Pith. sign in

REVIEW 3 major objections 4 minor 35 references

Continuum Robot State Estimation with Actuation Uncertainty

T0 review · 3 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read A discrete Cosserat rod factor graph with midpoint strain integration can jointly estimate shape, external loads, and actuation inputs of continuum robots in real time, while yielding manipulator Jacobians from the same linearized graph.

desk verdict Solid factor-graph framework for continuum robot state estimation; the Jacobian-extraction claim is oversold and the experimental reporting needs detail, but the core formulation is new and worth refereeing. read the letter →

arxiv 2601.04493 v3 pith:6GRS5VBJ submitted 2026-01-08 cs.RO

James M. Ferguson , Alan Kuntz , Tucker Hermans This is my paper · ORCID
classification cs.RO
keywords continuumrobotsstateestimationCosseratrodfactorgraphsactuationuncertaintytendon-drivenconcentrictubeforcesensing
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

This paper argues that continuum robots — thin flexible arms used in surgery — can be tracked in real time by reformulating the Cosserat rod equations as a discrete factor graph, with actuation treated as an uncertain variable rather than a fixed input. The key move is a midpoint strain integration rule that gives high numerical accuracy with only about 10–30 nodes along the backbone, keeping the optimization sparse. From the same linearized graph, the paper extracts manipulator Jacobians, so the estimated state can drive trajectory tracking. The authors show the approach on tendon-driven and parallel robots in simulation and on a surgical concentric tube robot, reporting mean tip position error of 1.95 mm and mean tip force error of 0.49 N at millisecond solve times.

What carries the argument

The central object is a discrete Cosserat rod factor graph: the backbone is divided into arclength nodes, each holding pose, stress, and wrench variables, connected by kinematics factors (midpoint strain rule) and mechanics factors (stress propagation with point loads). The midpoint strain rule, averaging endpoint stresses over each interval, is what buys accuracy with few nodes. A second mechanism is the actuation factor linking tendon tensions to backbone wrenches, plus platform constraint factors connecting multiple rods for parallel robots. The linearized graph's covariance blocks give the Jacobian.

What would settle it

A comparison experiment on a robot with a known distributed load (continuous contact along the shaft rather than point contact), with dense force sensors along the backbone: if estimated shape and force bias systematically with node count even at fine discretization, the midpoint-strain/discrete-load model is the culprit. Also, a calibration identifiability study: perturb stiffness and noise parameters and check whether the joint maximum-likelihood estimate recovers them uniquely; if not, the uncertainty envelopes are not trustworthy.

Watch

Extended reading notes

Core claim

The paper's central discovery is that the discrete Cosserat rod, when integrated with a midpoint strain rule, induces a sparse factor graph over SE(3) poses, internal stresses, and external wrenches; adding actuation variables as random nodes lets the graph jointly infer shape, external loads, and actuation inputs under uncertainty. A corollary is that the posterior covariance at the optimum directly yields the manipulator Jacobian relating actuation to tip motion, avoiding finite differences or separate sensitivity integration. The experiments show that this formulation matches a boundary-value-problem solver to better than 1.3% of robot length in open-loop simulation, and in hardware achie

Load-bearing premise

The load-bearing premise is that the discretized rod model—constant strain per interval computed from averaged endpoint stresses, loads concentrated at nodes, and a linear stiffness law—faithfully represents the physical robot, so any bias from distributed loads or calibration absorbing model error cannot be corrected by the factor graph.

Editorial extensions

If this is right

  • Real-time joint shape/load/actuation estimation becomes practical for tendon-driven, parallel, and concentric-tube continuum robots, with reported solve times around 2–24 ms.
  • Manipulator Jacobians come for free from the linearized factor graph, enabling closed-loop trajectory tracking without finite differences.
  • The same graph can run forward (known loads to shape) or inverse (observed shape to loads) by changing which priors and measurements are attached.
  • Actuation uncertainty, including backlash and torsion in endoscopic drive channels, is incorporated as noise on base poses rather than ignored.
  • Point-load discretization makes the model exact for tendon-disc loads and approximate for distributed loads, with accuracy controlled by node count.

