pith. sign in

arxiv: 2607.00673 · v1 · pith:XAD37DCHnew · submitted 2026-07-01 · 💻 cs.RO

Path Planning in Physically Viable World Models

Su Ann Low , Cheng-Hsi Hsiao , Xingjian Li , Adam J. Thorpe , Ufuk Topcu , Krishna Kumar This is my paper

Pith reviewed 2026-07-02 11:30 UTC · model grok-4.3

classification 💻 cs.RO
keywords path planningphysically viable world modelsterrain changesrobot navigation3D Gaussian splatsphysics simulationoutdoor environmentsflooding
0
0 comments X

The pith

Physically viable world models reveal long-horizon route failures in robot navigation that static maps miss.

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

The paper develops a world model that combines 3D scene reconstructions with physics simulations to predict how terrain changes like flooding will affect navigable routes. Robots can then use a terrain-aware planner to test planned paths against these modified environments. This matters because outdoor robots often operate on stale maps that do not account for physical alterations, potentially leading to unsafe or unreachable routes during execution. The evaluation on a real field site shows that these models identify failures and necessary reroutes not visible in the original reconstruction.

Core claim

The system augments reconstructed 3D Gaussian splat scenes with physics-based simulation to generate physically modified versions of the environment. A terrain-aware planner then accounts for the simulated events, obstacles, and deformations. This enables evaluation of whether planned routes remain feasible under future terrain changes, exposing long-horizon route failures and rerouting behavior not apparent when planning only on the original reconstructed environment.

What carries the argument

The physically viable world model that augments 3D reconstructions with physics-based simulation of terrain changes such as flooding.

If this is right

  • Planned routes may fail under simulated terrain changes even if feasible on the original map.
  • The planner can identify rerouting needs before deployment in changing environments.
  • Evaluation on multiple flooding severity levels measures changes in route and mission feasibility.
  • Robots can avoid committing to routes that become unsafe after physical events.

Where Pith is reading between the lines

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

  • Similar simulations could apply to other dynamic factors like seasonal vegetation or erosion.
  • Integration with online updates might allow continuous replanning as real changes occur.
  • This method reduces the need for frequent remapping of large sites.

Load-bearing premise

The physics simulation of terrain changes produces modified scenes realistic enough to match actual outcomes at the field site.

What would settle it

Measuring actual route feasibility after real flooding at the Texas site and comparing it to the simulated predictions.

Figures

Figures reproduced from arXiv: 2607.00673 by Adam J. Thorpe, Cheng-Hsi Hsiao, Krishna Kumar, Su Ann Low, Ufuk Topcu, Xingjian Li.

Figure 1
Figure 1. Figure 1: A physically viable world model generates scenes given different queries, and a [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Path planning results under different flood levels. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Simulating traversability of an Alaskan Village before and after flood. Top: the original [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Path planning in the sandbox. The red dashed circle highlights the landslide region. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Digital elevation map (DEM) of the environment. [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Terrain height maps under increasing flood levels. The first row corresponds to dry and [PITH_FULL_IMAGE:figures/full_fig_p014_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Illustration of the MPM update scheme. (1) Material points carry physical state (mass, [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗
read the original abstract

Robots deployed in unstructured outdoor environments often plan from scene reconstructions collected before deployment because operators cannot remap large or remote sites before every mission. As a result, robots must make long-horizon planning decisions using stale maps that assume the terrain remains unchanged, even though physical changes to the environment may render previously feasible routes unsafe or unreachable at execution time. We present a physically viable world model for evaluating what-if queries for robot navigation under future terrain change. The system augments reconstructed 3D Gaussian splat scenes with physics-based simulation to generate physically modified versions of the same environment without recollecting sensor data or rebuilding the map. We then implement a terrain-aware planner that accounts for physical events, obstacles, and deformations that are simulated by the world model. This allows robots and human operators to evaluate whether planned routes remain feasible before committing to a planned route, particularly in constrained environments where retreat or recovery may become impossible once conditions change. We evaluate the system on a real outdoor field site in Central Texas using simulated flooding across multiple severity levels. We measure route and mission feasibility as terrain conditions deteriorate under physically simulated interventions. Our results show that physically viable world models expose long-horizon route failures and rerouting behavior that are not apparent when planning only on the original reconstructed environment, allowing robots to evaluate how future terrain changes may affect route feasibility before deployment.

