Pith. sign in

REVIEW 3 major objections 3 minor 55 references

Representative Volume Element: Existence and Extent in Cracked Heterogeneous Medium

T0 review · 3 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper claims that two modifications—strain-periodic boundary conditions and fracture-energy-matched strain localisation—make representative-volume-element simulations of cracked fibre composites yield responses that are both size-insen

desk verdict Abstract promises a serious RVE size- and mesh-sensitivity study, but the full text attached is a different paper, so no scientific verdict is possible until the correct manuscript and a non-circular equalisation protocol are on the table. read the letter →

arxiv 2508.08320 v1 pith:SJQQK6AO submitted 2025-08-09 cs.CE physics.comp-ph

Hari Sankar R , Harpreet Singh This is my paper
classification cs.CEphysics.comp-ph
keywords representativevolumeelementperiodicboundaryconditionsmeshsensitivitysizefractureenergystrainlocalisationfibre-reinforcedcompositesdamageinitiation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper is trying to establish that two technical modifications to standard practice in microscale composite failure modelling remove the two usual excuses for distrusting results: simulated response that changes with the size of the representative volume element (RVE), and simulated response that changes with the mesh. The first modification is a set of modified periodic boundary conditions (MPBCs) that enforce periodicity of strain along with periodicity of displacement on the RVE boundary; the second is a rule that equalises the fracture energy seen in the computation with the physical fracture energy by choosing the bandwidth of strain localisation. To support these claims, the paper analyses 1,200 two-dimensional RVE samples spanning five fibre volume fractions and four RVE sizes under transverse loading, and reports that MPBCs attenuate the size effect on material-response precision, especially in the inelastic regime. A separate observation addresses damage initiation: a close pair of fibres is more likely to damage between the fibres when the line joining their centres is more closely aligned with the loading direction.

What carries the argument

The load-bearing device is the modified periodic boundary condition (MPBC): in addition to enforcing displacement periodicity on opposite faces of the RVE, it enforces periodicity of strain across the boundary, and this is the feature claimed to suppress size dependence. The companion device is a fracture-energy-matching rule: the strain-localisation bandwidth is chosen so that the fracture energy dissipated in the computation equals the physical fracture energy, and this is the feature claimed to suppress mesh dependence. Together they let the computed inelastic response be precise without requiring a large RVE or a fine localisation band.

What would settle it

Run the paper's MPBC procedure on one RVE size with the strain-localisation bandwidth set a priori (say, one element width) and again with it tuned to match a target fracture energy; if the resulting stress–strain curves differ, the claimed mesh independence is calibration-dependent. Then repeat the 1,200-sample campaign with a different random realisation or a 3D microstructure; if the spread across RVE sizes reappears, the size attenuation does not generalise beyond the tested 2D transverse configurations.

Watch

Extended reading notes

Core claim

The central claim is that RVE size sensitivity and mesh sensitivity in cracked heterogeneous media can be treated as boundary-condition and energy-accounting problems rather than as unavoidable sampling errors. MPBCs—which supplement displacement periodicity with strain periodicity across the domain boundary—are formulated and verified to attenuate the effect of RVE size on the precision of the material response, particularly in the inelastic regime. The companion mesh-sensitivity technique equalises the fracture energy observed from computational analysis with the real phenomenon, making the response independent of the bandwidth of strain localisation. The paper's supporting evidence is a c

Load-bearing premise

The method hinges on the premise that equalising one scalar, fracture energy, by choosing the strain-localisation bandwidth is sufficient to make the computed response match the real phenomenon, and that this bandwidth can be chosen without tuning it to the same response curves used to demonstrate mesh independence.