Reading between the lines

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

  • A natural extension, not drawn in the paper: the factor-graph structure should also support nonlinear constitutive laws, since the midpoint strain factor does not depend on linear stiffness; repeating the accuracy benchmark for nonlinear materials would test this.
  • The observed axial-force ill-conditioning suggests a testable design rule: a second sensing modality along the shaft, not just tip pose, may be needed to make 3D force sensing fully observable; the paper itself flags the z-direction weakness.
  • Because the Jacobian comes from posterior covariances, the same linearization could be used for information-driven palpation, selecting future contacts to minimize force-estimate uncertainty—an application the paper does not mention.
  • A follow-up validation not in the paper: leave-one-out calibration checks would reveal whether the jointly calibrated stiffness and noise parameters are identifiable or merely absorbing model error; if the latter, the reported uncertainty envelopes could be overconfident.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. This paper proposes a discrete Cosserat rod formulation for continuum robot state estimation, cast as a sparse factor graph on SE(3). The kinematics use a midpoint strain integration rule (Eqs. 7-8), and the mechanics are discretized by concentrating external wrenches at arclength nodes (Eqs. 9-12). Actuation is introduced as random variables with priors (Eq. 18) and tendon actuation factors (Eqs. 19-21); parallel continuum robots are modeled by connecting multiple rod graphs with platform constraints (Eqs. 24-25). The authors claim high numerical accuracy with few nodes (1% error at K=10, 0.1% with K<30), real-time solve times (2-24 ms), direct extraction of manipulator Jacobians from the linearized graph (Eq. 23), and experimental validation on a concentric tube robot with 1.95 mm tip position accuracy and 0.49 N tip force accuracy. They also disclose that axial z-force estimation is ill-conditioned (Sec. VI).

Significance. If the claims hold, the framework is a useful contribution: it extends prior SE(3) / GP-Cosserat estimators by explicitly modeling actuation uncertainty, covers tendon-driven, parallel, and concentric-tube architectures in one graph, and demonstrates real-time sparse inference. Strengths include the internal consistency of the core error models with the stated SE(3) conventions, the BVP comparisons that quantify discretization error, the real physical experiment on a concentric tube robot, and the honest disclosure of the z-force degeneracy. The main weaknesses are that the 'manipulator Jacobian' extraction in Eq. (23) is a posterior regression coefficient rather than the mechanical Jacobian when measurement factors are active, and that the experimental calibration lacks identifiability analysis. These issues are local and correctable, but they affect load-bearing claims in the abstract and in the control experiments.

major comments (3)
  1. [Sec. IV-A, Eq. (23)] Eq. (23) defines J_TKQ = Sigma_TKQ Sigma_QQ^{-1}, the best linear predictor of tip pose from actuation under the posterior distribution that includes the tip-position measurement factor (22) and the actuation prior (18). This is an estimation/feedback gain, not the mechanical manipulator Jacobian mapping commanded actuation to tip motion under fixed external loads. Measurement information attenuates the slope; in a scalar proxy T = aQ + w with measurement z = T + v, the posterior slope is a sigma_v^2 / (sigma_w^2 + sigma_v^2), not a. Since Sec. IV-B uses no tip measurements while Sec. IV-C tracks with them, the Jacobian used for control is measurement-contaminated. Please re-label Eq. (23) as a posterior feedback gain or specify that measurement factors must be removed when extracting the mechanical Jacobian, and compare against finite-difference Jacobians of the open-loop model.
  2. [Sec. VI] The experimental calibration performs a joint maximum-likelihood optimization over tracker pose, tube curvatures, tube bending/torsion stiffness, and all actuation noise parameters from a single calibration dataset. No identifiability analysis, parameter count, or cross-validation is reported. If these parameters are non-identifiable, the calibrated values can absorb modeling error, and the reported 1.95 mm / 0.49 N accuracy and the uncertainty envelopes may not generalize. Please report parameter counts, constraints, cross-validation, and the sensitivity of downstream estimates to calibration choices.
  3. [Sec. III-B / IV-B / V-A] The simulation benchmarks compare the MAP solution against deterministic BVP solvers of the same Cosserat equations, so they validate the discretization and optimization, not the physical model fidelity. The discrete-load approximation in Eq. (9) is exact only for point loads such as tendon discs; for distributed loads, accuracy depends on node count K, and no convergence study or experiment with distributed contact loading is provided. The claims of validation 'across multiple robot architectures' should be tempered, or supplemented with a test against an independent model or an experiment with distributed loading.
minor comments (4)
  1. [Fig. 10 caption] Typography: 'realtive' should be 'relative', and the trailing phrase 'mean force error).' is incomplete/duplicated.
  2. [Sec. III-E] The text refers to 'the midpoint noise model (7)', but Eq. (7) is a deterministic strain rule; the noise enters later through n_epsilon_k. Please correct the terminology.
  3. [Appendix I] The linearization of the actuation error e_Dd in Eq. (21) is deferred to 'source code', which is not available at review. Since the source is only promised 'upon acceptance', please include the full derivatives in the appendix or in a supplementary document.
  4. [Sec. VI] The z-force degeneracy is an important limitation and is mentioned only in passing at the end of Sec. VI. It should be stated in the abstract or conclusions with a quantitative comparison of prior vs. posterior variance, since it tempers the force-sensing claims.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivation is self-contained and benchmarks are independent cross-checks.

