Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

A Unified Framework for Probabilistic Dynamic-, Trajectory- and Vision-based Virtual Fixtures

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper claims that dynamical-system, trajectory, and visual-servoing virtual fixtures all reduce to probabilistic wrenches fused by one inverse-covariance rule, so a single controller spans autonomous, semi-automated, and manual haptic…

desk verdict A genuinely useful integration of probabilistic fixtures with real hardware validation, but the 'optimal' arbitration is a tuned weighting, not a derived probabilistic fusion. read the letter →

arxiv 2506.10239 v3 pith:LPWQSXEA submitted 2025-06-11 cs.RO

Maximilian Mühlbauer , Bernhard Weber , Sylvain Calinon , Freek Stulp , Alin Albu-Schäffer , João Silvério This is my paper · ORCID
classification cs.RO MSC 68T4093C85
keywords virtualfixturessharedcontrolproductofexpertskernelizedmovementprimitivesdynamicalsystemsvariableimpedanceRiemannianmanifoldsteleoperation
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 tries to establish that all robotic guidance aids known as virtual fixtures—coarse velocity guidance from learned dynamical systems, precise position-based trajectory guidance, and very precise visual-servoing guidance—can be written as Gaussian-distributed wrenches (combined force-and-torque commands) in the robot's end-effector space, and that a single product-of-experts fusion rule can arbitrate them according to their covariances. If true, this means a robot operator no longer needs hand-designed switching logic between assistance modes: the fixtures naturally take over and yield control based on their own uncertainty, from full autonomy to fine manual insertion. The paper also claims that the same covariances can be translated into a full, coupled stiffness matrix for impedance control, which the experiments report reduces measured interaction forces and improves usability for expert users across manual, semi-automated, and fully automated tasks. The broader point is that one controller, one arbitration formula, and one uncertainty representation can span the whole range of human-robot shared control.

What carries the argument

The central object is the probabilistic wrench: every fixture—the KMP-based dynamical-system VF, the position-based trajectory VF, and the visual-servoing VF—produces a Gaussian wrench $p(w_{VF,i}) = \mathcal{N}(\mu_{VF,i}, \Sigma_{VF,i})$ on its own manifold, which is transformed to a common cotangent space and fused by the product-of-experts formula of Eq. (15), $\hat{w} = (\sum_i \Sigma_{VF,i}^{-1})^{-1} \sum_i \Sigma_{VF,i}^{-1} \mu_{VF,i}$. This inverse-covariance weighting is what makes arbitration optimal in a least-squares sense and what lets the same covariance matrix drive the variable-impedance controller, which decomposes the rotated precision matrix into translational and rotational eigenscrew components, scales them by eigenvalue thresholds, and reassembles them into a fully populated stiffness matrix with positional-orientational couplings. For the new DS-based fixture, the machinery also includes a stabilizing policy $\mathcal{N}(\mu_{\mathrm{stab}}, \sigma_{\mathrm{stab}} I)$ that is activated by the same arbitration whenever the learned KMP policy's epistemic covariance becomes large, i.e., far from demonstrations.

What would settle it

Measure the actual tracking error or force deviation of each fixture over many trials and compare it with the covariance it outputs; if a fixture with a small covariance systematically produces larger errors than a fixture with a large covariance, the arbitration's weighting is not reflecting true authority, and the claimed optimality of Eq. (15) fails.

Watch

Extended reading notes

Core claim

Each virtual fixture outputs a wrench modeled as $w \sim \mathcal{N}(\mu,\Sigma)$ in the cotangent space of the end-effector manifold, and the framework fuses $N_{\mathrm{VF}}$ such wrenches by the Gaussian product $\hat{w} = (\sum_i \Sigma_{VF,i}^{-1})^{-1} \sum_i \Sigma_{VF,i}^{-1} \mu_{VF,i}$, which is the minimizer of the sum of squared Mahalanobis distances to the individual fixture wrenches. The covariance therefore serves double duty: it decides how much each fixture influences the combined wrench and, through a new variable-impedance scheme that decomposes the precision matrix into screw springs and torsional springs, it determines the robot's Cartesian stiffness, including couplings between translation and rotation. The paper introduces a probabilistic dynamical-system fixture based on kernelized movement primitives, whose epistemic uncertainty grows away from demonstrated data and triggers a stabilizing policy that steers the robot back, so that the same arbitration handles learned velocity fields, automated trajectory following, and vision-based corrections without explicit mode switching.

Load-bearing premise

The covariances of the fixtures are treated as trustworthy measures of their own authority; if those numbers are arbitrary or uncalibrated, the inverse-covariance fusion and the stiffness mapping reduce to heuristic weighting rather than a principled fusion.

Editorial extensions

If this is right

  • A single robot controller can now switch between fully autonomous execution, human-assisted semi-automation, and purely manual fine guidance without any discrete switching logic, because the fixtures arbitrate themselves through their covariances.
  • Covariance matrices that couple position and orientation can be realized as physical stiffness, so a human feels guidance along a geodesic between two detected targets instead of being pulled toward either one.
  • DS-based virtual fixtures can encode recurring motions (e.g., limit cycles) and multiple local dynamics, not only point attractors, enabling automated transport phases while keeping the operator able to intervene.
  • The framework reproduces the entire range of automation levels on different robots, with significantly lower measured interaction forces in user studies and comparable or better usability scores than an equal-weight baseline.
  • Because the same covariance drives arbitration, gating, and stiffness, only one object per fixture needs to be learned or set, which the authors argue simplifies programming of assistance.

