DeformX: A Versatile Co-Simulation Framework for Deformable Linear Objects

Chenhao Li; Henry Kou; Howie Choset; Lehong Wang; Lu Li; Xiang Fei; Yi Yang; Zilin Dai

arxiv: 2606.22116 · v1 · pith:LARWBO5Nnew · submitted 2026-06-20 · 💻 cs.RO

DeformX: A Versatile Co-Simulation Framework for Deformable Linear Objects

Yi Yang , Xiang Fei , Lehong Wang , Chenhao Li , Zilin Dai , Henry Kou , Lu Li , Howie Choset This is my paper

Pith reviewed 2026-06-26 11:52 UTC · model grok-4.3

classification 💻 cs.RO

keywords deformable linear objectscosserat rodco-simulationrobot manipulationsim-to-real transfermesh skinningphysics-based simulation

0 comments

The pith

DeformX combines a Cosserat rod physics engine with visual rendering to simulate deformable linear objects with both mechanical accuracy and visual fidelity for robotic tasks.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents DeformX as a co-simulation framework for wires, cables, and ropes that pairs a dedicated Cosserat rod engine with visual simulation to produce physically grounded deformations. The engine models bending, twisting, shear, self-collisions, and contacts with arbitrary meshes, while mesh skinning transfers those states onto imported CAD models for realism. The resulting simulations support synthetic data generation and policy learning, with reported improvements in downstream real-world segmentation accuracy and policy transfer error.

Core claim

DeformX integrates a dedicated Cosserat rod physics engine with visual simulation capabilities to enable DLO simulations that are both physically faithful and visually realistic, where the engine simulates dynamics and self-collisions plus contacts with free-form meshes and mesh skinning maps discrete rod deformations onto CAD models, providing one of the first frameworks that unifies realistic visualization, principled physics, and compatibility with robot learning pipelines.

What carries the argument

The Cosserat rod engine that computes dynamics, self-collisions, and contacts with arbitrary free-form meshes, paired with mesh skinning to map rod states onto visual CAD models.

If this is right

Synthetic data produced by the framework improves fine-tuning performance of image segmentation models on real photographs by 10.2 percent mAP@75.
Control policies trained entirely inside the simulation transfer to physical robot hardware with a mean target-hitting error of 6.6 cm.
The framework supports simulation of DLO interactions with arbitrary free-form meshes while maintaining both physical and visual consistency.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same engine-plus-skinning structure could be applied to other slender-object tasks such as knotting or cable routing without changing the core simulation loop.
If the rod engine runs at interactive rates, the framework could support online replanning loops that close the loop between perception and action on physical hardware.
Extending the contact model to include friction parameters derived from material measurements would allow direct comparison of simulated versus measured sliding behavior.

Load-bearing premise

The dedicated Cosserat rod model accurately captures the real-world bending, twisting, shear, and collision behaviors of slender elastic objects without significant inaccuracies.

What would settle it

Side-by-side measurements of a physical DLO's deformed shape and contact forces under controlled loads compared against the corresponding simulated outputs.

Figures

Figures reproduced from arXiv: 2606.22116 by Chenhao Li, Henry Kou, Howie Choset, Lehong Wang, Lu Li, Xiang Fei, Yi Yang, Zilin Dai.

**Figure 1.** Figure 1: DeformX is a co-simulation framework for deformable linear objects (DLOs) that integrates NVIDIA Isaac Sim with [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗

**Figure 2.** Figure 2: Cosserat rod modeling of DLOs. (a) Continuous [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Figure 3: Overview of our co-simulation framework. Compared to existing simulators for vision (left) and RL (right), our [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

**Figure 5.** Figure 5: Synthetic wire segmentation dataset generation [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗

**Figure 6.** Figure 6: Representative images from three settings across [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗

**Figure 7.** Figure 7: (a) Robot-driven rope experiment with one end [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗

**Figure 9.** Figure 9: All-data distribution of the per-image Jaccard score [PITH_FULL_IMAGE:figures/full_fig_p010_9.png] view at source ↗