full rationale

The paper's load-bearing derivations do not reduce to their inputs. The discrete Cosserat model (Eqs. 5-8) is a discretization of the continuous Cosserat equations, and its accuracy claim (Sec. III-E) is validated against a separate baseline BVP solver, which is a legitimate cross-check of numerical integration rather than a self-fulfilling fit. The tendon actuation model is based on an external citation [34], not on the authors' own prior work, and the parallel-robot and concentric-tube extensions reuse the same factor-graph formulation without importing any uniqueness theorem from self-citations. Experimental evaluation (Sec. VI) calibrates parameters on a separate calibration dataset and evaluates on a second dataset; the disclosed axial-force degeneracy shows the authors are not forcing agreement. The only potentially controversial step is Eq. (23), where the 'manipulator Jacobian' is defined as the posterior conditional regression coefficient J = Σ_{T_K Q} Σ_{QQ}^{-1}. This is a mathematical identity for a Gaussian factor graph, not a fitted prediction, and any concern that closed-loop measurement factors contaminate the Jacobian is a correctness/terminology issue, not circularity. Overall, no claim in the paper is equivalent by construction to its inputs, and self-citations are not load-bearing.

Assumptions & free parameters 6 free parameters · 7 assumptions · 0 invented entities

The ledger for the central estimation claims is small and mostly standard for the field; the main additions are the experimental calibration quantities (stiffnesses, curvatures, noise parameters, tracker pose), whose count and identifiability are not reported, and the hand-set simulation covariances. No new physical entities are introduced — the paper's inventions are inferential (new factors, a discretization rule, a Jacobian extraction).