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.

Referee Report

2 major / 2 minor

Summary. The paper proposes augmenting 3D Gaussian splat reconstructions of outdoor environments with physics-based simulation to create 'physically viable world models' that support what-if queries about terrain changes (e.g., flooding, deformations). A terrain-aware planner then evaluates route and mission feasibility under these simulated changes. The central claim is that this approach reveals long-horizon route failures and necessary reroutes that are invisible when planning solely on the original pre-deployment reconstruction. Evaluation consists of simulated flooding at multiple severity levels on a real Central Texas field site, with feasibility measured under the modified scenes.

Significance. If the physics simulation produces scenes whose traversability matches real-world terrain evolution, the method would offer a practical way to assess deployment risks in dynamic outdoor settings without repeated remapping. The use of existing reconstructions and external physics engines is a pragmatic strength. However, the absence of any reported quantitative validation of the simulated changes against actual physical outcomes at the site substantially reduces the significance of the reported feasibility results.

major comments (2)
  1. [Evaluation] Evaluation section: the central claim that physically viable world models 'expose long-horizon route failures and rerouting behavior that are not apparent when planning only on the original reconstructed environment' rests on the assumption that the physics-based flooding and deformation outputs produce modified scenes whose feasibility matches real terrain evolution. No quantitative comparison (water-depth maps, post-event DEMs, traversability measurements, or error metrics) to actual physical interventions at the Central Texas site is reported, leaving open the possibility that observed failures are simulation artifacts.
  2. [Evaluation] Evaluation section: no quantitative metrics, success rates, error analysis, or details on how route/mission feasibility was measured (e.g., binary feasibility, cost thresholds, or simulation parameters) are provided despite the abstract stating that 'we measure route and mission feasibility as terrain conditions deteriorate.' This absence makes it impossible to assess the magnitude or statistical reliability of the claimed difference between original and modified scenes.
minor comments (2)
  1. [Abstract] The abstract would be strengthened by including at least one concrete quantitative result (e.g., percentage of routes that become infeasible) rather than a purely qualitative statement of findings.
  2. [Method] Notation for the terrain-aware planner and the interface between the Gaussian splat and the physics simulator should be introduced with explicit equations or pseudocode in the method section to improve reproducibility.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for their thoughtful review and for highlighting important aspects of the evaluation. We respond to each major comment below and note where revisions or clarifications can be made.

read point-by-point responses

  1. Referee: [Evaluation] Evaluation section: the central claim that physically viable world models 'expose long-horizon route failures and rerouting behavior that are not apparent when planning only on the original reconstructed environment' rests on the assumption that the physics-based flooding and deformation outputs produce modified scenes whose feasibility matches real terrain evolution. No quantitative comparison (water-depth maps, post-event DEMs, traversability measurements, or error metrics) to actual physical interventions at the Central Texas site is reported, leaving open the possibility that observed failures are simulation artifacts.

    Authors: We acknowledge that the evaluation uses simulated flooding without direct quantitative validation against real post-event measurements at the site. The work focuses on demonstrating the planning consequences of what-if terrain modifications generated from an existing reconstruction and a standard physics engine, rather than on simulator validation. Because no real flooding or deformation data were collected at the Central Texas site, we cannot provide such comparisons. We will add an explicit limitations paragraph discussing the reliance on simulation fidelity and the distinction between demonstrating planning sensitivity and claiming predictive accuracy for real events. revision: partial

  2. Referee: [Evaluation] Evaluation section: no quantitative metrics, success rates, error analysis, or details on how route/mission feasibility was measured (e.g., binary feasibility, cost thresholds, or simulation parameters) are provided despite the abstract stating that 'we measure route and mission feasibility as terrain conditions deteriorate.' This absence makes it impossible to assess the magnitude or statistical reliability of the claimed difference between original and modified scenes.

    Authors: Feasibility is assessed by executing the terrain-aware planner on each modified scene and recording whether a collision-free path satisfying the mission constraints exists. We will expand the evaluation section to include the precise feasibility criterion (binary existence of a feasible path), the cost threshold used for replanning, the physics-simulation parameters (e.g., water level increments, material properties), and any aggregate statistics across the severity levels shown in the figures. This will allow readers to better gauge the magnitude of the observed differences. revision: yes

standing simulated objections not resolved
  • Absence of quantitative validation of simulated flooding against actual physical outcomes or post-event data at the Central Texas site, which was not collected during the study.