**Figure 10.** Figure 10: Distribution of the per-image Jaccard score [PITH_FULL_IMAGE:figures/full_fig_p010_10.png] view at source ↗

**Figure 11.** Figure 11: Hit-apple demonstration: temporal montage of two trials (rows) as a UR5e swings a rope through Initial [PITH_FULL_IMAGE:figures/full_fig_p011_11.png] view at source ↗

read the original abstract

Deformable linear objects (DLOs) such as wires, cables, and ropes are common in robotic manipulation tasks, yet simulating them with both visual realism and physical accuracy remains challenging. Existing visual simulation methods typically rely on procedural geometric primitives that lack physically grounded deformation behavior, while physics-based approaches with robot learning support often approximate DLOs as rigid-link chains or generic soft bodies, failing to accurately capture the bending, twisting, and shear mechanics of slender elastic structures. In this work, we introduce DeformX, a co-simulation framework that integrates a dedicated Cosserat rod physics engine with NVIDIA Isaac Sim, enabling DLO simulations that are both physically faithful and visually realistic. Our Cosserat rod engine simulates the dynamics and self-collisions of DLOs, and contact interactions with arbitrary free-form meshes. To achieve high-fidelity visualization, we employ mesh skinning to map discrete rod deformations onto imported CAD models. To the best of our knowledge, DeformX is the one of the first frameworks for DLO simulation that unifies realistic visualization, principled physics, and compatibility with robot learning pipelines. We demonstrate its versatility across synthetic data generation and policy learning for DLO manipulation, and validate visual and physical fidelity through comparisons against real-world experiments. Notably, fine-tuning Segment Anything Model 3 (SAM3) on DeformX-generated data yields a 10.2% mAP@75 improvement in real-image wire segmentation, and a rope-swinging policy trained entirely in DeformX achieves a mean target-hitting error of 6.6 cm on a UR5e manipulator in real-world trials, highlighting its strong sim-to-real transfer capability.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

DeformX integrates a Cosserat rod engine into Isaac Sim with mesh skinning and backs it with real-world validation on segmentation and policy transfer.

read the letter

The main point is that DeformX puts a dedicated Cosserat rod physics engine inside NVIDIA Isaac Sim and uses mesh skinning to map rod states onto imported CAD models for DLOs. The paper shows this setup handles bending, twisting, shear, self-collisions, and contacts with free-form meshes, then demonstrates it on two tasks: fine-tuning SAM3 on generated data for a 10.2% mAP@75 gain on real wire images, and training a rope-swinging policy that transfers to a UR5e with 6.6 cm mean error.

The work does a solid job on the integration details and the real-world comparisons. The experiments directly test the claimed mechanics and tie the performance numbers to the Cosserat component plus skinning, without obvious internal contradictions or circular definitions. The full manuscript supplies the implementation steps that make the unification argument concrete rather than just asserted.

Soft spots are minor and scoped. The novelty sits in the specific co-simulation architecture rather than a new physics model, so the "one of the first" phrasing holds only for this combination of tools and visualization layer. The policy result covers one task, which is enough to show transfer but leaves broader versatility as a claim that would need more examples to fully land. No load-bearing flaws appear in the reported validation.

This is for robotics people who need realistic DLO simulation inside existing learning pipelines. Anyone doing sim-to-real for cables, wires, or ropes will get practical value from the metrics and setup. It deserves serious peer review because the evidence supports the core claims and the framework fills a usable gap without relying on untested assumptions.

Referee Report

0 major / 0 minor

Summary. The manuscript introduces DeformX, a co-simulation framework that integrates a dedicated Cosserat rod physics engine with NVIDIA Isaac Sim for deformable linear objects (DLOs). The engine handles dynamics, self-collisions, and contacts with arbitrary free-form meshes; mesh skinning maps rod deformations to imported CAD models for visualization. The work positions DeformX as one of the first frameworks unifying realistic visualization, principled physics, and robot learning pipeline compatibility. It demonstrates utility via synthetic data generation for SAM3 fine-tuning (10.2% mAP@75 gain on real wire segmentation) and end-to-end policy learning for rope swinging (6.6 cm mean target-hitting error on a real UR5e after sim-only training), with direct real-world validation experiments.