Reading between the lines

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

  • If the covariances were calibrated from real sensing and tracking error rather than inherited from GMM input covariances, the same fusion rule would turn the framework into a proper multi-sensor Bayesian estimator; the paper does not make this calibration, and the DS fixture's covariance is constructed rather than measured.
  • The arbitration identity is generic: any collection of Gaussian experts that output wrenches—such as safety constraints, admittance filters, or learned residual policies—could be plugged into Eq. (15) without changing the controller, so the paper's unification likely extends beyond the three fixture types it tests.
  • A testable consequence the authors do not report: if the covariance truly reflects authority, then deliberately misreporting a fixture's covariance should measurably worsen task performance and increase forces; this could be checked in the same teleoperation setups used here.
  • The framework's reliance on hand-set hyperparameters (σ_stab for the stabilizing policy, eigenvalue thresholds for the stiffness) suggests that an automatic calibration from demonstrations would be the natural next step, as the discussion itself notes.
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

4 major / 5 minor

Summary. The paper proposes a unified probabilistic virtual fixture framework in which dynamical-system-based, position-based trajectory, and visual servoing fixtures are all represented as Gaussian wrenches and fused by a product-of-experts arbitration (Eq. 15). The authors introduce a KMP-based dynamical system VF with a stabilizing policy, a variable-impedance scheme that maps covariance precision into a fully populated Cartesian stiffness matrix, and extensions of trajectory and visual servoing fixtures to cylindrical and spherical manifolds. The framework is validated on three robot platforms, including a space-ready arm during parabolic flight and a user study with six expert operators, with quantitative comparisons of time, forces, tracking error, workload, and usability.

Significance. If the probabilistic semantics of the fused wrenches are made rigorous, the paper offers a practically valuable unification of three previously separate fixture types. The experimental breadth is a genuine strength: evaluations span manual, semi-automated, and fully automated modes, include repeated executions in microgravity, and report lower measured interaction forces and improved usability scores with the arbitration enabled. The coupled translational-rotational stiffness construction and the multi-manifold treatment are also useful contributions. However, the central 'optimal probabilistic arbitration' claim currently rests on covariance objects whose statistical meaning is not established, and the user-study statistics contain reporting inconsistencies that must be corrected.

major comments (4)
  1. [IV-A, Eqs. (10), (15), (28), (29), (33), and Appendix E] The load-bearing claim of optimal probabilistic arbitration is not supported because the covariance matrices entering Eq. (15) are not covariances of the wrenches being fused. Eq. (10) asserts w_i ~ N(mu_i, Sigma_i), but the DS fixture is defined as a Gaussian over velocity in Eq. (23), and the position-based and vision fixtures are defined as Gaussians over desired poses in Eqs. (29) and (33). The wrenches are produced through the impedance/velocity maps of Eq. (4) and Eq. (28), so the induced wrench covariance should be a pushforward of the pose or velocity covariance (at least D Sigma_DS D^T for the DS fixture and K Sigma_PB K^T in a linearized sense for the position-based fixture), not Sigma_i itself. Appendix E explicitly discards the velocity covariance and substitutes a pose-space covariance, which confirms that the covariance used in Eq. (15) is a pose/velocity covariance, not a wrench covariance. As written, the product-of-Gaussians solution is therefore not the posterior of a defined Gaussian wrench model, and the 'optimal' weighting is a unit-dependent heuristic. Please either derive the wrench-level covariance propagation from the underlying pose/velocity uncertainties or add an explicit calibration step and reframe the optimality claim accordingly.
  2. [IV-B, Eqs. (16)-(22), and Appendix A] The mapping from covariance to stiffness is presented as a translation of probabilistic uncertainty into impedance, but it is not derived from the uncertainty model. The precision matrix P = Sigma^{-1} has units of inverse pose or velocity variance, while the commanded stiffness has units of N/m and Nm/rad; the eigenvalue thresholds lambda^+ and lambda^- and the nominal stiffnesses k_nom are introduced as empirical parameters. No relation is established between the scales of the covariance and the scales of achievable stiffness, so the resulting K is a heuristic reshaping of the precision matrix rather than a calibrated impedance counterpart of the stated Gaussian wrench model. This is material because the variable-impedance behavior is a principal contribution and is used in all claimed 'optimal impedance gains' experiments. Please provide a derivation or an identification procedure that justifies interpreting the precision matrix as stiffness.
  3. [VIII-A, Table III] The reported Wilcoxon signed-rank p-values are not consistent with the stated design. Section VIII-A says the visual servoing stiffness experiment used six expert users with one repetition per condition; for n = 6, the two-sided exact p-value for W = 9 is about 0.84, and even W = 0 yields a two-sided p of about 0.031. The values W = 9 with p < 0.05 and W = 6 with p < 0.01 are therefore not reproducible. The same issue may affect Tables IV and V if the unit of analysis is repeated trials from the same users rather than independent subjects. Please report exact tests with the correct unit of analysis, or clarify how the listed p-values were computed.
  4. [V-B and Appendix A] The stabilizing policy's covariance is central to the DS-fixture arbitration, but its selection is not principled. Section V-B requires sigma_stab to be larger than the learned covariances near the demonstrations, while Appendix A recommends setting it smaller than the KMP uncertainty far from known points; Section VIII-E then acknowledges that sigma_stab = 0.04 m^2 is 'non-optimal' and generates forces when other fixtures should be active. This leaves the activation of the stabilizing policy, and hence the arbitration between learned and stabilizing policies, as a hand-tuned component. Please provide a data-based procedure for choosing sigma_stab or explicitly state that this part of the framework is heuristic.
minor comments (5)
  1. [VIII-G] The text states that 24 trials per condition were conducted but that one recording failed and the evaluation contains 23 trials; please clarify the actual N used in Table V and in the reported success rates.
  2. [Fig. 22] The caption contains 'M23', which appears to be a typographical error for 'M2'; please correct it.
  3. [Appendix A] The tuning guidance for the stabilizing policy covariance conflicts with Section V-B; please align the two descriptions or explicitly identify which recommendation applies in which regime.
  4. [VIII-C] The GP and KMP comparisons use different kernel hyperparameters and different covariance treatments, so the comparison is informative but not a strictly controlled ablation; please state this limitation explicitly.
  5. [Table VI] Several hyperparameters are introduced only in the text or figure captions; for reproducibility, please collect all final parameter values used in each experiment in a single table or appendix.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Eq. (15) is the closed-form solution of its own stated weighted least-squares objective, and the empirical claims are benchmarked externally (with/without arbitration, GP baseline, expert users).