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained

full rationale

The paper describes a system that augments pre-deployment 3D Gaussian reconstructions with external physics-based simulation to generate modified scenes, then applies a terrain-aware planner and compares feasibility metrics on original vs. simulated scenes. No equations, parameters, or claims reduce by construction to fitted inputs or self-citations; the evaluation relies on simulated interventions whose outputs are treated as independent inputs to the planner. The central empirical observation (long-horizon failures visible only under simulated change) follows directly from running the planner on two distinct scene versions without tautological redefinition. This is the most common honest non-finding for system-description papers that do not derive predictions from their own fitted quantities.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that physics simulation can generate realistic terrain modifications from static reconstructions without new sensor data.

axioms (1)
  • domain assumption Physics simulation accurately models real terrain changes such as flooding and deformations for the purpose of route evaluation
    Invoked to justify generating modified environments that affect path feasibility.

pith-pipeline@v0.9.1-grok · 5782 in / 1076 out tokens · 24149 ms · 2026-07-02T11:30:46.168338+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

44 extracted references · 22 canonical work pages · 6 internal anchors

  1. [1]

    A. J. Thorpe, S. Tretiakov, C.-H. Hsiao, S. A. Low, X. Li, H. Iqbal, N. P. Bhatt, U. Topcu, and K. Kumar. Physically viable world models: A case for query-conditioned embodied ai, 2026. URLhttps://arxiv.org/abs/2605.30542

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  2. [2]

    M. G. Andreu, V . Klemm, V . Patil, J. Tordesillas, and M. Hutter. Foci: Trajectory optimization on gaussian splats. In2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 1884–1891. IEEE, 2025

    2025

  3. [3]

    J. Frey, M. Patel, D. Atha, J. Nubert, D. Fan, A. Agha, C. Padgett, P. Spieler, M. Hutter, and S. Khattak. Roadrunner—learning traversability estimation for autonomous off-road driving. IEEE Transactions on Field Robotics, 1:192–212, 2024. doi:10.1109/TFR.2024.3464369

    work page doi:10.1109/tfr.2024.3464369 2024

  4. [4]

    E. Chen, C. Ho, M. Maulimov, C. Wang, and S. Scherer. Learning-on-the-drive: Self- supervised adaptation of visual offroad traversability models.2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2024. URLhttps://arxiv.org/ pdf/2306.15226

    work page arXiv 2024

  5. [5]

    M. G. Castro, S. Triest, W. Wang, J. M. Gregory, F. Sanchez, J. G. Rogers, and S. Scherer. How does it feel? self-supervised costmap learning for off-road vehicle traversability. In2023 IEEE International Conference on Robotics and Automation (ICRA), pages 931–938, 2023. doi:10.1109/ICRA48891.2023.10160856

    work page doi:10.1109/icra48891.2023.10160856 2023

  6. [6]

    X. Meng, N. Hatch, A. Lambert, A. Li, N. Wagener, M. Schmittle, J. Lee, W. Yuan, Z. Chen, S. Deng, G. Okopal, D. Fox, B. Boots, and A. Shaban. TerrainNet: Visual Modeling of Com- plex Terrain for High-speed, Off-road Navigation. InProceedings of Robotics: Science and Systems, Daegu, Republic of Korea, July 2023. doi:10.15607/RSS.2023.XIX.103

    work page doi:10.15607/rss.2023.xix.103 2023

  7. [7]

    Pentzer, S

    J. Pentzer, S. Brennan, and K. Reichard. On-line estimation of vehicle motion and power model parameters for skid-steer robot energy use prediction. In2014 American Control Conference, pages 2786–2791, 2014. doi:10.1109/ACC.2014.6859073

    work page doi:10.1109/acc.2014.6859073 2014

  8. [8]

    Kumar, Z

    A. Kumar, Z. Fu, D. Pathak, and J. Malik. RMA: Rapid motor adaptation for legged robots. In Robotics: Science and Systems, 2021

    2021

  9. [9]