free parameters (6)
  • tube bending/torsion stiffness K (inner and outer tubes) = not reported
    Fitted in the Sec VI joint maximum-likelihood calibration; the reported accuracy numbers (1.95 mm, 0.49 N) depend on these values.
  • tube pre-curvatures (natural curvature fields) = not reported
    Fitted in the Sec VI calibration ('tube curvatures'); these set the nominal strain ε-bar in Eq (3).
  • actuation noise covariances (base-pose priors, Σ_Q) = not reported
    Fitted in the Sec VI calibration ('all actuation noise parameters'); hand-chosen in simulation; the uncertainty envelopes in Figs 5-10 depend on them.
  • mechanics noise covariances Σ_T, Σ_S = 'small value' (Sec III-B)
    Hand-set to softly enforce equality (Sec III-B/D); the Laplace-approximation covariance and hence the extracted Jacobian (Eq 23) depend on them.
  • tip position measurement covariance Σ_p and tendon disturbance Σ_D = not reported in simulation parameters
    Hand-chosen in simulation; these control estimation behavior and the force-sensing results in Figs 5-6.
  • discretization node count K and spacing Δs = K=30 in simulations; K unspecified in experiment
    Design choices; the accuracy claims (1.3%, 0.06%, 1.95 mm) are reported at these discretizations, and the paper's '0.1% with less than K=30' claim depends on K.
assumptions (7)
  • domain assumption Cosserat rod linear constitutive law: generalized stress σ = K(ε − ε̄)
    Invoked in Eq (3); assumes linear elasticity with known stiffness K. Standard in the continuum-robotics literature [7], [34].
  • domain assumption External loading is concentrated at the K discretization nodes (Dirac-delta loads)
    Eq (9) replaces any distributed load by point wrenches at nodes; the paper argues a sufficiently fine discretization suffices, but this is an assumption about physical loading rather than a derived result.
  • domain assumption Per-interval strain is constant at the average of endpoint stresses (midpoint rule)
    Eqs (7)-(8); the entire numerical-accuracy claim rests on this integration rule. Its accuracy is shown empirically in simulation (Sec III-E), not derived.
  • domain assumption All noise is Gaussian with fixed covariances; posterior is summarized by the Laplace approximation at the MAP solution
    Sec III-D; enables the quadratic cost (16) and the Jacobian extraction (23). Assumes the posterior is well-approximated by a single Gaussian.
  • domain assumption Tendon force on a disc acts along the straight line through routing holes, tension-only, with friction modeled as small Gaussian noise
    Eqs (19)-(20), based on the friction model of [34]; friction and pulley effects beyond the noise term n_Dd are not modeled.
  • ad hoc to paper The posterior regression coefficient Σ_TKQ Σ^{-1}_QQ (Eq 23) equals the manipulator Jacobian
    Valid only in the linear-Gaussian limit; the coefficient depends on the graph's noise covariances and on which measurement factors are attached. The paper does not discuss this dependence or specify the graph used during tracking.
  • domain assumption The physical Virtuoso arm is adequately modeled as two serial Cosserat tubes (inner tube straight, no torsional tube interaction) with actuation uncertainty encoded only in base-pose covariances
    Sec VI; the paper explicitly lists 'torsionally interacting tubes' as future work, so the experimental validation excludes a known physical effect.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Continuum Robot State Estimation with Actuation Uncertainty." pith.science (2026). https://pith.science/paper/6GRS5VBJ

@misc{pith2026260104493,
  author       = {Pith},
  title        = {Pith review of: Continuum Robot State Estimation with Actuation Uncertainty},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6GRS5VBJ}},
  note         = {Machine review of arXiv:2601.04493}
}
read the original abstract

Continuum robots are flexible, slender manipulators well suited for confined surgical environments. In these settings, unknown interaction forces and model uncertainty significantly affect robot shape, motivating state estimation from external observations. Existing estimation methods either neglect actuation modeling or rely on simplified deterministic actuation models. In contrast, we jointly estimate robot shape, external loads, and actuation inputs using mechanically principled actuation priors. To achieve this, we present a discrete Cosserat rod formulation with piecewise-linear strain integration that provides high numerical accuracy while inducing a sparse factor graph structure for efficient nonlinear optimization. We extend the framework to tendon-driven and parallel robots in simulation and validate it experimentally on a surgical concentric tube robot. Overall, our approach enables principled real-time estimation across multiple robot architectures while providing direct access to manipulator Jacobians through the linearized factor graph.

Figures

Figures reproduced from arXiv: 2601.04493 by the authors.