full rationale

The derivation chain is not circular. Eq. (15) is explicitly presented as the closed-form solution of the minimization written directly above it; calling the result 'optimal' is definitional with respect to that stated objective, not a smuggled prediction. The covariance inputs are not fitted to the arbitration outputs: DS covariances come from KMP/GMM uncertainty (Appendix E), trajectory covariances from GMR, and vision covariances from assumed detections. The stabilizing-policy covariance is hand-set, and the paper openly states this requirement ('it has to be chosen such that Sigma_stab is larger than the covariances Sigma_DS,n ...'), so the activation behavior is a stated tuning rule rather than a prediction disguised as a result. Self-citations to [8] are prior published work used as a transparent foundation; the Gaussian-product arithmetic is standard PoE [55] and is restated, not imported as an unverified uniqueness or ansatz. The main statistically forced choices are calibrations (pose-space covariance used as a wrench-authority weight, manual hyperparameters), which affect validity and generalizability but do not make any claimed prediction equal to its inputs. External benchmarks (forces, time, TLX/SUS, success rate, GP comparison) provide independent support, as do the user studies. Appendix A's suggestion of likelihood-based tuning acknowledges the calibration limitation. Thus no circular step meets the evidentiary bar.

Assumptions & free parameters 12 free parameters · 5 assumptions · 0 invented entities

The central framework rests on a set of design choices rather than a physical law: Gaussian wrench models, covariance-as-authority, precision-to-stiffness mapping, and a zero-mean KMP prior. Many hyperparameters listed above must be set per experiment, making the reported behavior partly dependent on hand tuning, though the core operations (log/exp, PoE, impedance control) are standard.

free parameters (12)
  • KMP regularization lambda and lambda_c = lambda=0.05-0.1, lambda_c=10
    Control posterior mean and covariance; set per experiment, no automatic selection.
  • KMP scaling alpha = 0.1
    Scales KMP predictive covariance; central to arbitration weights.
  • RBF kernel length scale l = 0.03-0.3
    Determines extrapolation and epistemic-uncertainty growth for DS fixtures.
  • Stabilizing policy covariance sigma_stab = e.g., 0.04 m^2
    Hand-chosen larger than learned covariances near demonstrations so stabilizing policy acts only far from data.
  • Stabilizing policy default velocity x_dot_stab = 1 m/s
    Determines pull-back speed; manually selected.
  • Stiffness eigenvalue thresholds lambda_trans+-, lambda_rot+- = e.g., 1000/2500 N/m and 0.5/1.5
    Map precision eigenvalues to zero or full stiffness; set by hand per experiment.
  • Nominal stiffnesses k_trans,nom and k_rot,nom = e.g., 1000 N/m and 40 Nm/rad
    Spring stiffness magnitudes for eigenscrew reconstruction.
  • Distance adaptation d_min and d_max = e.g., 1 and 1.2
    Define where trajectory fixture precision is scaled down.
  • Visual-servoing length scale L and gating regularization gamma = gamma=1e-20, L per task
    Shape expert gating; hand-tuned until desired modulation is achieved (Appendix A).
  • Deadzone parameters l_dead, r_dead, l_add, r_dead,add = e.g., 5 mm and 0.2 rad
    Define zero-force regions near connectors.
  • Number of GMM components M = 2, 5, or 20
    Model complexity chosen per experiment.
  • Subsampling distance for demonstrations = 5 cm
    Data preprocessing; affects local density and kernel evaluation.
assumptions (5)
  • domain assumption Product of Experts fusion of Gaussian wrenches is a principled arbitration.
    Equation (15) treats inverse covariances as authority weights; paper does not validate Gaussianity of wrenches.
  • ad hoc to paper Covariance matrices computed in pose or velocity space represent uncertainty of wrenches.
    Appendix E constructs the DS covariance from the GMM input marginal; no pushforward through impedance or velocity mapping is derived.
  • standard math Precision matrix can be decomposed into eigenscrews and reassembled into a physically realizable stiffness matrix.
    Uses [42], [43], [57], [58] for spatial stiffness decomposition; the eigenvalue thresholding (19) is the paper's heuristic overlay.
  • standard math KMP prior with zero mean gives zero velocity far from demonstrations.
    KMP equations (8)-(9) with RBF kernel; used to justify the stabilizing policy.
  • domain assumption Wrench and covariance transformations between manifolds via Jacobian J_i,M are valid in the cotangent space.
    Appendix I gives J; correctness relies on coordinate conversions that are singular near r=0.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Unified Framework for Probabilistic Dynamic-, Trajectory- and Vision-based Virtual Fixtures." pith.science (2026). https://pith.science/paper/LPWQSXEA

@misc{pith2026250610239,
  author       = {Pith},
  title        = {Pith review of: A Unified Framework for Probabilistic Dynamic-, Trajectory- and Vision-based Virtual Fixtures},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LPWQSXEA}},
  note         = {Machine review of arXiv:2506.10239}
}
read the original abstract

Probabilistic Virtual Fixtures (VFs) enable the adaptive selection of the most suitable haptic feedback for each phase of a task, based on learned or perceived uncertainty. While keeping the human in the loop remains essential, for instance, to ensure high precision, partial automation of certain task phases is critical for productivity. We present a unified framework for probabilistic VFs that seamlessly switches between manual fixtures, semi-automated fixtures (with the human handling precise tasks), and full autonomy. We introduce a novel probabilistic Dynamical System-based VF for coarse guidance, enabling the robot to autonomously complete certain task phases while keeping the human operator in the loop. For tasks requiring precise guidance, we extend probabilistic position-based trajectory fixtures with automation, allowing for seamless human interaction, geometry-awareness and optimal impedance gains. For manual tasks requiring very precise guidance, we also extend visual servoing fixtures with the same geometry-awareness and impedance behavior. We validate our approach on different robots, including an evaluation with expert users, showcasing operation modes, the ease of programming fixtures and lower interaction forces and favorable usability compared to a baseline.

