Pith. sign in

REVIEW 5 major objections 6 minor 1 cited by

Embedding high-resolution touch across robotic hands enables adaptive human-like grasping

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read High-resolution touch embedded across a robotic hand's surface enables near-perfect adaptive multi-object grasping in real-world trials.

desk verdict A genuinely impressive tactile hand, but the key experiment confounds tactile feedback with any closed-loop control; the hardware deserves review, the headline claim needs a better control. read the letter →

arxiv 2412.14482 v3 pith:MOWGTX7E submitted 2024-12-19 cs.RO

Zihang Zhao , Wanlin Li , Yuyang Li , Tengyu Liu , Boren Li , Meng Wang , Kai Du , Hangxin Liu
show 4 more authors
Yixin Zhu Qining Wang Kaspar Althoefer Song-Chun Zhu
This is my paper · ORCID
classification cs.RO MSC 68T4093C85
keywords F-TACHandtactilesensingroboticmulti-objectgraspingclosed-loopcontrolgraspsynthesisembodiedAIbiomimetic
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 claims that a robotic hand covered in high-resolution touch sensors, the F-TAC Hand, can adapt in real time to unexpected shifts and collisions during multi-object grasping, where feedback-free planning fails about half the time. The authors built a hand whose 70% of the surface senses at 0.1 mm resolution, kept full 15-degree-of-freedom human-like motion, and paired it with a training-free generative grasp planner. In 600 real-world trials across 60 object combinations, the tactile-informed controller succeeded on every trial (M = 1.000) while the same plans without tactile feedback succeeded only 53.5% of the time. If the result holds, it is evidence that rich, full-hand touch—not just better kinematics or vision—is what lets dexterous hands behave robustly under real-world uncertainty.

What carries the argument

The load-bearing mechanism is the F-TAC Hand's dense, vision-based tactile array: 17 photometric-stereo sensors in six configurations, about 10,000 taxels/$cm^{2}$ at 0.1 mm spatial resolution, covering 70% of the palmar surface. Encoder-decoder networks trained on a physics-based image-formation simulator convert elastomer shading into surface geometry, and a Poisson solver produces normal maps used to estimate object pose and detect collisions. The grasp repertoire comes from sampling hand configurations from a Gibbs distribution over force-closure energy using a modified Metropolis-Adjusted Langevin Algorithm, which yields diverse, human-like grasps without training data.

What would settle it

Run the same 600-trial protocol with a vision-only replanning controller that estimates object positions and collisions from a camera; if it matches the 100% success rate, the claim that high-resolution tactile sensing is critical would be overturned. Alternatively, degrade the tactile resolution (e.g., to 1 mm) while keeping closed-loop control; if success stays 100%, the resolution's role is not load-bearing.

Watch

Extended reading notes

Core claim

The central discovery is a quantitative demonstration that closing the loop with whole-hand tactile feedback turns a fragile multi-object grasp plan into a reliable one. Under execution noise, the theoretical plan collides at a rate of 46.5% (M = 0.465, SD = 0.306) versus the 0% predicted, showing a substantial sim-to-real gap; with tactile-informed replanning, success becomes perfect (M = 1.000, SD = 0.000), compared with M = 0.535 without feedback, t(59) = 11.8, p = 2.1 × $10^{-17}$. The hand estimates the grasped object's pose from contact geometry, detects imminent collisions, and switches to an alternative grasp strategy within ~100 ms, adding about 1 s of processing in collision-free cases.

Load-bearing premise

The baseline receives no feedback of any kind, so the measured benefit could come from closing the loop with any sensor rather than from the high resolution and spatial coverage of touch.

Editorial extensions

If this is right

  • Multi-object transport by robot hands becomes robust to positioning and perception noise, so pick-and-place tasks can pick up several objects per trip.
  • High-resolution tactile coverage can be integrated across a dexterous hand without sacrificing range of motion, as shown by the Kapandji test and execution of all 33 human grasp types.
  • Grasp synthesis that needs no training data can produce human-like, diverse grasps across different hand topologies.
  • Closed-loop tactile control adds negligible overhead (~1 s) in collision-free runs, making the approach practical for real-time use.
  • The results support the broader thesis that physical sensing capability is a core ingredient of intelligent manipulation, not an add-on.