Editorial extensions

If this is right

  • For transverse loading of fibre composites, the paper implies that the inelastic stress–strain response from RVE analysis is no longer tied to the particular RVE size once MPBCs are used.
  • Equalising fracture energy makes the response independent of the strain-localisation bandwidth, so analysts do not have to tune the localisation band to get reproducible results.
  • Damage-initiation predictions in 2D RVEs can be guided by a geometric rule: between two closely spaced fibres, the damage-prone region is the ligament when the fibre-centre line is nearly aligned with the load.
  • The 1,200-sample database gives the size-attenuation claim statistical breadth across five fibre volume fractions and four RVE sizes.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the size attenuation transfers to other loadings and to 3D microstructures, MPBCs could justify using smaller, cheaper RVEs in multiscale failure analysis; the paper's evidence is limited to 2D transverse loading.
  • The mesh-independence claim is only as strong as the rule that chooses the strain-localisation bandwidth; if that bandwidth is calibrated against the same response curves used to demonstrate mesh independence, the independence is partly built in by construction.
  • The fibre-pair damage observation is incomplete as stated—the abstract gives no reference value for the threshold angle—so turning it into a quantitative design rule requires fixing that threshold.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 3 minor

Summary. The paper, as submitted, claims to address mesh and size sensitivity in RVE modelling of cracked fibre composites. It proposes (i) a set of modified periodic boundary conditions (MPBCs) intended to attenuate RVE-size effects on material-response precision, especially in the inelastic regime, and (ii) a mesh-sensitivity technique that equalises computational fracture energy with a real target, thereby making the response independent of the strain-localisation bandwidth. The claims are supported, according to the abstract, by 1,200 RVE samples spanning five fibre volume fractions and four RVE sizes under transverse loading, plus a qualitative observation about fibre-pair arrangements that promote damage. The body of the manuscript supplied for review, however, is arXiv:2508.08328, a robotics paper on whole-body dynamic grasping with legged manipulators. Consequently, none of the derivations, numerical protocols, convergence metrics, or statistical results asserted in the abstract can be examined.

Significance. If the size-attenuation and mesh-insensitivity claims were substantiated, they would be practically valuable for multiscale modelling of composites: RVE simulations that are both size-insensitive and mesh-insensitive would improve the reliability of transverse-load failure predictions. The scale of the numerical campaign (1,200 samples) is a strength if properly analysed. However, as submitted, the paper provides no quantitative evidence and no auditable derivation. The mesh-insensitivity claim also carries a latent circularity risk: if the strain-localisation bandwidth is calibrated so that the computed fracture energy matches a chosen target, then the reported independence from that bandwidth is partially enforced by construction rather than established as a model property. The current submission does not allow this risk to be resolved.

major comments (3)
  1. [Full text (supplied for review)] The body of the manuscript is a different paper: 'Whole-Body Coordination for Dynamic Object Grasping with Legged Manipulators' (arXiv:2508.08328), not the RVE/fracture paper announced in the abstract. This is not a minor formatting issue. It removes any possibility of auditing the MPBC formulation, the fracture-energy equalisation protocol, the damage-initiation criterion, and the 1,200-sample numerical matrix. The central claims of the abstract are therefore unsupported in the submitted artifact, and no scientific judgment on their correctness can be made.
  2. [Abstract, mesh-sensitivity claim] The abstract states only that the technique 'equalises the fracture energy observed from computational analysis with the real phenomenon'. It does not state how the target fracture energy is obtained or how the strain-localisation bandwidth is chosen. If the bandwidth is calibrated to the same simulated responses later used to demonstrate mesh independence, the independence is a consequence of the calibration. Moreover, matching a single scalar does not constrain peak stress, softening slope, or damage pattern, which are the quantities usually mesh-sensitive. A derivation and a parameter-free or independently determined calibration protocol are required; neither is present.
  3. [Abstract, size-attenuation claim] The MPBCs are introduced through the 'hypothesis' that strain periodicity supplements displacement periodicity, but no equations, boundary conditions, or enforcement method are given. The claim that the MPBCs attenuate RVE-size effects over 1,200 samples is accompanied by no numerical values, confidence intervals, or statistical summary. Without the formulation and the underlying convergence study, the size-attenuation claim cannot be checked.
minor comments (3)
  1. [Abstract, final sentence] The sentence 'the angle between the direction of loading and an imaginary line drawn between their centres is less' is incomplete; the threshold value or inequality is missing. As written, the damage-arrangement observation is unfalsifiable.
  2. [Abstract / reporting] The abstract reports the scale of the campaign (1,200 samples, five volume fractions, four RVE sizes) but no quantitative outcome. The manuscript should include concrete numbers, error bars, and a clear description of the sample generation, material properties, and element discretisation.
  3. [Submission integrity] The supplied full text does not correspond to the abstract's topic or authors. If this is a submission error, the correct manuscript must be uploaded; the current file cannot serve as the basis for review.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: abstract shows no self-referential reduction; full text is a different paper