Figures

Figures reproduced from arXiv: 2506.10239 by the authors.

Figure 1
Figure 1. Overview of our unified framework. We propose a new type of [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Manifolds used in this work inspired by and using the notation of [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. shows the different coordinate systems used by our method. An easy transfer between different object placements is possible through the task coordinate system. Coordinates for individual fixtures placed relative to this coordinate system allow us to exploit the properties of different geometries. A. Virtual Fixture Arbitration Previous works [8] use a PoE [55] to arbitrate different fix￾ture wrenches expressed in th… view at source ↗
Figures from the paper (15 more)
Figure 4
Figure 4. Figure 4: 2D motion policy using a KMP. Black arrows visualize velocities in [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 22
Figure 22. Figure 22: The result of DTW is comparable to the introduction [PITH_FULL_IMAGE:figures/full_fig_p008_22.png]
Figure 5
Figure 5. Figure 5: A torque-controlled 7-DoF manipulator (left) is used in hand-guided mode for evaluating individual fixtures in Sections VIII-A to VIII-C, as well as their combinations in Sections VIII-F and VIII-G. A dual arm setup with two torque-controlled 7-DoF manipulators (center…
Figure 7
Figure 7. Figure 7: Position-based fixture (Section VI) on the spherical manifold [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 10
Figure 10. Figure 10: Translational (upper plot) and rotational (lower plot) velocities [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]
Figure 9
Figure 9. Figure 9: DS-based VF using a KMP model for the robot in start [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 14
Figure 14. Figure 14: Left image: The space robot setup approaching the spring [PITH_FULL_IMAGE:figures/full_fig_p012_14.png]
Figure 15
Figure 15. Figure 15: Repeated execution of the spring pressing task (Fig. 14) with different [PITH_FULL_IMAGE:figures/full_fig_p012_15.png]
Figure 16
Figure 16. Figure 16: DS-based VF DS combined with a position-based fixture PB on the space robot setup. The DS-based VF guides the robot towards the start of the automated position-based fixture performing a docking of the interface mounted at the end effector to the interface mounted on …
Figure 17
Figure 17. Figure 17: Norm of precisions (top), target velocities and position offset (middle), raw (dashed, wVF,i) and arbitrated (solid lines, Σˆ VFΣ −1 VF,iwVF,i resp. wˆ, see (15)) fixture forces (bottom) of the stabilizing policy, DS-based and trajectory (wPB, waut from VI-D) fixtures…
Figure 18
Figure 18. Figure 18: Semi-automated CubeSat subsystem assembly [6]–[8]: A DS-based [PITH_FULL_IMAGE:figures/full_fig_p014_18.png]
Figure 19
Figure 19. Figure 19: Combination of dynamical system and visual servoing fixtures on [PITH_FULL_IMAGE:figures/full_fig_p014_19.png]
Figure 20
Figure 20. Figure 20: Picking up a bottle B for moving it into a crate C using position-based trajectory, velocity and visual servoing fixtures. The rotation￾ally symmetric bottle can be picked up from multiple angles thanks to a probabilistic trajectory fixture on M2 with variable stiffne…
Figure 21
Figure 21. Figure 21: Norm of precisions (top), target velocities and position offset (middle), raw (dashed, wVF,i) and arbitrated (solid lines, Σˆ VFΣ −1 VF,iwVF,i resp. wˆ, see (15)) fixture forces (bottom) of the stabilizing policy, DS-based, trajectory and visual servoing fixtures used…
Figure 22
Figure 22. Figure 22: 2D projection of xee and the closest poses µj and µj+1 on a trajectory (red points) in M1 (xcart) as well as M23 (xcyl). The x axis through xee as well as the circle with the same radius on S 1 × R2 × S3 is plotted in dashed gray, note that according to the manifold m…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Passive Variable Impedance For Shared Control

    cs.RO 2026-04 unverdicted novelty 6.0 of 10

    Variable impedance and matrix-valued arbitration in shared control can be passivated by either lever-scaling spring deflection or rate-limiting full stiffness matrices, without position drift.

Reference graph

Works this paper leans on

74 extracted references · 54 canonical work pages · cited by 1 Pith paper

  1. [1]

    Virtual fixtures: Perceptual tools for telerobotic manip- ulation,

    L. Rosenberg, “Virtual fixtures: Perceptual tools for telerobotic manip- ulation,” inProceedings of IEEE Virtual Reality Annual International Symposium, 1993, pp. 76–82

    work page 1993

  2. [2]

    Active con- straints/virtual fixtures: A survey,

    S. A. Bowyer, B. L. Davies, and F. R. y Baena, “Active con- straints/virtual fixtures: A survey,”IEEE Transactions on Robotics, vol. 30, no. 1, pp. 138–157, Feb 2014

    work page 2014

  3. [3]

    Continuous transitions between levels of autonomy based on virtual fixtures for surgical robotic systems,

    K. Hagmann, A. Hellings-Kuss, F. Steidle, F. Stulp, D. Leidner, and J. Klodmann, “Continuous transitions between levels of autonomy based on virtual fixtures for surgical robotic systems,” in2024 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS 2024 (submitted), 2024. [Online]. Available: https://elib.dlr.de/203851/

    work page 2024

  4. [4]

    Human-centric con- cept for a reconfigurable robotic system enabling low-volume assembly of photonic and quantum modules,