Significance. If the reported physical and visual fidelity claims hold, the framework addresses a recognized gap in DLO simulation by combining Cosserat-rod mechanics with high-fidelity rendering and learning compatibility inside a widely used robotics simulator. The concrete sim-to-real metrics (segmentation improvement and policy transfer error) provide falsifiable evidence of utility for manipulation tasks. The explicit implementation details and real-world comparison experiments strengthen the unification argument beyond prior approximations that treat DLOs as rigid chains or generic soft bodies.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment and recommendation to accept. The report accurately captures the contributions of DeformX in integrating Cosserat-rod physics with Isaac Sim for DLOs, along with the reported sim-to-real results on segmentation and policy transfer.

Circularity Check

0 steps flagged

No significant circularity; framework integrates external components with external validation

full rationale

The manuscript describes DeformX as a co-simulation integration of a pre-existing Cosserat rod engine, NVIDIA Isaac Sim, and standard mesh skinning, with no equations, derivations, or fitted parameters presented. Claims of unification and sim-to-real performance rest on implementation details plus direct real-world comparisons (SAM3 mAP improvement, 6.6 cm policy error), not on any self-referential definitions, self-citation chains, or renaming of known results. The 'to the best of our knowledge' novelty statement is non-load-bearing and does not invoke prior author work as a uniqueness theorem. No load-bearing step reduces to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no details on specific parameters, mathematical axioms, or new entities introduced beyond the framework name and components.

pith-pipeline@v0.9.1-grok · 5860 in / 1403 out tokens · 35863 ms · 2026-06-26T11:52:06.151448+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

30 extracted references · 7 canonical work pages

[1]

Robotic manipulation and sensing of deformable objects in domestic and industrial applications: a survey,

J. Sanchez, J.-A. Corrales, B.-C. Bouzgarrou, and Y . Mezouar, “Robotic manipulation and sensing of deformable objects in domestic and industrial applications: a survey,”The International Journal of Robotics Research, vol. 37, no. 7, pp. 688–716, 2018. [Online]. Available: https://doi.org/10.1177/0278364918779698

work page doi:10.1177/0278364918779698 2018
[2]

Handloom: Learned tracing of one-dimensional objects for inspection and manipulation,

V . Viswanath, K. Shivakumar, M. Parulekar, J. Ajmera, J. Kerr, J. Ichnowski, R. Cheng, T. Kollar, and K. Goldberg, “Handloom: Learned tracing of one-dimensional objects for inspection and manipulation,” inProceedings of The 7th Conference on Robot Learning, ser. Proceedings of Machine Learning Research, J. Tan, M. Toussaint, and K. Darvish, Eds., vol. 22...

2023
[3]

Generation of realistic synthetic cable images to train deep learning segmentation models,

P. MalvidoFresnillo, W. M. Mohammed, S. Vasudevan, J. A. PerezGarcia, and J. L. MartinezLastra, “Generation of realistic synthetic cable images to train deep learning segmentation models,” Machine Vision and Applications, vol. 35, no. 4, p. 84, Jun 2024. [Online]. Available: https://doi.org/10.1007/s00138-024-01562-y

work page doi:10.1007/s00138-024-01562-y 2024
[4]

Fastdlo: Fast deformable linear objects instance segmentation,

A. Caporali, K. Galassi, R. Zanella, and G. Palli, “Fastdlo: Fast deformable linear objects instance segmentation,”IEEE Robotics and Automation Letters, vol. 7, no. 4, pp. 9075–9082, 2022

2022
[5]

Auto-generated wires dataset for semantic segmentation with domain-independence,

R. Zanella, A. Caporali, K. Tadaka, D. D. Gregorio, and G. Palli, “Auto-generated wires dataset for semantic segmentation with domain-independence,”2021 International Conference on Computer, Control and Robotics (ICCCR), pp. 292–298, 2021. [Online]. Available: https://api.semanticscholar.org/CorpusID:231915434

