pith. sign in

arxiv: 2604.11981 · v1 · submitted 2026-04-13 · 💻 cs.RO

Bipedal-Walking-Dynamics Model on Granular Terrains

Xunjie Chen , Xinyan Huang , Peter Shan , Jingang Yi , Tao Liu This is my paper

Pith reviewed 2026-05-10 15:46 UTC · model grok-4.3

classification 💻 cs.RO
keywords bipedal locomotiongranular terrainsdynamics modelingfoot sinkagefoot slipcost of transportground reaction forcerobot walking
0
0 comments X

The pith

A dynamics model with three extra degrees of freedom predicts bipedal robot walking on granular media by estimating foot sinkage and slip.

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

The paper develops a dynamics model for bipedal robots walking on yielding surfaces such as sand. It adds three degrees of freedom to a foot-terrain interaction model to track how much the feet sink and slip. These effects matter because they change the forces the ground exerts back on the robot and raise the energy needed for walking. The model is tested on a physical biped robot, where it matches observed walking patterns, intrusion depths, and ground forces. If accurate, the approach gives robots a practical way to plan stable and efficient gaits on loose terrain without full particle simulations.

Core claim

We present a new dynamics-modeling approach to capture and predict bipedal-walking locomotion on granular media. A dynamic foot-terrain interaction model is integrated to compute the ground reaction force. The proposed granular dynamic model has three additional degrees of freedom to estimate foot sinkage and slip that are critical to capturing robot-walking kinematics and kinetics such as cost of transport. Experiments using a biped robotic walker on sand validate the model with gait profiles, media-intrusion prediction, and ground reaction force estimations.

What carries the argument

The dynamic foot-terrain interaction model that incorporates three additional degrees of freedom to estimate foot sinkage and slip while computing ground reaction forces during bipedal walking.

If this is right

  • The model allows analysis of bipedal kinetics and cost of transport on granular terrains.
  • It predicts foot-terrain rolling and intrusion effects during walking.
  • The approach supports development of locomotion control and optimization methods for bipedal robots on such surfaces.
  • Validation shows good agreement with robot gait profiles and measured ground reaction forces.

Where Pith is reading between the lines

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

  • If the three-DoF model generalizes well, it could be applied to other granular or soft terrains like gravel or mud with minimal recalibration.
  • Integrating this model into optimization loops might yield lower-energy gait patterns for robots operating in deserts or beaches.
  • Real-time versions of the model could enable adaptive stepping strategies that respond to detected sinkage during locomotion.

Load-bearing premise

The three additional degrees of freedom sufficiently represent the dominant effects of foot sinkage and slip on granular media without needing more complex particle-level simulations or terrain-specific parameter tuning.

What would settle it

Measurements from the biped robot showing that the predicted ground reaction forces deviate significantly from actual sensor readings across multiple gait cycles on sand would indicate the model fails to capture the dynamics accurately.

Figures

Figures reproduced from arXiv: 2604.11981 by Jingang Yi, Peter Shan, Tao Liu, Xinyan Huang, Xunjie Chen.

Figure 1
Figure 1. Figure 1: Schematic of the bipedal walking model with foot sinkage and slip on granular media. (a) [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Bipedal-walking-dynamics modeling on granular media with the inclusion of foot–terrain [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Experimental setup of the bipedal robot walking locomotion on granular media. (a) [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Comparisons of robot locomotion results on granular media within one stance phase. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 6
Figure 6. Figure 6: (a)-(c) Foot orientation angle ∆θr (negative sign presents decreasing values), the effective rolling radius of the stance foot, and the CoT of walking on sand and rigid ground during one stance phase, respectively. Shadowed areas represent the standard deviations. v¯ x CoM/ √ ghCoM = 0.1. (d) Velocity dependency of locomotion CoT of the bipedal robot and human on sand and rigid ground, respectively. Each C… view at source ↗
Figure 5
Figure 5. Figure 5: Comparisons of RMSE under differ￾ent forward-velocity commands. (a) Joint an￾gle q. (b) GRFs. (c) Intrusion variables xs , ys , and zs . Circle dots: the proposed model. Square dots: model with rigid-ground assump￾tions. References [1] J. Aguilar, T. Zhang, F. Qian, M. Kingsbury, B. McIn￾roe, N. Mazouchova, C. Li, R. Maladen, C. Gong, M. Travers, R. L. Hatton, H. Choset, P. B. Umban￾howar, and D. I. Goldma… view at source ↗
read the original abstract