    A. Wicht, T. Franke, A. Hahn, N. H ˚akansson, C. K ¨urbis, R. Smol, T. Hulin, T. Eiband, P. Lehner, M. M ¨uhlbauer, K. Nottensteiner, R. Pietschmann, B. Thaler, D. Thaler, and J. Bosse, “Human-centric con- cept for a reconfigurable robotic system enabling low-volume assembly of photonic and quantum modules,” in2025 IEEE/SICE International Symposium on Sys...

    work page 2025

  5. [5]

    Ai-enabled cyber–physical in- orbit factory - ai approaches based on digital twin technology for robotic 17 small satellite production,

    F. Leutert, D. Bohlig, F. Kempf, K. Schilling, M. M ¨uhlbauer, B. Ayan, T. Hulin, F. Stulp, A. Albu-Sch¨affer, V . Kutscher, C. Plesker, T. Dasbach, S. Damm, R. Anderl, and B. Schleich, “Ai-enabled cyber–physical in- orbit factory - ai approaches based on digital twin technology for robotic 17 small satellite production,”Acta Astronautica, vol. 217, pp. 1...

    work page 2024

  6. [6]

    Ai-based robust and failure- tolerant processes for in-orbit manufacturing of modular small satellites,

    M. M ¨uhlbauer, F. Leutert, C. Plesker, L. X. Wiedmann, A. M. Giordano, J. Silverio, T. Hulin, F. Stulp, J. V oges, R. Knobloch, K. Schilling, B. Schleich, and A. O. Albu-Sch ¨affer, “Ai-based robust and failure- tolerant processes for in-orbit manufacturing of modular small satellites,” in75th International Astronautical Congress, IAC 2024, 10 2024

    work page 2024

  7. [7]

    Multi-phase multi-modal haptic teleoperation,

    M. M ¨uhlbauer, F. Steinmetz, F. Stulp, T. Hulin, and A. Albu-Sch ¨affer, “Multi-phase multi-modal haptic teleoperation,” in2022 IEEE/RSJ Inter- national Conference on Intelligent Robots and Systems (IROS). IEEE, 10 2022

    work page 2022

  8. [8]

    A probabilistic approach to multi-modal adaptive virtual fixtures,

    M. M ¨uhlbauer, T. Hulin, B. Weber, S. Calinon, F. Stulp, A. Albu- Sch¨affer, and J. Silv ´erio, “A probabilistic approach to multi-modal adaptive virtual fixtures,”IEEE Robotics and Automation Letters, vol. 9, no. 6, pp. 5298–5305, 2024

    2024

Show all 74 references
  1. [9]

    Dexterous underwater manipulation from onshore locations: Streamlining efficiencies for remotely operated underwater vehicles,

    A. Birk, G. Antonelli, P. Di Lillo, E. Simetti, G. Casalino, G. Indiveri, L. Ostuni, A. Turetta, A. Caffaz, P. Weiss, T. Gobert, T. Doernbach, B. Chemisky, J. Gancet, T. Siedel, S. Govindaraj, X. Martinez, P. Letier, C. Mueller, T. Luczynski, A. Gomez Chavez, D. Koehntopp, A. ...

    2018

  2. [10]

    Movement imitation with nonlinear dynamical systems in humanoid robots,

    A. Ijspeert, J. Nakanishi, and S. Schaal, “Movement imitation with nonlinear dynamical systems in humanoid robots,” in2002 IEEE In- ternational Conference on Robotics and Automation. IEEE, 05 2002, pp. 1398–1403

    2002

  3. [11]

    Dynamical movement primitives: Learning attractor models for motor behaviors,

    A. J. Ijspeert, J. Nakanishi, H. Hoffmann, P. Pastor, and S. Schaal, “Dynamical movement primitives: Learning attractor models for motor behaviors,”Neural Computation, vol. 25, no. 2, pp. 328–373, 2013

    2013

  4. [12]

    Dynamic movement primitives in robotics: A tutorial survey,

    M. Saveriano, F. J. Abu-Dakka, A. Kramberger, and L. Peternel, “Dynamic movement primitives in robotics: A tutorial survey,”The International Journal of Robotics Research, vol. 42, no. 13, pp. 1133– 1184, 2023

    2023

  5. [13]

    Non-parametric regression for robot learning on manifolds,

    P. C. Lopez-Custodio, K. Bharath, A. Kucukyilmaz, and S. P. Preston, “Non-parametric regression for robot learning on manifolds,” 2024. [Online]. Available: https://arxiv.org/abs/2310.19561

    2024 arXiv

  6. [14]

    Implicit man- ifold gaussian process regression,

    B. Fichera, S. Borovitskiy, A. Krause, and A. G. Billard, “Implicit man- ifold gaussian process regression,” inAdvances in Neural Information Processing Systems, A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, Eds., vol. 36. Curran Associates, Inc., 2023, ...

    2023

  7. [15]

    Learning stable nonlinear dy- namical systems with gaussian mixture models,

    S. M. Khansari-Zadeh and A. Billard, “Learning stable nonlinear dy- namical systems with gaussian mixture models,”IEEE Transactions on Robotics, vol. 27, no. 5, pp. 943–957, 2011

    2011

  8. [16]

    Linearization and identification of multiple- attractor dynamical systems through laplacian eigenmaps,

    B. Fichera and A. Billard, “Linearization and identification of multiple- attractor dynamical systems through laplacian eigenmaps,”J. Mach. Learn. Res., vol. 23, no. 1, Jan 2022

    2022

  9. [17]

    Learning dynamical systems encoding non-linearity within space curvature,

    ——, “Learning dynamical systems encoding non-linearity within space curvature,” 2024. [Online]. Available: https://arxiv.org/abs/2403.11948

    2024 arXiv

  10. [18]

    Neural contractive dynamical systems,

    H. B. Mohammadi, S. Hauberg, G. Arvanitidis, N. Figueroa, G. Neu- mann, and L. Rozo, “Neural contractive dynamical systems,” inThe Twelfth International Conference on Learning Representations, 2023

    2023

  11. [19]