full rationale

The supplied full text is arXiv:2508.08328 (a legged-manipulation grasping paper), not the target RVE manuscript, so the claimed derivation chain (MPBC formulation, strain-localization bandwidth choice, fracture-energy equalization, and the 1,200-sample campaign) cannot be inspected. On the abstract alone, the central claims are not circular. The mesh-sensitivity technique 'equalises the fracture energy observed from computational analysis with the real phenomenon'; read literally, this anchors the regularization to an external material property rather than to the same simulated response curves later used to demonstrate bandwidth independence. No fitted-input-called-prediction reduction is exhibited. The MPBCs are introduced explicitly as a hypothesis and then tested on 1,200 samples spanning five fibre volume fractions and four RVE sizes; that is an independent empirical check, not a restatement of the hypothesis. The damage-initiation observation ('the angle ... is less' with no stated threshold) is under-specified and unfalsifiable as written, which is a correctness/rigor concern, not circularity. No self-citations, imported uniqueness theorems, or ansatz-smuggling citations appear in the abstract. Therefore no circular step can be quoted or exhibited, and the appropriate finding is no significant circularity.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

Provisional, abstract-only audit. The two free parameters are the equalisation bandwidth (calibration status unstated) and the unstated angle threshold in the fibre-pair damage criterion. The axioms are standard homogenisation and damage-mechanics scaffolding: displacement periodicity, statistical representativeness of the RVE realisations, and a smeared-crack fracture-energy representation. No new physical entities are introduced; MPBCs are a boundary-condition modification, not an entity.

free parameters (2)
  • Strain-localisation bandwidth (regularisation length) for fracture-energy equalisation = not stated in abstract
    The mesh-sensitivity technique equalises computational and real fracture energy through a choice of localisation-band width; the abstract does not say whether this length is derived from theory or calibrated to the response being predicted.
  • Fibre-pair alignment angle threshold for damage initiation = not stated
    The abstract's damage-initiation observation requires the angle between loading direction and the fibre-centre line to be below some threshold ('is less'), but the threshold is never given, leaving the criterion under-specified.
assumptions (3)
  • domain assumption The displacement field is periodic across the RVE boundary (standard PBC assumption carried over into MPBCs).
    Conventional PBCs are the stated starting point of the modification; MPBCs are framed as displacement periodicity plus strain periodicity, so displacement periodicity is assumed from prior RVE practice.
  • domain assumption RVE realisations at five fibre volume fractions and four sizes are statistically representative of the composite microstructure.
    The 1,200-sample verification treats the generated random arrangements as capturing the material response; representativeness of these realisations is assumed rather than demonstrated in the abstract.
  • domain assumption Fracture energy is the correct quantity to equalise between the computational smeared crack and the real phenomenon.
    The mesh-sensitivity fix presupposes a smeared-crack representation in which fracture energy per area maps to dissipation through the localisation band; this is standard crack-band theory, but the abstract does not name it.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Representative Volume Element: Existence and Extent in Cracked Heterogeneous Medium." pith.science (2026). https://pith.science/paper/SJQQK6AO

@misc{pith2026250808320,
  author       = {Pith},
  title        = {Pith review of: Representative Volume Element: Existence and Extent in Cracked Heterogeneous Medium},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SJQQK6AO}},
  note         = {Machine review of arXiv:2508.08320}
}
read the original abstract