Bipeds have demonstrated high agility and mobility in unstructured environments such as sand. The yielding of such granular media brings significant sinkage and slip of the bipedal feet, leading to uncertainty and instability of walking locomotion. We present a new dynamics-modeling approach to capture and predict bipedal-walking locomotion on granular media. A dynamic foot-terrain interaction model is integrated to compute the ground reaction force (GRF). The proposed granular dynamic model has three additional degree-of-freedom (DoF) to estimate foot sinkage and slip that are critical to capturing robot-walking kinematics and kinetics such as cost of transport (CoT). Using the new model, we analyze bipedal kinetics, CoT, and foot-terrain rolling and intrusion affects. Experiments are conducted using a biped robotic walker on sand to validate the proposed dynamic model with robot-gait profiles, media-intrusion prediction, and GRF estimations. This new dynamics model can further serve as an enabling tool for locomotion control and optimization of bipedal robots to efficiently walk on granular terrains.

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

3 major / 2 minor

Summary. The paper proposes a new dynamics-modeling approach for bipedal walking on granular media that integrates a foot-terrain interaction model with three additional degrees of freedom to estimate foot sinkage and slip. This enables computation of ground reaction forces (GRF) and analysis of kinematics, kinetics, and cost of transport (CoT). The model is claimed to be validated experimentally on sand using a biped robot, comparing gait profiles, media intrusion, and GRF.

Significance. If the 3-DoF reduced-order model proves predictive without media-specific recalibration, it would offer a computationally efficient tool for simulation, control, and optimization of legged robots on yielding surfaces, addressing a practical gap in unstructured terrain locomotion where rigid-body assumptions break down.

major comments (3)
  1. [Abstract / Experimental validation] Abstract and experimental validation section: The central claim of predictive accuracy for gait profiles, intrusion, and GRF rests on experimental validation, yet no quantitative metrics (e.g., RMS error, R², or mean absolute deviation between model and measured GRF/CoT) are reported; without these, the strength of the validation cannot be assessed and the claim remains unverified.
  2. [Foot-terrain interaction model] Foot-terrain interaction model section: The assertion that three additional DoFs suffice to capture dominant nonlinear sinkage, slip, and rolling effects (without particle-level DEM fidelity or material-specific parameters) is load-bearing for generalization; the manuscript provides no sensitivity analysis to particle size distribution, compaction history, or cross-media tests to support this reduction.
  3. [Model formulation] Model derivation: No explicit equations or fitting procedure for the additional DoFs are shown in the provided text, preventing evaluation of whether predictions are independent of hidden calibration or circularly dependent on the same experimental data used for validation.
minor comments (2)
  1. [Model description] Notation for the three additional DoFs and their coupling to the biped dynamics is introduced without a clear diagram or variable table, making it hard to follow the integration with standard rigid-body equations.
  2. [Results] The abstract mentions analysis of 'foot-terrain rolling and intrusion affects' but the full manuscript should include a dedicated results subsection with time-series plots of predicted vs. measured intrusion depth to support the claims.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed comments on our manuscript. These have helped us clarify the presentation of our model and validation results. We provide point-by-point responses below and have revised the manuscript to address the concerns where possible.

read point-by-point responses

  1. Referee: [Abstract / Experimental validation] Abstract and experimental validation section: The central claim of predictive accuracy for gait profiles, intrusion, and GRF rests on experimental validation, yet no quantitative metrics (e.g., RMS error, R², or mean absolute deviation between model and measured GRF/CoT) are reported; without these, the strength of the validation cannot be assessed and the claim remains unverified.

    Authors: We agree that quantitative metrics strengthen the assessment of predictive accuracy. In the revised manuscript, we have added RMS error, R², and mean absolute deviation values in the experimental validation section for comparisons of model-predicted versus measured GRF, foot intrusion depths, gait profiles, and CoT. These metrics show good agreement (e.g., R² values above 0.85 across key quantities), supporting the central claims. revision: yes

  2. Referee: [Foot-terrain interaction model] Foot-terrain interaction model section: The assertion that three additional DoFs suffice to capture dominant nonlinear sinkage, slip, and rolling effects (without particle-level DEM fidelity or material-specific parameters) is load-bearing for generalization; the manuscript provides no sensitivity analysis to particle size distribution, compaction history, or cross-media tests to support this reduction.