Figure 1
Figure 1. Left: Our approach fuses uncertain actuation inputs and ex￾ternal backbone measurements with a prior mechanics model. Right: Example snapshot results from our experiments. When interaction forces are known (or assumed zero), we estimate a distribution for robot shape (top). In the more realistic case where loads are unknown, our approach estimates a conditional distribution for external loading, given backbone obser… view at source ↗
Figure 2
Figure 2. Factor graph representation of the discrete Cosserat rod model, the core graph module in our robots. Internal factors (red) encode noisy Cosserat equation constraints between state variables. To make the problem well-posed, appropriate boundary conditions (e.g. fixed base pose and known wrenches) must be specified as prior factors (green). Boundary factors (yellow) enforce consistency between the tip wrenches and ti… view at source ↗
Figure 4
Figure 4. Factor graph representation of the discrete Cosserat rod model with tendon actuation (showing one node between each disc and only 4 discs for clarity). The nodes and factors (red) along the rod are a condensed version of the Cosserat rod graph (see [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figures from the paper (6 more)
Figure 5
Figure 5. Figure 5: Example snapshots from our tendon robot simulations illustrating different operating modes. Left: Forward kinematics with actuation/modeling uncertainty (i.e. no external wrenches). The red ellipses illustrate the position components of pose uncertainty (2 − σ) along t…
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: Left: Factor graph representation of a parallel robot with base actuation [17]. Multiple Cosserat graphs (see [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: Parallel robot simulation results (24.02 ms mean solve time over 900 time steps). Left: Estimated tip force over the trajectory. Right: Overall position and force uncertainty p Tr(Σ) throughout the trajectory. V. PARALLEL CONTINUUM ROBOTS Given the base Cosserat rod gr…
Figure 9
Figure 9. Figure 9: Surgical concentric tube robot experiment setup. The robot was teleoperated to touch a silicone object several times, simulating soft tissue palpation. The object was attached to a force/torque sensor to measure interaction forces, and a magnetic tracking coil was atta…
Figure 10
Figure 10. Figure 10: Experimental concentric tube robot results Left: Forward model predicting shape given measured actuation and tip forces (2.41 ms mean solve time over 900 time steps). Mean tip position accuracy relative to the tracker was 1.95 mm. Right: Inverse model predicting loads…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 2 linked inside Pith

  1. [1]

    Continuum robots for medical applications: A survey,

    J. Burgner-Kahrs, D. C. Rucker, and H. Choset, “Continuum robots for medical applications: A survey,”IEEE Transactions on Robotics, vol. 31, no. 6, pp. 1261–1280, 2015

    2015

  2. [2]

    Continuum robots: An overview,

    M. Russo, S. M. H. Sadati, X. Dong, A. Mohammad, I. D. Walker, C. Bergeles, K. Xu, and D. A. Axinte, “Continuum robots: An overview,”Advanced Intelligent Systems, vol. 5, no. 5, 2023

    2023

  3. [3]

    T. D. Barfoot,State estimation for robotics. Cambridge University Press, 2017

    2017

  4. [4]

    Factor graphs: Exploiting structure in robotics,

    F. Dellaert, “Factor graphs: Exploiting structure in robotics,”Annual Review of Control, Robotics, and Autonomous Systems, vol. 4, no. 1, pp. 141–166, 2021

    2021

  5. [5]

    Continuum reconfigurable parallel robots for surgery: Shape sensing and state estimation with uncertainty,

    P. L. Anderson, A. W. Mahoney, and R. J. Webster, “Continuum reconfigurable parallel robots for surgery: Shape sensing and state estimation with uncertainty,”IEEE robotics and automation letters, vol. 2, no. 3, pp. 1617–1624, 2017

    2017

  6. [6]

    Unified shape and external load state estimation for continuum robots,

    J. M. Ferguson, D. C. Rucker, and R. J. Webster, “Unified shape and external load state estimation for continuum robots,”IEEE Transac- tions on Robotics, vol. 40, pp. 1813–1827, 2024

    2024

  7. [7]

    Continuum robot state estimation using gaussian process regression on se (3),

    S. Lilge, T. D. Barfoot, and J. Burgner-Kahrs, “Continuum robot state estimation using gaussian process regression on se (3),”The International Journal of Robotics Research, vol. 41, no. 13-14, pp. 1099–1120, 2022

    2022

  8. [8]

    State estimation for continuum multi-robot systems on se (3),

    ——, “State estimation for continuum multi-robot systems on se (3),” IEEE Transactions on Robotics, 2024

    2024

Show all 35 references
  1. [9]

    Incorporating control inputs in continuous- time gaussian process state estimation for robotics,

    S. Lilge and T. D. Barfoot, “Incorporating control inputs in continuous- time gaussian process state estimation for robotics,”Robotica, vol. 43, no. 3, pp. 1067–1086, 2025

    2025

  2. [10]