Acknowledging the ever-increasing demand for composites in the engineering industry, this paper focuses on the failure of composites at the microscale and augmenting the use of multiscale modelling techniques to make them better for various applications. This work aims to increase the representativeness of the volume element by attenuating the mesh and size sensitivities in representative volume element (RVE) modelling. A technique to alleviate mesh sensitivity in RVE modelling is proposed, which equalises the fracture energy observed from computational analysis with the real phenomenon, thereby keeping the response independent of the bandwidth of strain localisation. Based on the hypothesis that ensuring periodicity of strain, in addition to displacement periodicity across the domain boundary and supplementing the capability of periodic boundary conditions (PBCs) to attenuate the size dependency in RVE modelling, a set of modified PBCs (MPBCs) are formulated. One thousand two hundred RVE samples falling into combinations of five fibre volume fractions and four RVE sizes are analysed under transverse loading, and the ability of MPBCs to attenuate the effect of RVE size on the precision of material response, particularly in the inelastic regime, is verified. This work also focuses on various factors affecting damage initiation in 2D composite RVEs. The arrangement of a pair of fibres with their members placed close to each other, such that the angle between the direction of loading and an imaginary line drawn between their centres is less, is observed to make the region between them more favourable to damage.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

55 extracted references · 45 canonical work pages

  1. [1]

    Akinola, I.; Xu, J.; Song, S.; and Allen, P. K. 2021. Dynamic grasping with reachability and motion aware- ness. In2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 9422–9429

    work page 2021

  2. [2]

    Bharadhwaj, H.; Mottaghi, R.; Gupta, A.; and Tulsiani, S. 2024. Track2act: Predicting point tracks from inter- net videos enables generalizable robot manipulation. In European Conference on Computer Vision, 306–324

    work page 2024

  3. [3]

    Burgess-Limerick, B.; Lehnert, C.; Leitner, J.; and Corke, P. 2022. DGBench: An Open-Source, Re- producible Benchmark for Dynamic Grasping. In IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS 2022, Kyoto, Japan, October 23-27, 2022, 3218–3224

    work page 2022

  4. [4]

    Cai, J.; Su, J.; Zhou, Z.; Cheng, H.; Chen, Q.; and Wang, M. Y . 2022. V olumetric-based Contact Point Detection for 7-DoF Grasping. InConference on Robot Learning (CoRL)

    work page 2022

  5. [5]

    Calli, B.; Walsman, A.; Singh, A.; Srinivasa, S.; Abbeel, P.; and Dollar, A. M. 2015. Benchmarking in Manipu- lation Research: Using the Yale-CMU-Berkeley Object and Model Set.IEEE Robotics and Automation Maga- zine, 22(3): 36–52

    work page 2015

  6. [6]

    Chen, Y .; Lin, Y .; Xu, R.; and Vela, P. A. 2023. Keypoint-graspnet: Keypoint-based 6-dof grasp gen- eration from the monocular rgb-d input. In2023 IEEE International Conference on Robotics and Automation (ICRA), 7988–7995. IEEE

    work page 2023

  7. [7]

    Cheng, Y .; Li, L.; Xu, Y .; Li, X.; Yang, Z.; Wang, W.; and Yang, Y . 2023. Segment and track anything.arXiv preprint arXiv:2305.06558

    arXiv 2023

  8. [8]

    Dai, Q.; Zhang, J.; Li, Q.; Wu, T.; Dong, H.; Liu, Z.; Tan, P.; and Wang, H. 2022. Domain Randomization- Enhanced Depth Simulation and Restoration for Per- ceiving and Grasping Specular and Transparent Objects. InEuropean Conference on Computer Vision (ECCV)

    work page 2022

Show all 55 references
  1. [9]

    Fang, H.; Wang, C.; Gou, M.; and Lu, C. 2020. GraspNet-1Billion: A Large-Scale Benchmark for Gen- eral Object Grasping. In2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2020, Seattle, WA, USA, June 13-19, 2020, 11441–11450

    2020

  2. [10]

    Fang, H.-S.; Wang, C.; Fang, H.; Gou, M.; Liu, J.; Yan, H.; Liu, W.; Xie, Y .; and Lu, C. 2023. Anygrasp: Robust and efficient grasp perception in spatial and temporal domains.IEEE Transactions on Robotics, 39(5): 3929– 3945

    2023

  3. [11]

    Fu, Z.; Cheng, X.; and Pathak, D. 2022. Deep Whole- Body Control: Learning a Unified Policy for Manipula- tion and Locomotion. InConference on Robot Learn- ing, CoRL 2022, 14-18 December 2022, Auckland, New Zealand, 138–149