    Puma: Deep metric imitation learning for stable motion primitives,

    R. P ´erez-Dattari, C. Della Santina, and J. Kober, “Puma: Deep metric imitation learning for stable motion primitives,”Advanced Intelligent Systems, vol. 6, no. 11, 2024

    2024

  12. [20]

    Stable motion primitives via imitation and contrastive learning,

    R. P ´erez-Dattari and J. Kober, “Stable motion primitives via imitation and contrastive learning,”IEEE Transactions on Robotics, vol. 39, no. 5, pp. 3909–3928, 2023

    2023

  13. [21]

    Co-manipulation with a library of virtual guiding fixtures,

    G. Raiola, S. S. Restrepo, P. Chevalier, P. Rodriguez-Ayerbe, X. Lamy, S. Tliba, and F. Stulp, “Co-manipulation with a library of virtual guiding fixtures,”Autonomous Robots, vol. 42, no. 5, pp. 1037–1051, 11 2017

    2017

  14. [22]

    A geometric op- timal control approach for imitation and generalization of manipulation skills,

    B. Ti, A. Razmjoo, Y . Gao, J. Zhao, and S. Calinon, “A geometric op- timal control approach for imitation and generalization of manipulation skills,”Robotics and Autonomous Systems, vol. 164, p. 104413, 2023

    2023

  15. [23]

    Ilosa: Interactive learning of stiffness and attractors,

    G. Franzese, A. Meszaros, L. Peternel, and J. Kober, “Ilosa: Interactive learning of stiffness and attractors,” in2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 09 2021, pp. 7778–7785

    2021

  16. [24]

    Learning to pick at non-zero- velocity from interactive demonstrations,

    A. Meszaros, G. Franzese, and J. Kober, “Learning to pick at non-zero- velocity from interactive demonstrations,”IEEE Robotics and Automa- tion Letters, vol. 7, no. 3, pp. 6052–6059, 2022

    2022

  17. [25]

    Bayesian gaussian mixture model for robotic policy imitation,

    E. Pignat and S. Calinon, “Bayesian gaussian mixture model for robotic policy imitation,”IEEE Robotics and Automation Letters, vol. 4, no. 4, pp. 4452–4458, 2019

    2019

  18. [26]

    Kinesthetic guidance utilizing dmp synchronization and assistive virtual fixtures for progressive automation,

    D. Papageorgiou, F. Dimeas, T. Kastritsi, and Z. Doulgeri, “Kinesthetic guidance utilizing dmp synchronization and assistive virtual fixtures for progressive automation,”Robotica, vol. 38, no. 10, pp. 1824–1841, 2019

    2019

  19. [27]

    Motion encoding with asynchronous trajectories of repetitive teleoperation tasks and its extension to human-agent shared teleoperation,

    A. Pervez, H. Latifee, J.-H. Ryu, and D. Lee, “Motion encoding with asynchronous trajectories of repetitive teleoperation tasks and its extension to human-agent shared teleoperation,”Autonomous Robots, vol. 43, no. 8, pp. 2055–2069, 2019

    2019

  20. [28]

    Closed-loop variable stiffness control of dynamical systems,

    X. Chen, Y . Michel, and D. Lee, “Closed-loop variable stiffness control of dynamical systems,” in2020 IEEE-RAS 20th International Confer- ence on Humanoid Robots (Humanoids). IEEE, 07 2021

    2021

  21. [29]

    A shared control approach based on first-order dynamical systems and closed-loop variable stiffness control,

    H. Xue, Y . Michel, and D. Lee, “A shared control approach based on first-order dynamical systems and closed-loop variable stiffness control,” Journal of Intelligent & Robotic Systems, vol. 109, no. 4, 2023

    2023

  22. [30]

    A dynamical system approach to motion and force generation in contact tasks,

    W. Amanhoud, M. Khoramshahi, and A. Billard, “A dynamical system approach to motion and force generation in contact tasks,” inProceed- ings of Robotics: Science and Systems, June 2019

    2019

  23. [31]

    A digital twin approach for contextual assistance for surgeons during surgical robotics training,

    K. Hagmann, A. Hellings-Kuß, J. Klodmann, R. Richter, F. Stulp, and D. Leidner, “A digital twin approach for contextual assistance for surgeons during surgical robotics training,”Frontiers in Robotics and AI, vol. 8, 2021

    2021

  24. [32]

    Method for generating real-time interactive virtual fixture for shared teleoperation in unknown environments,

    V . Pruks and J.-H. Ryu, “Method for generating real-time interactive virtual fixture for shared teleoperation in unknown environments,”The International Journal of Robotics Research, vol. 41, no. 9–10, pp. 925– 951, 2022

    2022

  25. [33]

    Vision- assisted control for manipulation using virtual fixtures,

    A. Bettini, P. Marayong, S. Lang, A. Okamura, and G. Hager, “Vision- assisted control for manipulation using virtual fixtures,”IEEE Transac- tions on Robotics, vol. 20, no. 6, pp. 953–966, 2004

    2004

  26. [34]

    Programming by demonstration for shared control with an application in teleoperation,

    M. J. A. Zeestraten, I. Havoutis, and S. Calinon, “Programming by demonstration for shared control with an application in teleoperation,” IEEE Robotics and Automation Letters, vol. 3, no. 3, pp. 1848–1855, 2018

    2018

  27. [35]

    Haptic virtual fixtures for robot-assisted manipulation,

    J. J. Abbott, P. Marayong, and A. M. Okamura, “Haptic virtual fixtures for robot-assisted manipulation,” inRobotics Research: Results of the 12th International Symposium ISRR. Springer, 2007, pp. 49–64

    2007

  28. [36]

    Dynamic guidance with pseudoadmittance virtual fixtures,

    Z. Pezzementi, A. M. Okamura, and G. D. Hager, “Dynamic guidance with pseudoadmittance virtual fixtures,” inProceedings 2007 IEEE International Conference on Robotics and Automation, 2007, pp. 1761– 1767