2021
[6]

Iscute: Instance segmentation of cables using text embedding,

S. Kozlovsky, O. Joglekar, and D. D. Castro, “Iscute: Instance segmentation of cables using text embedding,” 2024. [Online]. Available: https://arxiv.org/abs/2402.11996

arXiv 2024
[7]

Dexdlo: Learning goal- conditioned dexterous policy for dynamic manipulation of deformable linear objects,

S. Zhaole, J. Zhu, and R. B. Fisher, “Dexdlo: Learning goal- conditioned dexterous policy for dynamic manipulation of deformable linear objects,” in2024 IEEE International Conference on Robotics and Automation (ICRA), 2024, pp. 16 009–16 015

2024
[8]

Iterative residual policy: for goal-conditioned dynamic manipulation of deformable objects,

C. Chi, B. Burchfiel, E. Cousineau, S. Feng, and S. Song, “Iterative residual policy: for goal-conditioned dynamic manipulation of deformable objects,” 2022. [Online]. Available: https://arxiv.org/ abs/2203.00663

arXiv 2022
[9]

Real2sim2real: Self-supervised learning of physical single-step dynamic actions for planar robot casting,

V . Lim, H. Huang, L. Y . Chen, J. Wang, J. Ichnowski, D. Seita, M. Laskey, and K. Goldberg, “Real2sim2real: Self-supervised learning of physical single-step dynamic actions for planar robot casting,” in 2022 International Conference on Robotics and Automation (ICRA), 2022, pp. 8282–8289

2022
[10]

Flexible sim- ulation of deformable models using discontinuous galerkin fem,

P. Kaufmann, S. Martin, M. Botsch, and M. Gross, “Flexible sim- ulation of deformable models using discontinuous galerkin fem,” Graphical Models, vol. 71, no. 4, pp. 153–167, 2009

2009
[11]

An efficient fe model of slender members for crash analysis of cable barriers,

Q. Wang, H. Fang, N. Li, D. C. Weggel, and G. Wen, “An efficient fe model of slender members for crash analysis of cable barriers,” Engineering structures, vol. 52, pp. 240–256, 2013

2013
[12]

Physically based real-time interactive assembly simulation of cable harness,

N. Lv, J. Liu, X. Ding, J. Liu, H. Lin, and J. Ma, “Physically based real-time interactive assembly simulation of cable harness,”Journal of Manufacturing Systems, vol. 43, pp. 385–399, 2017

2017
[13]

Modeling elastic beams using dynamic splines,

P. P. Valentini and E. Pennestr `ı, “Modeling elastic beams using dynamic splines,”Multibody system dynamics, vol. 25, no. 3, pp. 271– 284, 2011

2011
[14]

Nonlinear dynamics of elastic rods using the cosserat theory: Modelling and simulation,

D. Cao and R. W. Tucker, “Nonlinear dynamics of elastic rods using the cosserat theory: Modelling and simulation,”International Journal of Solids and Structures, vol. 45, no. 2, pp. 460–477, 2008. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S0020768307003253

2008
[15]

Discrete cosserat rod models based on the difference geometry of framed curves for interactive simulation of flexible cables,

J. Linn and K. Dreßler, “Discrete cosserat rod models based on the difference geometry of framed curves for interactive simulation of flexible cables,” inMath for the Digital Factory. Springer, 2017, pp. 289–319

2017
[16]

Discrete cosserat approach for multisection soft manipulator dynamics,

F. Renda, F. Boyer, J. Dias, and L. Seneviratne, “Discrete cosserat approach for multisection soft manipulator dynamics,”IEEE Transac- tions on Robotics, vol. 34, no. 6, pp. 1518–1533, 2018

2018
[17]