    2022

  4. [12]

    Ha, H.; Gao, Y .; Fu, Z.; Tan, J.; and Song, S. 2024. UMI on Legs: Making Manipulation Policies Mobile with Manipulation-Centric Whole-body Controllers. In Proceedings of the 2024 Conference on Robot Learning

    2024

  5. [13]

    Huang, B.; Yu, J.; and Jain, S. 2023. EARL: Eye-on- hand reinforcement learner for dynamic grasping with active pose estimation. In2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2963–2970

    2023

  6. [14]

    Kumar, A.; Fu, Z.; Pathak, D.; and Malik, J. 2021. Rma: Rapid motor adaptation for legged robots.arXiv preprint arXiv:2107.04034

    2021 arXiv

  7. [15]

    R.; Villa-Renteria, I.; Tang, J

    Li, C.; Zhang, R.; Wong, J.; Gokmen, C.; Srivastava, S.; Martín-Martín, R.; Wang, C.; Levine, G.; Lingelbach, M.; Sun, J.; Anvari, M.; Hwang, M.; Sharma, M.; Aydin, A.; Bansal, D.; Hunter, S.; Kim, K.; Lou, A.; Matthews, C. R.; Villa-Renteria, I.; Tang, J. H.; Tang, C.; Xia, F...

    2022

  8. [16]

    Liu, J.; Zhang, R.; Fang, H.-S.; Gou, M.; Fang, H.; Wang, C.; Xu, S.; Yan, H.; and Lu, C. 2023. Target- referenced reactive grasping for dynamic objects. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 8824–8833

    2023

  9. [17]

    Liu, M.; Chen, Z.; Cheng, X.; Ji, Y .; Qiu, R.; Yang, R.; and Wang, X. 2024. Visual Whole-Body Control for Legged Loco-Manipulation. InConference on Robot Learning, 6-9 November 2024, Munich, Germany, 234– 257

    2024

  10. [18]

    Long, J.; Wang, Z.; Li, Q.; Cao, L.; Gao, J.; and Pang, J

  11. [19]

    Ma, H.; and Huang, D. 2023. Towards scale balanced 6-dof grasp detection in cluttered scenes. InConference on robot learning, 2004–2013

    2023

  12. [20]

    Ma, Y .; Farshidian, F.; Miki, T.; Lee, J.; and Hutter, M. 2022. Combining learning-based locomotion pol- icy with model-based manipulation for legged mobile manipulators.IEEE Robotics and Automation Letters, 7(2): 2377–2384

    2022

  13. [21]

    Makoviychuk, V .; Wawrzyniak, L.; Guo, Y .; Lu, M.; Storey, K.; Macklin, M.; Hoeller, D.; Rudin, N.; All- shire, A.; Handa, A.; et al. 2021. Isaac gym: High performance gpu-based physics simulation for robot learning.arXiv preprint arXiv:2108.10470

    2021 arXiv

  14. [22]

    Mei, Y .; Wang, Y .; Zheng, S.; and Jin, Q. 2024. Quadrupedgpt: Towards a versatile quadruped agent in open-ended worlds.arXiv preprint arXiv:2406.16578

    2024 arXiv

  15. [23]

    Menon, A.; Cohen, B.; and Likhachev, M. 2014. Mo- tion planning for smooth pickup of moving objects. In 2014 IEEE International Conference on Robotics and Automation (ICRA), 453–460

    2014

  16. [24]

    Mittal, M.; Hoeller, D.; Farshidian, F.; Hutter, M.; and Garg, A. 2022. Articulated object interaction in un- known scenes with whole-body mobile manipulation. In2022 IEEE/RSJ international conference on intelli- gent robots and systems (IROS), 1647–1654. IEEE

    2022

  17. [25]

    Mu, T.; Ling, Z.; Xiang, F.; Yang, D.; Li, X.; Tao, S.; Huang, Z.; Jia, Z.; and Su, H. 2021. ManiSkill: General- izable Manipulation Skill Benchmark with Large-Scale Demonstrations. InProceedings of the Neural Informa- tion Processing Systems Track on Datasets and Bench- marks...