    2007

  29. [37]

    Impedance control: An approach to manipulation,

    N. Hogan, “Impedance control: An approach to manipulation,” in1984 American Control Conference. IEEE, 07 1984

    1984

  30. [38]

    Bilateral teleoperation with adaptive impedance control for contact tasks,

    Y . Michel, R. Rahal, C. Pacchierotti, P. R. Giordano, and D. Lee, “Bilateral teleoperation with adaptive impedance control for contact tasks,”IEEE Robotics and Automation Letters, vol. 6, no. 3, pp. 5429– 5436, 2021

    2021

  31. [39]

    Orientation control with variable stiffness dynamical systems,

    Y . Michel, M. Saveriano, F. J. Abu-Dakka, and D. Lee, “Orientation control with variable stiffness dynamical systems,” in2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 10 2023

    2023

  32. [40]

    Passive arbitration in adaptive shared control of robots with variable force and stiffness scaling,

    R. Balachandran, M. De Stefano, H. Mishra, C. Ott, and A. Albu- Sch¨affer, “Passive arbitration in adaptive shared control of robots with variable force and stiffness scaling,”Mechatronics, vol. 90, p. 102930, 2023

    2023

  33. [41]

    A learning-based shared control architecture for interactive task execution,

    F. Abi-Farraj, T. Osa, N. P. J. Peters, G. Neumann, and P. R. Gior- dano, “A learning-based shared control architecture for interactive task execution,” in2017 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 05 2017

    2017

  34. [42]

    Achieving an Arbitrary Spatial Stiffness with Springs Connected in Parallel,

    S. Huang and J. M. Schimmels, “Achieving an Arbitrary Spatial Stiffness with Springs Connected in Parallel,”Journal of Mechanical Design, vol. 120, no. 4, pp. 520–526, 12 1998

    1998

  35. [43]

    The eigenscrew decomposition of spatial stiffness matrices,

    S. Huang and J. Schimmels, “The eigenscrew decomposition of spatial stiffness matrices,”IEEE Transactions on Robotics and Automation, vol. 16, no. 2, pp. 146–156, 2000

    2000

  36. [44]

    Autonomy in physical human-robot interaction: A brief survey,

    M. Selvaggio, M. Cognetti, S. Nikolaidis, S. Ivaldi, and B. Siciliano, “Autonomy in physical human-robot interaction: A brief survey,”IEEE Robotics and Automation Letters, vol. 6, no. 4, pp. 7989–7996, 2021

    2021

  37. [45]

    Enhancing bilateral teleoperation using camera-based online vir- tual fixtures generation,

    M. Selvaggio, G. Notomista, F. Chen, B. Gao, F. Trapani, and D. Cald- well, “Enhancing bilateral teleoperation using camera-based online vir- tual fixtures generation,” in2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 10 2016

    2016

  38. [46]

    A haptic shared-control architecture for guided multi-target robotic grasping,

    F. Abi-Farraj, C. Pacchierotti, O. Arenz, G. Neumann, and P. R. Giordano, “A haptic shared-control architecture for guided multi-target robotic grasping,”IEEE Transactions on Haptics, vol. 13, no. 2, pp. 270–285, 2020

    2020

  39. [47]

    An approach for imitation learning on Riemannian manifolds,

    M. J. A. Zeestraten, I. Havoutis, J. Silv ´erio, S. Calinon, and D. G. Cald- well, “An approach for imitation learning on Riemannian manifolds,” IEEE Robotics and Automation Letters, vol. 2, no. 3, pp. 1240–1247, 2017. 18

    2017

  40. [48]

    Gaussians on riemannian manifolds: Applications for robot learning and adaptive control,

    S. Calinon, “Gaussians on riemannian manifolds: Applications for robot learning and adaptive control,”IEEE Robotics & Automation Magazine, vol. 27, no. 2, pp. 33–45, 2020

    2020

  41. [49]

    J. M. Lee,Introduction to Smooth Manifolds. Springer New York, 2003

    2003

  42. [50]

    A tutorial on task-parameterized movement learning and retrieval,

    S. Calinon, “A tutorial on task-parameterized movement learning and retrieval,”Intelligent Service Robotics, vol. 9, no. 1, pp. 1–29, Sep 2015

    2015

  43. [51]

    Adaptive mixtures of local experts,

    R. A. Jacobs, M. I. Jordan, S. J. Nowlan, and G. E. Hinton, “Adaptive mixtures of local experts,”Neural Computation, vol. 3, no. 1, pp. 79–87, 1991

    1991

  44. [52]

    C. K. Williams and C. E. Rasmussen,Gaussian processes for machine learning. MIT press Cambridge, MA, 2006, vol. 2

    2006

  45. [53]

    Kernelized movement primitives,

    Y . Huang, L. Rozo, J. Silv ´erio, and D. G. Caldwell, “Kernelized movement primitives,”The International Journal of Robotics Research, vol. 38, no. 7, pp. 833–852, 2019

    2019

  46. [54]

    Geometry-aware bayesian optimization in robotics using riemannian mat ´ern kernels,

    N. Jaquier, V . Borovitskiy, A. Smolensky, A. Terenin, T. Asfour, and L. Rozo, “Geometry-aware bayesian optimization in robotics using riemannian mat ´ern kernels,” inProceedings of the 5th Conference on Robot Learning, ser. Proceedings of Machine Learning Research, vol

  47. [55]

    Products of experts,

    G. Hinton, “Products of experts,” in9th International Conference on Artificial Neural Networks: ICANN ’99, vol. 1999. IEEE, 1999, pp. 1–6

    1999

  48. [56]

    Impedance control on arbitrary surfaces for ultrasound scanning using discrete differential geometry,