    Authors: The 3-DoF reduction is derived from continuum-level granular mechanics to capture dominant effects without per-material recalibration. We have added a dedicated discussion subsection on model assumptions, limitations, and parameter sensitivity to particle properties and compaction. While we did not perform new cross-media experiments (which would require substantial additional resources), the sand-based validation demonstrates practical utility, and we explicitly note the need for future multi-media testing. revision: partial

  3. Referee: [Model formulation] Model derivation: No explicit equations or fitting procedure for the additional DoFs are shown in the provided text, preventing evaluation of whether predictions are independent of hidden calibration or circularly dependent on the same experimental data used for validation.

    Authors: The governing equations for the three additional DoFs (vertical sinkage, horizontal slip, and rolling) appear in Section 3.2, along with the GRF computation. The fitting uses independent single-foot intrusion and slip calibration experiments conducted separately from the bipedal walking trials. We have expanded the text to explicitly present the equations, detail the fitting procedure, and emphasize the separation between calibration and validation datasets to eliminate any ambiguity. revision: yes

Circularity Check

0 steps flagged

No circularity: model introduced with independent experimental validation on sand

full rationale

The provided abstract and description introduce a 3-DoF foot-terrain interaction model to estimate sinkage/slip and compute GRF, then validate it against robot gait profiles, media intrusion, and GRF data collected on sand. No equations, parameter-fitting procedures, self-citations, or uniqueness theorems are shown that would reduce any claimed prediction to the inputs by construction. The derivation chain therefore remains self-contained and externally falsifiable via the reported experiments rather than tautological.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review yields no explicit equations, so free parameters, axioms, and invented entities cannot be enumerated; the model implicitly assumes standard rigid-body dynamics plus a simplified granular interaction that is not detailed.

pith-pipeline@v0.9.0 · 5494 in / 1138 out tokens · 96519 ms · 2026-05-10T15:46:42.495397+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

30 extracted references · 30 canonical work pages

  1. [1]

    A review on locomotion robophysics: The study of movement at the intersection of robotics, soft matter and dynamical systems,

    J. Aguilar, T. Zhang, F. Qian, M. Kingsbury, B. McIn- roe, N. Mazouchova, C. Li, R. Maladen, C. Gong, M. Travers, R. L. Hatton, H. Choset, P. B. Umban- howar, and D. I. Goldman, “A review on locomotion robophysics: The study of movement at the intersection of robotics, soft matter and dynamical systems,” Rep. Prog. Phys., vol. 79, 2016, article 110001

    work page 2016

  2. [2]

    Maneu- vering on non-Newtonian fluidic terrain: A survey of animal and bio-inspired robot locomotion techniques on soft yielding grounds,

    S. Godon, M. Kruusmaa, and A. Ristolainen, “Maneu- vering on non-Newtonian fluidic terrain: A survey of animal and bio-inspired robot locomotion techniques on soft yielding grounds,”Front. Robotics AI, vol. 10, 2023, article 1113881

    work page 2023

  3. [3]

    Learning quadrupedal locomotion on de- formable terrain,

    S. Choi, G. Ji, J. Park, H. Kim, J. Mun, J. H. Lee, and J. Hwangbo, “Learning quadrupedal locomotion on de- formable terrain,” Sci. Robot., vol. 8, no. 74, 2023, arti- cle eade2256

    work page 2023

  4. [4]

    Cost of transport esti- mation for legged robot based on terrain features infer- ence from aerial scan,

    M. Pr ´agr, P. ˇC´ıˇzek, and J. Faigl, “Cost of transport esti- mation for legged robot based on terrain features infer- ence from aerial scan,” in Proc. IEEE/RSJ Int. Conf. In- tell. Robot. Syst., Madrid, Spain, 2018, pp. 1745–1750

    work page 2018

  5. [5]

    Comparative study of terrame- chanics properties of spherical and cylindrical feet for planetary legged robots on deformable terrain,

    H. Yang, C. Zhang, L. Ding, Q. Wei, H. Gao, G. Liu, L. Ge, and Z. Deng, “Comparative study of terrame- chanics properties of spherical and cylindrical feet for planetary legged robots on deformable terrain,” J. Ter- ramech., vol. 113, 2024, article 100968

    work page 2024

  6. [6]

    STAF: Interaction-based design and evalua- tion of sensorized terrain-adaptive foot for legged robot traversing on soft slopes,