    2021

  18. [26]

    Pan, G.; Ben, Q.; Yuan, Z.; Jiang, G.; Ji, Y .; Li, S.; Pang, J.; Liu, H.; and Xu, H. 2024. RoboDuet: Whole-body Legged Loco-Manipulation with Cross-Embodiment Deployment.arXiv preprint arXiv:2403.17367

    2024 arXiv

  19. [27]

    Portela, T.; Cramariuc, A.; Mittal, M.; and Hutter, M

  20. [28]

    B.; Ji, Y .; and Agrawal, P

    Portela, T.; Margolis, G. B.; Ji, Y .; and Agrawal, P

  21. [29]

    Whole-body end-effector pose tracking.arXiv preprint arXiv:2409.16048

    arXiv

  22. [30]

    Qin, R.; Ma, H.; Gao, B.; and Huang, D. 2023. RGB-D grasp detection via depth guided learning with cross- modal attention. In2023 IEEE International Confer- ence on Robotics and Automation (ICRA), 8003–8009

    2023

  23. [31]

    In2024 IEEE International Conference on Robotics and Automation (ICRA), 15366–15372

    Learning force control for legged manipulation. In2024 IEEE International Conference on Robotics and Automation (ICRA), 15366–15372

  24. [32]

    R.; Su, H.; Mo, K.; and Guibas, L

    Qi, C. R.; Su, H.; Mo, K.; and Guibas, L. J. 2017. Point- net: Deep learning on point sets for 3d classification and segmentation. InProceedings of the IEEE conference on computer vision and pattern recognition, 652–660

    2017

  25. [33]

    J.; and Bagnell, D

    Ross, S.; Gordon, G. J.; and Bagnell, D. 2011. A Re- duction of Imitation Learning and Structured Prediction to No-Regret Online Learning. InProceedings of the Fourteenth International Conference on Artificial Intel- ligence and Statistics, AISTATS 2011, Fort Lauderdale, USA, ...

    2011

  26. [34]

    Qiu, R.-Z.; Hu, Y .; Song, Y .; Yang, G.; Fu, Y .; Ye, J.; Mu, J.; Yang, R.; Atanasov, N.; Scherer, S.; et al. 2024. Learning generalizable feature fields for mobile manip- ulation.arXiv preprint arXiv:2403.07563

    2024 arXiv

  27. [35]

    A.; Yang, G.; Liu, M.; Ji, M.; Jia, C.; Yang, R.; Zou, X.; et al

    Qiu, R.-Z.; Song, Y .; Peng, X.; Suryadevara, S. A.; Yang, G.; Liu, M.; Ji, M.; Jia, C.; Yang, R.; Zou, X.; et al

  28. [36]

    WildLMa: Long Horizon Loco-Manipulation in the Wild.arXiv preprint arXiv:2411.15131

    arXiv

  29. [37]

    Sundermeyer, M.; Mousavian, A.; Triebel, R.; and Fox, D. 2021. Contact-graspnet: Efficient 6-dof grasp gener- ation in cluttered scenes. In2021 IEEE International Conference on Robotics and Automation (ICRA), 13438– 13444

    2021

  30. [38]

    Schulman, J.; Wolski, F.; Dhariwal, P.; Radford, A.; and Klimov, O. 2017. Proximal Policy Optimization Algorithms. arXiv:1707.06347

    2017 arXiv

  31. [39]

    V .; and Hutter, M

    Sleiman, J.-P.; Farshidian, F.; Minniti, M. V .; and Hutter, M. 2021. A unified mpc framework for whole-body dynamic locomotion and manipulation.IEEE Robotics and Automation Letters, 6(3): 4688–4695

    2021

  32. [40]

    Sun, J.; Zhou, L.; Geng, B.; Zhang, Y .; and Li, Y

  33. [41]

    IEEE/ASME transactions on mechatronics

    Leg state estimation for quadruped robot by using probabilistic model with proprioceptive feedback. IEEE/ASME transactions on mechatronics

  34. [42]

