiting accuracy, stability, and generalization. This paper presents SoMA, a 3D Gaussian Splat simulator for soft-body manipulation. SoMA couples deformable dynamics, environmental forces, and robot joint actions in a unified latent neural space for end-to-end real-to-sim simulation. Modeling interactions over learned Gaussian splats enables controllable, stable long-horizon manipulation and generalization beyond observed trajectories without predefined physical models. SoMA improves resimulation accuracy and generalization on real-world robot manipulation by 20%, enabling stable simulation of complex tasks such as long-horizon cloth folding.
format Preprint
id arxiv_https___arxiv_org_abs_2602_02402
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SoMA: A Real-to-Sim Neural Simulator for Robotic Soft-body Manipulation
Huang, Mu
Wang, Hui
Ren, Kerui
Xu, Linning
Zhou, Yunsong
Yu, Mulin
Dai, Bo
Pang, Jiangmiao
Robotics
Artificial Intelligence
Computer Vision and Pattern Recognition
Applied Physics
68T05, 68T40, 68U05
I.2.9; I.2.6; I.2.10
Simulating deformable objects under rich interactions remains a fundamental challenge for real-to-sim robot manipulation, with dynamics jointly driven by environmental effects and robot actions. Existing simulators rely on predefined physics or data-driven dynamics without robot-conditioned control, limiting accuracy, stability, and generalization. This paper presents SoMA, a 3D Gaussian Splat simulator for soft-body manipulation. SoMA couples deformable dynamics, environmental forces, and robot joint actions in a unified latent neural space for end-to-end real-to-sim simulation. Modeling interactions over learned Gaussian splats enables controllable, stable long-horizon manipulation and generalization beyond observed trajectories without predefined physical models. SoMA improves resimulation accuracy and generalization on real-world robot manipulation by 20%, enabling stable simulation of complex tasks such as long-horizon cloth folding.
title SoMA: A Real-to-Sim Neural Simulator for Robotic Soft-body Manipulation
topic Robotics
Artificial Intelligence
Computer Vision and Pattern Recognition
Applied Physics
68T05, 68T40, 68U05
I.2.9; I.2.6; I.2.10
url https://arxiv.org/abs/2602.02402