    C. Yao, G. Shi, P. Xu, S. Lyu, Z. Qiang, Z. Zhu, L. Ding, and Z. Jia, “STAF: Interaction-based design and evalua- tion of sensorized terrain-adaptive foot for legged robot traversing on soft slopes,” IEEE/ASME Trans. Mecha- tronics, vol. 29, no. 6, pp. 4039–4050, 2024

    work page 2024

  7. [7]

    A terradynamics of legged locomotion on granular media,

    C. Li, T. Zhang, and D. I. Goldman, “A terradynamics of legged locomotion on granular media,” Science, vol. 339, no. 6126, pp. 1408–1412, 2013

    work page 2013

  8. [8]

    Robophysical study of jumping dynamics on granular media,

    J. Aguilar and D. I. Goldman, “Robophysical study of jumping dynamics on granular media,” Nat. Phys. , vol. 12, no. 3, pp. 278–283, 2016

    work page 2016

  9. [9]

    A dy- namic foot–terrain interaction model for biped walking on granular media,

    X. Chen, X. Huang, J. Yi, J. W. Shan, and T. Liu, “A dy- namic foot–terrain interaction model for biped walking on granular media,” IEEE/ASME Trans. Mechatronics, 2025

    work page 2025

  10. [10]

    A stability region criterion for flat-footed bipedal walking on de- formable granular terrain,

    X. Xiong, A. D. Ames, and D. I. Goldman, “A stability region criterion for flat-footed bipedal walking on de- formable granular terrain,” inProc. IEEE/RSJ Int. Conf. Intell. Robot. Syst., Vancouver, Canada, 2017, pp. 4552– 4559

    work page 2017

  11. [11]

    Foot shape- dependent resistive force model for bipedal walkers on granular terrains,

    X. Chen, A. Anikode, J. Yi, and T. Liu, “Foot shape- dependent resistive force model for bipedal walkers on granular terrains,” in Proc. IEEE Int. Conf. Robot. Au- tom., Yokohama, Japan, 2024, pp. 13 093–13 099

    work page 2024

  12. [12]

    Bio- inspired swing leg control for spring-mass robots run- ning on ground with unexpected height disturbance,

    H. Vejdani, Y . Blum, M. Daley, and J. Hurst, “Bio- inspired swing leg control for spring-mass robots run- ning on ground with unexpected height disturbance,” Bioinspir. Biomim., vol. 8, no. 4, 2013, article 046006

    work page 2013

  13. [13]

    Design and control of braver: A bipedal robot actuated via proprioceptive electric motors,

    Z. Zhu, W. Zhu, G. Zhang, T. Chen, Y . Li, X. Rong, R. Song, D. Qin, Q. Hua, and S. Ma, “Design and control of braver: A bipedal robot actuated via proprioceptive electric motors,”Auton. Robots, vol. 47, no. 8, pp. 1229– 1243, 2023

    work page 2023

  14. [14]

    Slip walking over rough terrain via h-lip stepping and backstepping-barrier function in- spired quadratic program,

    X. Xiong and A. Ames, “Slip walking over rough terrain via h-lip stepping and backstepping-barrier function in- spired quadratic program,” IEEE Robot. Automat. Lett., vol. 6, no. 2, pp. 2122–2129, 2021

    work page 2021

  15. [15]

    Predicting human walking gaits with a simple planar model,

    A. E. Martin and J. P. Schmiedeler, “Predicting human walking gaits with a simple planar model,” J. Biomech., vol. 47, no. 6, pp. 1416–1421, 2014

    work page 2014

  16. [16]

    Developing equations of motion for a planar biped walker with nonuniform foot shape,

    C. H. Rodman and A. E. Martin, “Developing equations of motion for a planar biped walker with nonuniform foot shape,” IFAC PapersOnLine, vol. 54, no. 20, pp. 455–462, 2021

    work page 2021

  17. [17]

    Bipedal model and hybrid zero dynamics of human walking with foot slip,

    M. Trkov, K. Chen, and J. Yi, “Bipedal model and hybrid zero dynamics of human walking with foot slip,” ASME J. Computat. Nonlinear Dyn., vol. 14, no. 10, 2019, arti- cle 101002

    work page 2019

  18. [18]

    Inertial sensor- based slip detection in human walking,

    M. Trkov, K. Chen, J. Yi, and T. Liu, “Inertial sensor- based slip detection in human walking,” IEEE Trans. Automat. Sci. Eng., vol. 16, no. 3, pp. 1399–1411, 2019

    work page 2019

  19. [19]