    Vision based 3-d shape sensing of flexible manipulators,

    D. B. Camarillo, K. E. Loewke, C. R. Carlson, and J. K. Salisbury, “Vision based 3-d shape sensing of flexible manipulators,” inIEEE International Conference on Robotics and Automation, 2008, pp. 2940–2947

    2008

  3. [11]

    Real-time shape estimation for continuum robots using vision,

    M. W. Hannan and I. D. Walker, “Real-time shape estimation for continuum robots using vision,”Robotica, vol. 23, no. 5, pp. 645– 651, 2005

    2005

  4. [12]

    Optimizing curvature sen- sor placement for fast, accurate shape sensing of continuum robots,

    B. Kim, J. Ha, F. C. Park, and P. E. Dupont, “Optimizing curvature sen- sor placement for fast, accurate shape sensing of continuum robots,” inIEEE international conference on robotics and automation, 2014, pp. 5374–5379

    2014

  5. [13]

    Shape sensing based on longitudinal strain measurements considering elongation, bending, and twisting,

    V . Modes, T. Ortmaier, and J. Burgner-Kahrs, “Shape sensing based on longitudinal strain measurements considering elongation, bending, and twisting,”IEEE Sensors Journal, vol. 21, no. 5, pp. 6712–6723, 2020

    2020

  6. [14]

    Fbg-based shape sensing tubes for continuum robots,

    S. C. Ryu and P. E. Dupont, “Fbg-based shape sensing tubes for continuum robots,” inIEEE International Conference on Robotics and Automation, 2014, pp. 3531–3537

    2014

  7. [15]

    Shape estimation and control of a soft continuum robot under external payloads,

    Q. Zhao, J. Lai, K. Huang, X. Hu, and H. K. Chu, “Shape estimation and control of a soft continuum robot under external payloads,” IEEE/ASME Transactions on Mechatronics, vol. 27, no. 5, pp. 2511– 2522, 2021

    2021

  8. [16]

    Learning the complete shape of concentric tube robots,

    A. Kuntz, A. Sethi, R. J. Webster, and R. Alterovitz, “Learning the complete shape of concentric tube robots,”IEEE transactions on medical robotics and bionics, vol. 2, no. 2, pp. 140–147, 2020

    2020

  9. [17]

    Parallel continuum robots: Modeling, analysis, and actuation-based force sensing,

    C. B. Black, J. Till, and D. C. Rucker, “Parallel continuum robots: Modeling, analysis, and actuation-based force sensing,”IEEE Trans- actions on Robotics, vol. 34, no. 1, pp. 29–47, 2017

    2017

  10. [18]

    Estimating forces along continuum robots,

    V . Aloi, K. T. Dang, E. J. Barth, and C. Rucker, “Estimating forces along continuum robots,”IEEE Robotics and Automation Letters, vol. 7, no. 4, pp. 8877–8884, 2022

    2022

  11. [19]

    Deflection-based force sensing for continuum robots: A probabilistic approach,

    D. C. Rucker and R. J. Webster, “Deflection-based force sensing for continuum robots: A probabilistic approach,” inIEEE/RSJ Interna- tional Conference on Intelligent Robots and Systems, 2011, pp. 3764– 3769

    2011

  12. [20]

    An ultrasound-based localization algorithm for catheter ablation guidance in the left atrium,

    A. Brij Koolwal, F. Barbagli, C. Carlson, and D. Liang, “An ultrasound-based localization algorithm for catheter ablation guidance in the left atrium,”The International Journal of Robotics Research, vol. 29, no. 6, pp. 643–665, 2010

    2010

  13. [21]

    Multi-modal local- ization algorithm for catheter interventions,

    J. A. Borgstadt, M. R. Zinn, and N. J. Ferrier, “Multi-modal local- ization algorithm for catheter interventions,” inIEEE international conference on robotics and automation, 2015, pp. 5350–5357

    2015

  14. [22]

    Force from shape—estimating the location and magnitude of the external force on flexible instruments,

    Q. Qiao, G. Borghesan, J. De Schutter, and E. Vander Poorten, “Force from shape—estimating the location and magnitude of the external force on flexible instruments,”IEEE Transactions on Robotics, vol. 37, no. 5, pp. 1826–1833, 2021