A review of techniques for modeling flexible cables,

N. Lv, J. Liu, H. Xia, J. Ma, and X. Yang, “A review of techniques for modeling flexible cables,”Computer-Aided Design, vol. 122, p. 102826, 2020. [Online]. Available: https: //www.sciencedirect.com/science/article/pii/S0010448520300191

2020
[18]

Elastica: A compliant mechanics environment for soft robotic control,

N. Naughton, J. Sun, A. Tekinalp, T. Parthasarathy, G. Chowdhary, and M. Gazzola, “Elastica: A compliant mechanics environment for soft robotic control,”IEEE Robotics and Automation Letters, vol. 6, no. 2, pp. 3389–3396, 2021

2021
[19]

Multistage cable routing through hierarchical imitation learning,

J. Luo, C. Xu, X. Geng, G. Feng, K. Fang, L. Tan, S. Schaal, and S. Levine, “Multistage cable routing through hierarchical imitation learning,”IEEE Transactions on Robotics, vol. 40, pp. 1476–1491, 2024

2024
[20]

Nonlinear problems of elasticity (stuart s. antman),

M. Renardy, “Nonlinear problems of elasticity (stuart s. antman),” SIAM Review, vol. 37, no. 4, pp. 637–637, 1995. [Online]. Available: https://doi.org/10.1137/1037152

work page doi:10.1137/1037152 1995
[21]

and Dudte, L

M. Gazzola, L. H. Dudte, A. G. McCormick, and L. Mahadevan, “Forward and inverse problems in the mechanics of soft filaments,” Royal Society Open Science, vol. 5, no. 6, p. 171628, 06 2018. [Online]. Available: https://doi.org/10.1098/rsos.171628

work page doi:10.1098/rsos.171628 2018
[22]

Discrete elastic rods,

M. Bergou, M. Wardetzky, S. Robinson, B. Audoly, and E. Grinspun, “Discrete elastic rods,” inACM SIGGRAPH 2008 Papers, ser. SIGGRAPH ’08. New York, NY , USA: Association for Computing Machinery, 2008. [Online]. Available: https://doi.org/10.1145/1399504.1360662

work page doi:10.1145/1399504.1360662 2008
[23]

Elastica: Cosserat rods,

Gazzola Lab, “Elastica: Cosserat rods,” Website, developed and maintained by the Gazzola Lab at the University of Illinois at Urbana- Champaign. [Online]. Available: https://www.cosseratrods.org/
[24]

Gazzolalab/pyelastica,

A. Tekinalp, S. H. Kim, Y . Bhosale, T. Parthasarathy, N. Naughton, A. Albazroun, R. Joon, S. Cui, I. Nasiriziba, M. St ¨olzle, C.-H. C. Shih, and M. Gazzola, “Gazzolalab/pyelastica,” 2024. [Online]. Available: https://doi.org/10.5281/zenodo.7658871

work page doi:10.5281/zenodo.7658871 2024
[25]

Stable cosserat rods,

J. Hsu, T. Wang, K. Wu, and C. Yuksel, “Stable cosserat rods,” inProceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers, ser. SIGGRAPH Conference Papers ’25. New York, NY , USA: Association for Computing Machinery, 2025. [Online]. Available: https://doi.org/10.1145/3721238.3730618

work page doi:10.1145/3721238.3730618 2025
[26]

Skinning mesh animations,

D. L. James and C. D. Twigg, “Skinning mesh animations,”ACM Transactions on Graphics (TOG), vol. 24, no. 3, pp. 399–407, 2005

2005
[27]

White coexflex 60a tpu,

COEX 3D, “White coexflex 60a tpu,” accessed: 2026-03-05. [Online]. Available: https://coex3d.com/products/white-coexflex-60a-tpu

2026
[28]

Data center assets pack,

NVIDIA Corporation, “Data center assets pack,” Online documentation, NVIDIA Omniverse, accessed: 2026-03-06. [Online]. Available: https://docs.omniverse.nvidia.com/usd/latest/ usd content samples/downloadable packs.html

2026
[29]

Trackdlo: Tracking deformable linear objects under occlusion with motion coherence,