Reading between the lines

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

  • The control experiment compares tactile feedback against no feedback at all; a stronger test would pit high-resolution touch against a vision-based replanning loop to confirm that rich tactile information, and not just closed-loop feedback, drives the gain.
  • The perfect success rate (M = 1.000, SD = 0.000) on 60 combinations suggests a ceiling effect; perturbing objects more aggressively or adding sensor noise could reveal the resolution at which touch's advantage degrades.
  • The same sensing architecture could enable in-hand manipulation and slip detection during transport, not only grasp replanning, because contact geometry is reconstructed at the finger level.
  • If confirmed, the design offers a replicable platform for studying embodied tactile intelligence: modular hand, physics-based sensor calibration, and training-free grasp synthesis could transfer to other hand morphology.
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

5 major / 6 minor

Summary. The paper presents F-TAC Hand, a biomimetic five-fingered robotic hand with 17 vision-based tactile sensors covering 70% of its palmar surface at a claimed 0.1 mm resolution. The hardware is matched with a generative grasp planner that samples hand configurations from a Gibbs distribution using a differentiable force-closure energy, and with a closed-loop controller that uses tactile readings to estimate in-hand object poses and replan multi-object grasps. The central empirical claim is that tactile-informed closed-loop adaptation achieves near-perfect multi-object grasping under execution noise (M = 1.000, SD = 0.000) whereas a non-tactile condition succeeds only 53.5% of the time (t(59) = 11.8, p = 2.1e-17), based on 60 object combinations and 600 real-world trials. The paper also reports kinematic validation (Kapandji score, all 33 Feix grasp types) and a sensor simulator calibrated with a physics-based image formation model.

Significance. If the central comparison were unconfounded, the paper would provide a strong existence proof that dense, high-resolution tactile sensing plus closed-loop replanning yields large robustness gains in a challenging multi-object manipulation task. The hardware contribution is substantial: integrating 17 photometric-stereo sensors into a 15-DoF anthropomorphic hand without losing range of motion is a genuine engineering achievement, and the authors provide open data and code as well as a calibrated, physics-based sensor simulator that may be reusable by the community. The reported 600-trial evaluation is a serious effort. However, the significance of the headline cognitive claim—that rich tactile embodiment per se is critical—is not yet established by the experimental design, because the comparison condition is open-loop and therefore confounds tactile feedback with feedback in general.

major comments (5)
  1. [Results, 'We quantified the impact of tactile sensing'; Methods, 'Context-sensitive motor controls'] The Figure 6d comparison does not isolate high-resolution tactile sensing. The non-tactile condition receives no feedback of any kind: the Methods state that 'Initial grasps were programmed from a disembodied AI perspective, using theoretically optimal strategies without considering environmental dynamics,' and the non-tactile-informed approach cannot replan after the first grasp. The tactile condition therefore changes two variables at once—it adds a sensor modality and it closes the feedback loop. The observed 0.465 success gap could plausibly be recovered by a vision-based or even a binary-contact closed-loop controller, which would leave the title and Discussion claim about the critical role of high-resolution touch unsupported. Please add a closed-loop control condition that uses an alternative feedback modality (or lower-resolution/binary touch) to disentangle sensing richness from loop closure, or explicitly reframe the conclusion as demonstrating the value of closed-loop tactile-informed replanning rather than of high-resolution touch specifically. The stress-test concern therefore lands.
  2. [Results, Figure 6c and 6d] The identical t(59) = 11.8 and p = 2.1e-17 are reported for two different comparisons: collision rate under perfect vs. real-world execution (means 0.000 and 0.465) and success rate with vs. without tactile-informed adaptation (means 1.000 and 0.535). These can only coincide if the paired differences have exactly the same mean and standard deviation, which is not obvious from the reported per-condition means and SDs. Please report the paired-difference statistics (mean difference, SD of the difference, and the 95% confidence interval) for each comparison, and make the trial-level data available in the Zenodo deposit so the t-tests can be reproduced.
  3. [Results, Figure 6c] The 'theoretical predictions' of zero collisions (M = 0.000, SD = 0.000) are true by construction: the theoretical planner assumes perfect execution and therefore cannot register a collision. The significant t-test in panel (c) quantifies execution noise against a constant, not the value of tactile sensing, and presenting it as evidence for the tactile system's contribution overstates its informativeness. Please reframe this panel as a sanity check on the presence of execution noise and move the causal argument entirely to the controlled comparison in panel (d).
  4. [Results, 'The tactile-informed approach achieved perfect adaptation'] The perfect success rate (M = 1.000, SD = 0.000) in the tactile condition is a zero-variance outcome over 60 replicates, and the success variable is binary per trial. A paired t-test on such a condition relies on the assumption that the paired differences are approximately normal, which is difficult to justify for a Bernoulli-derived proportion with a ceiling at 1.0. Please report a non-parametric paired test (e.g., Wilcoxon signed-rank) and an effect size with a confidence interval to accompany the t-test, and disclose how the 10 trials per object combination were aggregated into the 60 replicate proportions.
  5. [Supplementary Information S5] The claim that the grasp-generation algorithm produces 'human-like' diversity rests on human annotations of 3,450 generated grasps into 19 Feix categories, but no inter-annotator agreement statistic is reported. Given that the annotation protocol explicitly merges several Feix types and asks annotators for a two-step classification, please report Cohen's kappa (or equivalent) for the broad and fine categories, and state the number of annotators and the instructions used for ambiguous cases.
minor comments (6)
  1. [Results, Figure 6 caption] The caption states that the identical p-values 'derive from one-sided paired t-tests (t(59) = 11.8) comparing perfect vs. real-world execution (c) and with vs. without tactile-informed adaptation (d)', but it does not specify whether the tests were one-sided in the same direction for both panels; this should be stated explicitly.
  2. [Methods, 'Context-sensitive motor controls'] The four-ball illustration is helpful, but the text does not define how the 'success' of a multi-object grasp is scored in the reported statistics; please state the exact criterion (e.g., all planned objects transported, no drops, no collisions) in the Methods.
  3. [Discussion, first paragraph] The Discussion refers to 'unprecedented integration' and 'exceptional sensing' but does not quantify the trade-off between tactile coverage and sensor thickness or mechanical compliance; adding a few sentences on how the 70% coverage affects the hand's payload and joint range would strengthen the hardware claim.
  4. [Supplementary Information S3.1] The training details say the learning rate was 'halved every five epochs' but do not give the total number of epochs or the convergence criterion; adding these would make the training reproducible.
  5. [Figure 1] The schematic in Figure 1 includes human physiological benchmarks (afferents/cm^2) that are informative, but the font size for the '10000 taxels/cm²' annotation is very small; please enlarge it or move it to a table.
  6. [Abstract and Results] The phrase 'p < 0.0001' in the Abstract is weaker than the exact p = 2.1e-17 reported in the Results; for consistency, report the exact value or a standardized effect size in the Abstract as well.