    W. Ward, S. Etter, T. Ingebrand, C. Ellis, A. J. Thorpe, and U. Topcu. Online adaptation of terrain-aware dynamics for planning in unstructured environments, 2025. URLhttps: //arxiv.org/abs/2506.04484

    work page arXiv 2025

  10. [10]

    W. Ward, S. Etter, J. Quattrociocchi, C. Ellis, A. J. Thorpe, and U. Topcu. Zero to autonomy in real-time: Online adaptation of dynamics in unstructured environments, 2026. URLhttps: //arxiv.org/abs/2509.12516

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  11. [11]

    Williams, A

    G. Williams, A. Aldrich, and E. A. Theodorou. Model predictive path integral control: From theory to parallel computation.Journal of Guidance, Control, and Dynamics, 40(2):344–357,

  12. [12]

    URLhttps://doi.org/10.2514/1.G001921

    doi:10.2514/1.G001921. URLhttps://doi.org/10.2514/1.G001921

    work page doi:10.2514/1.g001921

  13. [13]

    Williams, P

    G. Williams, P. Drews, B. Goldfain, J. M. Rehg, and E. A. Theodorou. Information-theoretic model predictive control: Theory and applications to autonomous driving.Trans. Rob., 34(6): 1603–1622, Dec. 2018. ISSN 1552-3098. doi:10.1109/TRO.2018.2865891. URLhttps: //doi.org/10.1109/TRO.2018.2865891

    work page doi:10.1109/tro.2018.2865891 2018

  14. [14]

    C. Finn, P. Abbeel, and S. Levine. Model-agnostic meta-learning for fast adaptation of deep networks. In D. Precup and Y . W. Teh, editors,Proceedings of the 34th International Con- ference on Machine Learning, volume 70 ofProceedings of Machine Learning Research, pages 1126–1135. PMLR, 06–11 Aug 2017. URLhttps://proceedings.mlr.press/ v70/finn17a.html. 9

    2017

  15. [15]

    Clavera, A

    I. Clavera, A. Nagabandi, S. Liu, R. S. Fearing, P. Abbeel, S. Levine, and C. Finn. Learning to adapt in dynamic, real-world environments through meta-reinforcement learning. InInter- national Conference on Learning Representations, 2019. URLhttps://openreview.net/ forum?id=HyztsoC5Y7

    2019

  16. [16]

    Pantic, C

    M. Pantic, C. Cadena, R. Siegwart, and L. Ott. Sampling-free obstacle gradients and reactive planning in neural radiance fields (nerf).arXiv preprint arXiv:2205.01389, 2022

    work page arXiv 2022

  17. [17]

    Adamkiewicz, T

    M. Adamkiewicz, T. Chen, A. Caccavale, R. Gardner, P. Culbertson, J. Bohg, and M. Schwa- ger. Vision-only robot navigation in a neural radiance world.IEEE Robotics and Automation Letters, 7(2):4606–4613, 2022

    2022

  18. [18]

    T. Chen, P. Culbertson, and M. Schwager. Catnips: Collision avoidance through neural implicit probabilistic scenes.IEEE Transactions on Robotics, 40:2712–2728, 2024

    2024

  19. [19]

    T. Chen, O. Shorinwa, J. Bruno, A. Swann, J. Yu, W. Zeng, K. Nagami, P. Dames, and M. Schwager. Splat-nav: Safe real-time robot navigation in gaussian splatting maps.IEEE Transactions on Robotics, 2025

    2025

  20. [20]

    World Models

    D. Ha and J. Schmidhuber. World models.arXiv preprint arXiv:1803.10122, 2(3):440, 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  21. [21]

    Hafner, J

    D. Hafner, J. Pasukonis, J. Ba, and T. Lillicrap. Mastering diverse control tasks through world models.Nature, 640(8059):647–653, 2025

    2025

  22. [22]

    Hafner, T

    D. Hafner, T. Lillicrap, I. Fischer, R. Villegas, D. Ha, H. Lee, and J. Davidson. Learning latent dynamics for planning from pixels. InInternational conference on machine learning, pages 2555–2565. PMLR, 2019

    2019

  23. [23]

    V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning

    M. Assran, A. Bardes, D. Fan, Q. Garrido, R. Howes, M. Muckley, A. Rizvi, C. Roberts, K. Sinha, A. Zholus, et al. V-jepa 2: Self-supervised video models enable understanding, prediction and planning.arXiv preprint arXiv:2506.09985, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  24. [24]