    Dynamics of shear bands in a dense granular material forced by a slowly moving rigid body,

    E. Hamm, F. Tapia, and F. Melo, “Dynamics of shear bands in a dense granular material forced by a slowly moving rigid body,” Phy. Rev. E, vol. 84, no. 4, 2011, ariticle 041304

    work page 2011

  20. [20]

    Impact-activated solidification of dense suspensions via dynamic jam- ming fronts,

    S. R. Waitukaitis and H. M. Jaeger, “Impact-activated solidification of dense suspensions via dynamic jam- ming fronts,” Nature, vol. 487, no. 7406, pp. 205–209, 2012

    work page 2012

  21. [21]

    Proprioceptive-based whole-body disturbance re- jection control for dynamic motions in legged robots,

    Z. Zhu, G. Zhang, Z. Sun, T. Chen, X. Rong, A. Xie, and Y . Li, “Proprioceptive-based whole-body disturbance re- jection control for dynamic motions in legged robots,” IEEE Robot. Automat. Lett. , vol. 8, no. 11, pp. 7703– 7710, 2023

    work page 2023

  22. [22]

    STANCE: Locomotion adaptation over soft terrain,

    S. Fahmi, M. Focchi, A. Radulescu, G. Fink, V . Bara- suol, and C. Semini, “STANCE: Locomotion adaptation over soft terrain,” IEEE Trans. Robotics, vol. 36, no. 2, pp. 443–457, 2020

    work page 2020

  23. [23]

    3-D underactuated bipedal walking via h-lip based gait synthesis and stepping stabi- lization,

    X. Xiong and A. Ames, “3-D underactuated bipedal walking via h-lip based gait synthesis and stepping stabi- lization,”IEEE Trans. Robotics, vol. 38, no. 4, pp. 2405– 2425, 2022

    work page 2022

  24. [24]

    Biomechanical compari- son of human walking locomotion on solid ground and sand,

    C. Zhu, X. Chen, and J. Yi, “Biomechanical compari- son of human walking locomotion on solid ground and sand,” ASME J. Biomech. Eng. , vol. 147, no. 4, 2025, article 041005

    work page 2025

  25. [25]

    Balance gait controller for a bipedal robotic walker with foot slip,

    M. Mihalec and J. Yi, “Balance gait controller for a bipedal robotic walker with foot slip,” IEEE/ASME Trans. Mechatronics, vol. 28, no. 4, pp. 2012–2019, 2023

    work page 2012

  26. [26]

    A comparative collision-based analysis of hu- man gait,

    D. V . Lee, T. N. Comanescu, M. T. Butcher, and J. E. Bertram, “A comparative collision-based analysis of hu- man gait,” in Proc. R. Soc. B: Biol. Sci. , vol. 280, no. 1771, 2013, article 20131779

    work page 2013

  27. [27]

    Me- chanics and energetics of human locomotion on sand,

    T. M. Lejeune, P. A. Willems, and N. C. Heglund, “Me- chanics and energetics of human locomotion on sand,” J. Exp. Biol., vol. 201, no. 13, pp. 2071–2080, 1998

    work page 2071

  28. [28]

    The Pinocchio C++ library: A fast and flexible implementation of rigid body dynamics algorithms and their analytical deriva- tives,

    J. Carpentier, G. Saurel, G. Buondonno, J. Mirabel, F. Lamiraux, O. Stasse, and N. Mansard, “The Pinocchio C++ library: A fast and flexible implementation of rigid body dynamics algorithms and their analytical deriva- tives,” in Proc. IEEE/SICE Int. Symp. Syst. Integration , 2019, pp. 614–619

    work page 2019

  29. [29]

    Benchmarking and workload analysis of robot dynamics algorithms,

    S. M. Neuman, T. Koolen, J. Drean, J. E. Miller, and S. Devadas, “Benchmarking and workload analysis of robot dynamics algorithms,” in Proc. IEEE/RSJ Int. Conf. Intell. Robot. Syst. , Macau, China, 2019, pp. 5235–5242

    work page 2019

  30. [30]

    A reduced-order mud re- action force model for robotic foot-mud interactions,

    X. Chen, J. Yi, and J. Shan, “A reduced-order mud re- action force model for robotic foot-mud interactions,” in Proc. IEEE/ASME Int. Conf. Adv. Intelli. Mechatronics, Boston, MA, USA, 2024, pp. 1341–1346

    work page 2024