J. Xiang, H. Dinkel, H. Zhao, N. Gao, B. Coltin, T. Smith, and T. Bretl, “Trackdlo: Tracking deformable linear objects under occlusion with motion coherence,”IEEE Robotics and Automation Letters, vol. 8, no. 10, pp. 6179–6186, 2023

2023
[30]

Depth anything 3: Recovering the visual space from any views,

H. Lin, S. Chen, J. Liew, D. Y . Chen, Z. Li, G. Shi, J. Feng, and B. Kang, “Depth anything 3: Recovering the visual space from any views,” 2025. [Online]. Available: https://arxiv.org/abs/2511.10647 VII. APPENDIX This appendix details theWireSeg-36kdataset, the per- scene behaviour of the SAM3 baseline, RL formulation of the planar hit-target rope-swingi...

Pith/arXiv arXiv 2025

[1] [1]

Robotic manipulation and sensing of deformable objects in domestic and industrial applications: a survey,

J. Sanchez, J.-A. Corrales, B.-C. Bouzgarrou, and Y . Mezouar, “Robotic manipulation and sensing of deformable objects in domestic and industrial applications: a survey,”The International Journal of Robotics Research, vol. 37, no. 7, pp. 688–716, 2018. [Online]. Available: https://doi.org/10.1177/0278364918779698

work page doi:10.1177/0278364918779698 2018

[2] [2]

Handloom: Learned tracing of one-dimensional objects for inspection and manipulation,

V . Viswanath, K. Shivakumar, M. Parulekar, J. Ajmera, J. Kerr, J. Ichnowski, R. Cheng, T. Kollar, and K. Goldberg, “Handloom: Learned tracing of one-dimensional objects for inspection and manipulation,” inProceedings of The 7th Conference on Robot Learning, ser. Proceedings of Machine Learning Research, J. Tan, M. Toussaint, and K. Darvish, Eds., vol. 22...

2023

[3] [3]

Generation of realistic synthetic cable images to train deep learning segmentation models,

P. MalvidoFresnillo, W. M. Mohammed, S. Vasudevan, J. A. PerezGarcia, and J. L. MartinezLastra, “Generation of realistic synthetic cable images to train deep learning segmentation models,” Machine Vision and Applications, vol. 35, no. 4, p. 84, Jun 2024. [Online]. Available: https://doi.org/10.1007/s00138-024-01562-y

work page doi:10.1007/s00138-024-01562-y 2024

[4] [4]

Fastdlo: Fast deformable linear objects instance segmentation,

A. Caporali, K. Galassi, R. Zanella, and G. Palli, “Fastdlo: Fast deformable linear objects instance segmentation,”IEEE Robotics and Automation Letters, vol. 7, no. 4, pp. 9075–9082, 2022

2022

[5] [5]

Auto-generated wires dataset for semantic segmentation with domain-independence,

R. Zanella, A. Caporali, K. Tadaka, D. D. Gregorio, and G. Palli, “Auto-generated wires dataset for semantic segmentation with domain-independence,”2021 International Conference on Computer, Control and Robotics (ICCCR), pp. 292–298, 2021. [Online]. Available: https://api.semanticscholar.org/CorpusID:231915434

2021

[6] [6]

Iscute: Instance segmentation of cables using text embedding,

S. Kozlovsky, O. Joglekar, and D. D. Castro, “Iscute: Instance segmentation of cables using text embedding,” 2024. [Online]. Available: https://arxiv.org/abs/2402.11996

arXiv 2024

[7] [7]

Dexdlo: Learning goal- conditioned dexterous policy for dynamic manipulation of deformable linear objects,

S. Zhaole, J. Zhu, and R. B. Fisher, “Dexdlo: Learning goal- conditioned dexterous policy for dynamic manipulation of deformable linear objects,” in2024 IEEE International Conference on Robotics and Automation (ICRA), 2024, pp. 16 009–16 015

2024

[8] [8]

Iterative residual policy: for goal-conditioned dynamic manipulation of deformable objects,