    2021

  15. [23]

    Statics and dynamics of continuum robots with general tendon routing and external loading,

    D. C. Rucker and R. J. Webster III, “Statics and dynamics of continuum robots with general tendon routing and external loading,” IEEE Transactions on Robotics, vol. 27, no. 6, pp. 1033–1044, 2011

    2011

  16. [24]

    Isogeometric collocation methods for cosserat rods and rod structures,

    O. Weeger, S.-K. Yeung, and M. L. Dunn, “Isogeometric collocation methods for cosserat rods and rod structures,”Computer Methods in Applied Mechanics and Engineering, vol. 316, pp. 100–122, 2017

    2017

  17. [25]

    Solving cosserat rod models via collocation and the magnus expansion,

    A. L. Orekhov and N. Simaan, “Solving cosserat rod models via collocation and the magnus expansion,” inIEEE/RSJ International Conference on Intelligent Robots and Systems, 2020, pp. 8653–8660

    2020

  18. [26]

    Control space reduction and real-time accurate modeling of continuum manipulators using ritz and ritz–galerkin methods,

    S. H. Sadati, S. E. Naghibi, I. D. Walker, K. Althoefer, and T. Nanayakkara, “Control space reduction and real-time accurate modeling of continuum manipulators using ritz and ritz–galerkin methods,”IEEE Robotics and Automation Letters, vol. 3, no. 1, pp. 328–335, 2017

    2017

  19. [27]

    A geometrically exact model for soft continuum robots: The finite element deformation space formulation,

    S. Grazioso, G. Di Gironimo, and B. Siciliano, “A geometrically exact model for soft continuum robots: The finite element deformation space formulation,”Soft robotics, vol. 6, no. 6, pp. 790–811, 2019

    2019

  20. [28]

    A unified method for solving inverse, forward, and hybrid manipulator dynamics using factor graphs,

    M. Xie and F. Dellaert, “A unified method for solving inverse, forward, and hybrid manipulator dynamics using factor graphs,”arXiv preprint arXiv:1911.10065, 2019

    1911 arXiv

  21. [29]

    A factor-graph approach for optimization problems with dynamics constraints,

    M. Xie, A. Escontrela, and F. Dellaert, “A factor-graph approach for optimization problems with dynamics constraints,”arXiv preprint arXiv:2011.06194, 2020

    2011 arXiv

  22. [30]

    A general framework for modeling and dynamic simulation of multibody systems using factor graphs,

    J.-L. Blanco-Claraco, A. Leanza, and G. Reina, “A general framework for modeling and dynamic simulation of multibody systems using factor graphs,”Nonlinear Dynamics, vol. 105, no. 3, pp. 2031–2053, 2021

  23. [31]

    Motion planning as probabilistic inference using gaussian processes and factor graphs

    J. Dong, M. Mukadam, F. Dellaert, and B. Boots, “Motion planning as probabilistic inference using gaussian processes and factor graphs.” inRobotics: Science and Systems, vol. 12, no. 4, 2016, pp. 10–15 607

    2016

  24. [32]

    A stochastic framework for continuous-time state estimation of continuum robots,

    S. Teetaert, S. Lilge, J. Burgner-Kahrs, and T. D. Barfoot, “A stochastic framework for continuous-time state estimation of continuum robots,” arXiv preprint arXiv:2510.01381, 2025

    2025

  25. [33]

    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

  26. [34]

    A general friction model of discrete interactions for tendon actuated dexterous manipulators,

    A. Gao, Y . Zou, Z. Wang, and H. Liu, “A general friction model of discrete interactions for tendon actuated dexterous manipulators,” Journal of Mechanisms and Robotics, vol. 9, no. 4, p. 041019, 2017

    2017

  27. [35]

    Computing jacobians and compliance matrices for externally loaded continuum robots,

    D. C. Rucker and R. J. Webster, “Computing jacobians and compliance matrices for externally loaded continuum robots,” in2011 IEEE international conference on robotics and automation. IEEE, 2011, pp. 945–950

    2011

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.