    Bruce, M

    J. Bruce, M. Dennis, A. Edwards, J. Parker-Holder, Y . Shi, E. Hughes, M. Lai, A. Mavalankar, R. Steigerwald, C. Apps, Y . Aytar, S. Bechtle, F. Behbahani, S. Chan, N. Heess, L. Gonzalez, S. Osindero, S. Ozair, S. Reed, J. Zhang, K. Zolna, J. Clune, N. de Freitas, S. Singh, and T. Rockt¨aschel. Genie: Generative interactive environments.arXiv preprint arX...

    work page arXiv 2024

  25. [25]

    PaLM-E: An Embodied Multimodal Language Model

    D. Driess, F. Xia, M. S. M. Sajjadi, C. Lynch, A. Chowdhery, B. Ichter, A. Wahid, J. Tompson, Q. Vuong, T. Yu, W. Huang, Y . Chebotar, P. Sermanet, D. Duckworth, S. Levine, V . Vanhoucke, K. Hausman, M. Toussaint, K. Greff, A. Zeng, I. Mordatch, and P. Florence. Palm-e: An embodied multimodal language model. InarXiv preprint arXiv:2303.03378, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  26. [26]

    Brohan, N

    A. Brohan, N. Brown, J. Carbajal, Y . Chebotar, J. Dabis, C. Finn, K. Gopalakrishnan, K. Haus- man, A. Herzog, J. Hsu, J. Ibarz, B. Ichter, A. Irpan, T. Jackson, S. Jesmonth, N. Joshi, R. Ju- lian, D. Kalashnikov, Y . Kuang, I. Leal, K.-H. Lee, S. Levine, Y . Lu, U. Malla, D. Manjunath, I. Mordatch, O. Nachum, C. Parada, J. Peralta, E. Perez, K. Pertsch, ...

    work page doi:10.15607/rss.2023.xix.025 2023

  27. [27]

    $\pi_0$: A Vision-Language-Action Flow Model for General Robot Control

    K. Black, N. Brown, D. Driess, A. Esmail, M. Equi, C. Finn, N. Fusai, L. Groom, K. Haus- man, B. Ichter, S. Jakubczak, T. Jones, L. Ke, S. Levine, A. Li-Bell, M. Mothukuri, S. Nair, K. Pertsch, L. X. Shi, J. Tanner, Q. Vuong, A. Walling, H. Wang, and U. Zhilinsky.π 0: A vision-language-action flow model for general robot control, 2026. URLhttps://arxiv. o...

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  28. [28]

    Language models as zero-shot planners: Extracting actionable knowledge for embodied agents,

    W. Huang, P. Abbeel, D. Pathak, and I. Mordatch. Language models as zero-shot planners: Extracting actionable knowledge for embodied agents, 2022. URLhttps://arxiv.org/ abs/2201.07207

    work page arXiv 2022

  29. [29]

    Byravan, F

    A. Byravan, F. Leeb, F. Meier, and D. Fox. Se3-pose-nets: Structured deep dynamics models for visuomotor control. In2018 IEEE International Conference on Robotics and Automation (ICRA), pages 3339–3346, 2018. doi:10.1109/ICRA.2018.8461184

    work page doi:10.1109/icra.2018.8461184 2018

  30. [30]

    Battaglia, R

    P. Battaglia, R. Pascanu, M. Lai, D. Jimenez Rezende, and k. kavukcuoglu. Interaction net- works for learning about objects, relations and physics. In D. Lee, M. Sugiyama, U. Luxburg, I. Guyon, and R. Garnett, editors,Advances in Neural Information Processing Systems, vol- ume 29. Curran Associates, Inc., 2016. URLhttps://proceedings.neurips.cc/paper_ fil...

    2016

  31. [31]

    Battaglia, R

    P. Battaglia, R. Pascanu, M. Lai, D. J. Rezende, and K. kavukcuoglu. Interaction networks for learning about objects, relations and physics. InProceedings of the 30th International Conference on Neural Information Processing Systems, NIPS’16, page 4509–4517, Red Hook, NY , USA, 2016. Curran Associates Inc. ISBN 9781510838819

    2016

  32. [32]

    Sanchez-Gonzalez, J

    A. Sanchez-Gonzalez, J. Godwin, T. Pfaff, R. Ying, J. Leskovec, and P. Battaglia. Learning to simulate complex physics with graph networks. In H. D. III and A. Singh, editors,Proceedings of the 37th International Conference on Machine Learning, volume 119 ofProceedings of Machine Learning Research, pages 8459–8468. PMLR, 13–18 Jul 2020. URLhttps:// procee...