    M. Dyck, A. Sachtler, J. Klodmann, and A. Albu-Schaffer, “Impedance control on arbitrary surfaces for ultrasound scanning using discrete differential geometry,”IEEE Robotics and Automation Letters, vol. 7, no. 3, pp. 7738–7746, 2022

    2022

  49. [57]

    The principal axes decom- position of spatial stiffness matrices,

    G. Chen, H. Wang, Z. Lin, and X. Lai, “The principal axes decom- position of spatial stiffness matrices,”IEEE Transactions on Robotics, vol. 31, no. 1, pp. 191–207, 2015

    2015

  50. [58]

    Comments on “the principal axes decomposition of spatial stiffness matrices

    S. Huang and J. M. Schimmels, “Comments on “the principal axes decomposition of spatial stiffness matrices”,”IEEE Transactions on Robotics, vol. 31, no. 6, pp. 1561–1564, 2015

    2015

  51. [59]

    Learning control,

    S. Calinon and D. Lee, “Learning control,” inHumanoid robotics: A reference. Springer Netherlands, 2017, pp. 1–52

    2017

  52. [60]

    Toward accurate dynamic time warping in linear time and space,

    S. Salvador and P. Chan, “Toward accurate dynamic time warping in linear time and space,”Intelligent Data Analysis, vol. 11, no. 5, pp. 561–580, 2007

    2007

  53. [61]

    C. M. Bishop,Pattern Recognition and Machine Learning, ser. Infor- mation Science and Statistics. Springer, 2006

    2006

  54. [62]

    Mixture of experts on riemannian manifolds for visual-servoing fixtures,

    M. M ¨uhlbauer, F. Stulp, A. O. Albu-Sch ¨affer, and J. Silv ´erio, “Mixture of experts on riemannian manifolds for visual-servoing fixtures,” in 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems, Workshop on Probabilistic Robotics in the Age of Deep Lea...

    2022

  55. [63]

    Development of nasa-tlx (task load index): Results of empirical and theoretical research,

    S. G. Hart and L. E. Staveland, “Development of nasa-tlx (task load index): Results of empirical and theoretical research,” inAdvances in psychology. Elsevier, 1988, vol. 52, pp. 139–183

    1988

  56. [64]

    Sus-a quick and dirty usability scale,

    J. Brookeet al., “Sus-a quick and dirty usability scale,”Usability evaluation in industry, vol. 189, no. 194, pp. 4–7, 1996

    1996

  57. [65]

    Passive variable impedance for shared control,

    M. M ¨uhlbauer, N. Werner, R. Balachandran, T. Hulin, J. Silv ´erio, F. Stulp, and A. Albu-Sch ¨affer, “Passive variable impedance for shared control,”arXiv preprint arXiv:2604.20557, 2026. [Online]. Available: https://arxiv.org/abs/2604.20557

    2026 arXiv

  58. [66]

    Learning potential functions from human demonstrations with encapsulated dynamic and compliant behaviors,

    S. M. Khansari-Zadeh and O. Khatib, “Learning potential functions from human demonstrations with encapsulated dynamic and compliant behaviors,”Autonomous Robots, vol. 41, no. 1, pp. 45–69, 2015

    2015

  59. [67]

    A passive admittance controller to enforce remote center of motion and tool spatial constraints with application in hands-on surgical procedures,

    T. Kastritsi and Z. Doulgeri, “A passive admittance controller to enforce remote center of motion and tool spatial constraints with application in hands-on surgical procedures,”Robotics and Autonomous Systems, vol. 152, p. 104073, 2022

    2022

  60. [68]

    A passive phri controller for assisting the user in partially known tasks,

    D. Papageorgiou, T. Kastritsi, Z. Doulgeri, and G. A. Rovithakis, “A passive phri controller for assisting the user in partially known tasks,” IEEE Transactions on Robotics, vol. 36, no. 3, pp. 802–815, 2020

    2020

  61. [69]

    A probabilistic approach for learning and adapting shared control skills with the human in the loop,

    G. Quere, F. Stulp, D. Filliat, and J. Silv ´erio, “A probabilistic approach for learning and adapting shared control skills with the human in the loop,” in2024 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 05 2024, pp. 15 728–15 734

    2024

  62. [70]

    Visual-servoing based mixture of experts fixture learning,

    J. K ¨uhn¨ol, “Visual-servoing based mixture of experts fixture learning,” Master’s thesis, Technical University of Munich, 2024. [Online]. Available: https://elib.dlr.de/210096/

    2024

  63. [71]

    Cartesian impedance control of redundant robots: recent results with the dlr-light- weight-arms,

    A. Albu-Sch ¨affer, C. Ott, U. Frese, and G. Hirzinger, “Cartesian impedance control of redundant robots: recent results with the dlr-light- weight-arms,” in2003 IEEE International Conference on Robotics and Automation, vol. 3, 2003, pp. 3704–3709 vol.3. TABLE VI: KEY NOTATION...

    2003

  64. [73]

    For smallr, the angleθis not well-defined

    Conversion betweenS 1 ×R2 ×S 3 andR 3 ×S 3:Position quantities can be converted fromM 1 toM 2 using θc =x/r, θ s =y/r, r= p x2 +y 2, z=z whereθ c andθ s denote the angleθaround the z axis as a complex number. For smallr, the angleθis not well-defined. The quaternion has to be ...

  65. [74]

    The quater- nion expressing the orientation has to be rotated by the inverse of the rotationq align aligning[0,0,1]with[s 0, s1, s2]: qM3 =q −1 align ·q M1

    Conversion betweenS 2 ×R1 ×S 3 andR 3 ×S 3:Position quantities can be converted fromM 1 toM 3 using s0 =x/r, s 1 =y/r, s 2 =z/r, r= p x2 +y 2 +z 2 For very smallr, valuess i are not well-defined. The quater- nion expressing the orientation has to be rotated by the inverse of t...

  66. [164]

    PMLR, 08–11 Nov 2022, pp. 794–805

    2022

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.