C. Chi, B. Burchfiel, E. Cousineau, S. Feng, and S. Song, “Iterative residual policy: for goal-conditioned dynamic manipulation of deformable objects,” 2022. [Online]. Available: https://arxiv.org/ abs/2203.00663

arXiv 2022

[9] [9]

Real2sim2real: Self-supervised learning of physical single-step dynamic actions for planar robot casting,

V . Lim, H. Huang, L. Y . Chen, J. Wang, J. Ichnowski, D. Seita, M. Laskey, and K. Goldberg, “Real2sim2real: Self-supervised learning of physical single-step dynamic actions for planar robot casting,” in 2022 International Conference on Robotics and Automation (ICRA), 2022, pp. 8282–8289

2022

[10] [10]

Flexible sim- ulation of deformable models using discontinuous galerkin fem,

P. Kaufmann, S. Martin, M. Botsch, and M. Gross, “Flexible sim- ulation of deformable models using discontinuous galerkin fem,” Graphical Models, vol. 71, no. 4, pp. 153–167, 2009

2009

[11] [11]

An efficient fe model of slender members for crash analysis of cable barriers,

Q. Wang, H. Fang, N. Li, D. C. Weggel, and G. Wen, “An efficient fe model of slender members for crash analysis of cable barriers,” Engineering structures, vol. 52, pp. 240–256, 2013

2013

[12] [12]

Physically based real-time interactive assembly simulation of cable harness,

N. Lv, J. Liu, X. Ding, J. Liu, H. Lin, and J. Ma, “Physically based real-time interactive assembly simulation of cable harness,”Journal of Manufacturing Systems, vol. 43, pp. 385–399, 2017

2017

[13] [13]

Modeling elastic beams using dynamic splines,

P. P. Valentini and E. Pennestr `ı, “Modeling elastic beams using dynamic splines,”Multibody system dynamics, vol. 25, no. 3, pp. 271– 284, 2011

2011

[14] [14]

Nonlinear dynamics of elastic rods using the cosserat theory: Modelling and simulation,

D. Cao and R. W. Tucker, “Nonlinear dynamics of elastic rods using the cosserat theory: Modelling and simulation,”International Journal of Solids and Structures, vol. 45, no. 2, pp. 460–477, 2008. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S0020768307003253

2008

[15] [15]

Discrete cosserat rod models based on the difference geometry of framed curves for interactive simulation of flexible cables,

J. Linn and K. Dreßler, “Discrete cosserat rod models based on the difference geometry of framed curves for interactive simulation of flexible cables,” inMath for the Digital Factory. Springer, 2017, pp. 289–319

2017

[16] [16]

Discrete cosserat approach for multisection soft manipulator dynamics,

F. Renda, F. Boyer, J. Dias, and L. Seneviratne, “Discrete cosserat approach for multisection soft manipulator dynamics,”IEEE Transac- tions on Robotics, vol. 34, no. 6, pp. 1518–1533, 2018

2018

[17] [17]

A review of techniques for modeling flexible cables,

N. Lv, J. Liu, H. Xia, J. Ma, and X. Yang, “A review of techniques for modeling flexible cables,”Computer-Aided Design, vol. 122, p. 102826, 2020. [Online]. Available: https: //www.sciencedirect.com/science/article/pii/S0010448520300191

2020

[18] [18]

Elastica: A compliant mechanics environment for soft robotic control,

N. Naughton, J. Sun, A. Tekinalp, T. Parthasarathy, G. Chowdhary, and M. Gazzola, “Elastica: A compliant mechanics environment for soft robotic control,”IEEE Robotics and Automation Letters, vol. 6, no. 2, pp. 3389–3396, 2021

2021

[19] [19]

Multistage cable routing through hierarchical imitation learning,

J. Luo, C. Xu, X. Geng, G. Feng, K. Fang, L. Tan, S. Schaal, and S. Levine, “Multistage cable routing through hierarchical imitation learning,”IEEE Transactions on Robotics, vol. 40, pp. 1476–1491, 2024

2024

[20] [20]