    2020

  33. [33]

    Raissi, P

    M. Raissi, P. Perdikaris, and G. Karniadakis. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial dif- ferential equations.Journal of Computational Physics, 378:686–707, 2019. ISSN 0021-

    2019

  34. [34]

    Physics-Informed Neural Networks: A Deep Learning Framework for Solving Forward and Inverse Problems Involving Nonlinear Partial Differential Equations

    doi:https://doi.org/10.1016/j.jcp.2018.10.045. URLhttps://www.sciencedirect. com/science/article/pii/S0021999118307125

    work page doi:10.1016/j.jcp.2018.10.045 2018

  35. [35]

    Z. Li, N. B. Kovachki, K. Azizzadenesheli, B. liu, K. Bhattacharya, A. Stuart, and A. Anand- kumar. Fourier neural operator for parametric partial differential equations. InInternational Conference on Learning Representations, 2021. URLhttps://openreview.net/forum? id=c8P9NQVtmnO

    2021

  36. [36]

    Kovachki, Z

    N. Kovachki, Z. Li, B. Liu, K. Azizzadenesheli, K. Bhattacharya, A. Stuart, and A. Anand- kumar. Neural operator: learning maps between function spaces with applications to pdes.J. Mach. Learn. Res., 24(1), Jan. 2023. ISSN 1532-4435

    2023

  37. [37]

    T. Xie, Z. Zong, Y . Qiu, X. Li, Y . Feng, Y . Yang, and C. Jiang. Physgaussian: Physics- integrated 3d gaussians for generative dynamics, 2024. URLhttps://arxiv.org/abs/ 2311.12198

    work page arXiv 2024

  38. [38]

    Abou-Chakra, K

    J. Abou-Chakra, K. Rana, F. Dayoub, and N. Suenderhauf. Physically embodied gaussian splatting: A realtime correctable world model for robotics. In8th Annual Conference on Robot Learning, 2024. URLhttps://openreview.net/forum?id=AEq0onGrN2

    2024

  39. [39]

    Borycki, W

    P. Borycki, W. Smolak-Dy ˙zewska, J. Waczy´nska, M. Mazur, S. Tadeja, and P. Spurek. Gasp: Gaussian splatting for physics-based simulations.Computer Vision and Image Understanding, 265:104699, 2026. ISSN 1077-3142. doi:https://doi.org/10.1016/j.cviu.2026.104699. URL https://www.sciencedirect.com/science/article/pii/S1077314226000664

    work page doi:10.1016/j.cviu.2026.104699 2026

  40. [40]

    T. Xie, Z. Zong, Y . Qiu, X. Li, Y . Feng, Y . Yang, and C. Jiang. Physgaussian: Physics- integrated 3d gaussians for generative dynamics. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4389–4398, 2024. 11

    2024

  41. [41]

    W ¨achter and L

    A. W ¨achter and L. T. Biegler. On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming.Mathematical programming, 106(1):25–57, 2006

    2006

  42. [42]

    M. Macklin. Warp: A High-performance Python Framework for GPU Simulation and Graph- ics, Mar. 2022. URLhttps://github.com/NVIDIA/warp. NVIDIA GPU Technology Con- ference (GTC)

    2022

  43. [43]

    Tordesillas and J

    J. Tordesillas and J. P. How. Minvo basis: Finding simplexes with minimum volume enclosing polynomial curves.Computer-Aided Design, 151:103341, 2022

    2022

  44. [44]

    Knapitsch, J

    A. Knapitsch, J. Park, Q.-Y . Zhou, and V . Koltun. Tanks and temples: Benchmarking large- scale scene reconstruction.ACM Transactions on Graphics, 36(4), 2017. 12 Figure 5: Digital elevation map (DEM) of the environment. A Additional Experiments A.1 Central Texas Hill Country We evaluate the system on a real outdoor field site, an approximately500×500ft ...

    2017