    Xie, P.; Chen, S.; Chen, Q.; Tang, W.; Hu, D.; Dai, Y .; Chen, R.; and Wang, G. 2024. GAP-RL: Grasps as Points for RL Towards Dynamic Object Grasping. IEEE Robotics and Automation Letters

    2024

  35. [43]

    Wang, J.; Rajabov, J.; Xu, C.; Zheng, Y .; and Wang, H

  36. [44]

    Yang, R.; Zhang, M.; Hansen, N.; Xu, H.; and Wang, X. 2022. Learning Vision-Guided Quadrupedal Loco- motion End-to-End with Cross-Modal Transformers. In International Conference on Learning Representations

    2022

  37. [45]

    Wen, H.; Yan, J.; Peng, W.; and Sun, Y . 2022. Trans- Grasp: Grasp Pose Estimation of a Category of Ob- jects by Transferring Grasps from Only One Labeled Instance. InProceedings of the European Conference on Computer Vision (ECCV)

    2022

  38. [46]

    Wu, T.; Zhong, F.; Geng, Y .; Wang, H.; Zhu, Y .; Wang, Y .; and Dong, H. 2022. Grasparl: Dynamic grasping via adversarial reinforcement learning.arXiv preprint arXiv:2203.02119

    2022 arXiv

  39. [47]

    Xia, Y .; Ding, R.; Qin, Z.; Zhan, G.; Zhou, K.; Yang, L.; Dong, H.; and Cremers, D. 2024. TARGO: Benchmark- ing Target-driven Object Grasping under Occlusions. arXiv:2407.06168

    2024 arXiv

  40. [48]

    Zhang, J.; Gireesh, N.; Wang, J.; Fang, X.; Xu, C.; Chen, W.; Dai, L.; and Wang, H. 2024. GAMMA: Graspability-Aware Mobile MAnipulation Policy Learn- ing based on Online Grasping Pose Fusion. InIEEE International Conference on Robotics and Automation, ICRA 2024, Yokohama, Japa...

    2024

  41. [49]

    Xie, P.; Chen, S.; Chen, Q.; Tang, W.; Hu, D.; Dai, Y .; Chen, R.; and Wang, G. 2025. GAP-RL: Grasps as Points for RL Towards Dynamic Object Grasping. IEEE Robotics Autom. Lett., 10(1): 40–47

    2025

  42. [51]

    Yokoyama, N.; Clegg, A.; Truong, J.; Undersander, E.; Yang, T.-Y .; Arnaud, S.; Ha, S.; Batra, D.; and Rai, A. 2023. Asc: Adaptive skill coordination for robotic mobile manipulation.IEEE Robotics and Automation Letters, 9(1): 779–786

    2023

  43. [52]

    Youm, D.; Jung, H.; Kim, H.; Hwangbo, J.; Park, H.- W.; and Ha, S. 2023. Imitating and finetuning model predictive control for robust and symmetric quadrupedal locomotion.IEEE Robotics and Automation Letters, 8(11): 7799–7806

    2023

  44. [53]

    Zeng, R.; Zhou, D.; Liang, Q.; Liu, J.; Li, H.; Huang, C.; Li, J.; Hu, X.; and Sun, F. 2024. Video2Reward: Gener- ating Reward Function from Videos for Legged Robot Behavior Learning.arXiv preprint arXiv:2412.05515

    2024 arXiv

  45. [55]

    cheating

    Zhang, T.; Guo, S.; Xiong, X.; Li, W.; Qi, Z.; and Lou, Y . 2023. Dynamic object tracking for quadruped manipulator with spherical image-based approach. In 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 727–734. IEEE. Additional Benchmark Deta...

    2023

  46. [2024]

    InThe Twelfth International Conference on Learning Represen- tations, ICLR 2024, Vienna, Austria, May 7-11, 2024

    Hybrid Internal Model: Learning Agile Legged Locomotion with Simulated Robot Response. InThe Twelfth International Conference on Learning Represen- tations, ICLR 2024, Vienna, Austria, May 7-11, 2024

    2024

  47. [2025]

    arXiv:2411.06782

    QuadWBG: Generalizable Quadrupedal Whole- Body Grasping. arXiv:2411.06782

    arXiv

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.