Nonlinear problems of elasticity (stuart s. antman),

M. Renardy, “Nonlinear problems of elasticity (stuart s. antman),” SIAM Review, vol. 37, no. 4, pp. 637–637, 1995. [Online]. Available: https://doi.org/10.1137/1037152

work page doi:10.1137/1037152 1995

[21] [21]

and Dudte, L

M. Gazzola, L. H. Dudte, A. G. McCormick, and L. Mahadevan, “Forward and inverse problems in the mechanics of soft filaments,” Royal Society Open Science, vol. 5, no. 6, p. 171628, 06 2018. [Online]. Available: https://doi.org/10.1098/rsos.171628

work page doi:10.1098/rsos.171628 2018

[22] [22]

Discrete elastic rods,

M. Bergou, M. Wardetzky, S. Robinson, B. Audoly, and E. Grinspun, “Discrete elastic rods,” inACM SIGGRAPH 2008 Papers, ser. SIGGRAPH ’08. New York, NY , USA: Association for Computing Machinery, 2008. [Online]. Available: https://doi.org/10.1145/1399504.1360662

work page doi:10.1145/1399504.1360662 2008

[23] [23]

Elastica: Cosserat rods,

Gazzola Lab, “Elastica: Cosserat rods,” Website, developed and maintained by the Gazzola Lab at the University of Illinois at Urbana- Champaign. [Online]. Available: https://www.cosseratrods.org/

[24] [24]

Gazzolalab/pyelastica,

A. Tekinalp, S. H. Kim, Y . Bhosale, T. Parthasarathy, N. Naughton, A. Albazroun, R. Joon, S. Cui, I. Nasiriziba, M. St ¨olzle, C.-H. C. Shih, and M. Gazzola, “Gazzolalab/pyelastica,” 2024. [Online]. Available: https://doi.org/10.5281/zenodo.7658871

work page doi:10.5281/zenodo.7658871 2024

[25] [25]

Stable cosserat rods,

J. Hsu, T. Wang, K. Wu, and C. Yuksel, “Stable cosserat rods,” inProceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers, ser. SIGGRAPH Conference Papers ’25. New York, NY , USA: Association for Computing Machinery, 2025. [Online]. Available: https://doi.org/10.1145/3721238.3730618

work page doi:10.1145/3721238.3730618 2025

[26] [26]

Skinning mesh animations,

D. L. James and C. D. Twigg, “Skinning mesh animations,”ACM Transactions on Graphics (TOG), vol. 24, no. 3, pp. 399–407, 2005

2005

[27] [27]

White coexflex 60a tpu,

COEX 3D, “White coexflex 60a tpu,” accessed: 2026-03-05. [Online]. Available: https://coex3d.com/products/white-coexflex-60a-tpu

2026

[28] [28]

Data center assets pack,

NVIDIA Corporation, “Data center assets pack,” Online documentation, NVIDIA Omniverse, accessed: 2026-03-06. [Online]. Available: https://docs.omniverse.nvidia.com/usd/latest/ usd content samples/downloadable packs.html

2026

[29] [29]

Trackdlo: Tracking deformable linear objects under occlusion with motion coherence,

J. Xiang, H. Dinkel, H. Zhao, N. Gao, B. Coltin, T. Smith, and T. Bretl, “Trackdlo: Tracking deformable linear objects under occlusion with motion coherence,”IEEE Robotics and Automation Letters, vol. 8, no. 10, pp. 6179–6186, 2023

2023

[30] [30]

Depth anything 3: Recovering the visual space from any views,

H. Lin, S. Chen, J. Liew, D. Y . Chen, Z. Li, G. Shi, J. Feng, and B. Kang, “Depth anything 3: Recovering the visual space from any views,” 2025. [Online]. Available: https://arxiv.org/abs/2511.10647 VII. APPENDIX This appendix details theWireSeg-36kdataset, the per- scene behaviour of the SAM3 baseline, RL formulation of the planar hit-target rope-swingi...

Pith/arXiv arXiv 2025