Dynamic Manipulation of Deformable Objects in 3D: Simulation, Benchmark and Learning Strategy

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lan, Guanzhou, Yang, Yuqi, Mathew, Anup Teejo, Nie, Feiping, Wang, Rong, Li, Xuelong, Renda, Federico, Zhao, Bin
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910964244283392
author Lan, Guanzhou
Yang, Yuqi
Mathew, Anup Teejo
Nie, Feiping
Wang, Rong
Li, Xuelong
Renda, Federico
Zhao, Bin
author_facet Lan, Guanzhou
Yang, Yuqi
Mathew, Anup Teejo
Nie, Feiping
Wang, Rong
Li, Xuelong
Renda, Federico
Zhao, Bin
contents Goal-conditioned dynamic manipulation is inherently challenging due to complex system dynamics and stringent task constraints, particularly in deformable object scenarios characterized by high degrees of freedom and underactuation. Prior methods often simplify the problem to low-speed or 2D settings, limiting their applicability to real-world 3D tasks. In this work, we explore 3D goal-conditioned rope manipulation as a representative challenge. To mitigate data scarcity, we introduce a novel simulation framework and benchmark grounded in reduced-order dynamics, which enables compact state representation and facilitates efficient policy learning. Building on this, we propose Dynamics Informed Diffusion Policy (DIDP), a framework that integrates imitation pretraining with physics-informed test-time adaptation. First, we design a diffusion policy that learns inverse dynamics within the reduced-order space, enabling imitation learning to move beyond naïve data fitting and capture the underlying physical structure. Second, we propose a physics-informed test-time adaptation scheme that imposes kinematic boundary conditions and structured dynamics priors on the diffusion process, ensuring consistency and reliability in manipulation execution. Extensive experiments validate the proposed approach, demonstrating strong performance in terms of accuracy and robustness in the learned policy.
format Preprint
id arxiv_https___arxiv_org_abs_2505_17434
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dynamic Manipulation of Deformable Objects in 3D: Simulation, Benchmark and Learning Strategy
Lan, Guanzhou
Yang, Yuqi
Mathew, Anup Teejo
Nie, Feiping
Wang, Rong
Li, Xuelong
Renda, Federico
Zhao, Bin
Robotics
Artificial Intelligence
Goal-conditioned dynamic manipulation is inherently challenging due to complex system dynamics and stringent task constraints, particularly in deformable object scenarios characterized by high degrees of freedom and underactuation. Prior methods often simplify the problem to low-speed or 2D settings, limiting their applicability to real-world 3D tasks. In this work, we explore 3D goal-conditioned rope manipulation as a representative challenge. To mitigate data scarcity, we introduce a novel simulation framework and benchmark grounded in reduced-order dynamics, which enables compact state representation and facilitates efficient policy learning. Building on this, we propose Dynamics Informed Diffusion Policy (DIDP), a framework that integrates imitation pretraining with physics-informed test-time adaptation. First, we design a diffusion policy that learns inverse dynamics within the reduced-order space, enabling imitation learning to move beyond naïve data fitting and capture the underlying physical structure. Second, we propose a physics-informed test-time adaptation scheme that imposes kinematic boundary conditions and structured dynamics priors on the diffusion process, ensuring consistency and reliability in manipulation execution. Extensive experiments validate the proposed approach, demonstrating strong performance in terms of accuracy and robustness in the learned policy.
title Dynamic Manipulation of Deformable Objects in 3D: Simulation, Benchmark and Learning Strategy
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2505.17434