Circularity Check

1 steps flagged · score 2.0 of 10

Minor self-definitional baseline: the zero-collision 'theoretical prediction' is true by construction of the planner, but the central tactile-vs-non-tactile comparison is empirical and not circular.

  1. self definitional [Results, Adaptive behaviors of F-TAC Hand, Figure 6c discussion]
    "The observed collision rate in real-world execution ( M = 0.465, SD= 0.306) differed significantly from theoretical predictions ( M = 0.000, SD= 0.000), highlighting the substantial gap between simulation and reality, t(59) = 11.8, p= 2.1×10 −17 (Figure 6c)."

    The 'theoretical prediction' of zero collisions is not an independent prediction; it is the planned output of the grasp planner, which by construction enforces no penetration through the physical plausibility energy Ephy (Eq. 4) that penalizes hand-object penetration and joint-limit violations. Under perfect execution, the planned grasp has zero collisions by definition of the planning constraints. The sim-to-real gap is therefore a comparison between a definitional baseline and measured execution, not a test of a predictive theory. This is a minor circularity: it demonstrates execution noise, but does not itself establish that high-resolution touch, rather than closed-loop feedback in general, is the causal factor.

full rationale

The paper's central claim is an empirical hardware evaluation: F-TAC Hand with tactile-informed closed-loop control achieves near-perfect multi-object grasping success (M = 1.000) versus a non-tactile open-loop baseline (M = 0.535) across 600 trials. That comparison is measured, not derived, so it is not circular. The only definitional element is the 'theoretical prediction' of zero collisions, which is true by construction because the planner's energy function (Eq. 4) penalizes hand-object penetration and joint-limit violations; this is a tautological baseline rather than a predictive theory. The non-tactile baseline being open-loop is a control confound (it does not isolate high-resolution touch from closed-loop feedback generally), but that is an experimental-design limitation, not a circularity. The paper does not rely on load-bearing self-citations: the cited force-closure estimator, ADELM, and grasp-taxonomy results are external tools or benchmarks, and no uniqueness theorem is invoked from the authors' own prior work. Overall, the central result retains independent empirical content, with only a minor self-definitional baseline inflating the reported sim-to-real gap.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim is an empirical hardware result. The grasp planner relies on a Gibbs energy formulation with hand-tuned weights, and the sensor model uses calibrated optical parameters; these are fitted inputs, not the target result. No theoretical entities are postulated; the hardware itself provides independent evidence via video and trials.

free parameters (6)
  • Energy weights lambda_grasp, lambda_phy = not stated
    Weights in Eq (2) balancing grasp quality and physical plausibility; they shape the generated grasp distribution and are not derived from first principles.
  • Force closure relaxation parameters (lambda0, delta, epsilon, mu) = not stated
    Parameters in Eq (S10)-(S11) softening the force closure constraints; chosen by the authors for tractable optimization.
  • Sensor model parameters (eta_k, mu_k, p_bar_k, sigma, rho_c) = calibrated per sensor type
    Fitted in Eq (S4) via Levenburg-Marquardt to real sensor readings; used to synthesize training data for the geometry reconstruction network.
  • Contact map thresholds (epsilon_1=0.0001, epsilon_2=0.05) = 0.0001, 0.05
    Empirically chosen in Supplementary Section S6 to truncate and scale contact distances for grasp classification clustering.
  • DPS network weights = learned
    Encoder-decoder network trained on simulated tactile images to reconstruct contact geometry; a learned component of the sensing pipeline.
  • MALA sampling hyperparameters = not stated in main text
    Grasp exploration uses modified Metropolis-Adjusted Langevin Algorithm; details such as step size and iteration count are only partially reported in Supplementary Section S7.
assumptions (5)
  • domain assumption Force closure is a valid criterion for grasp stability and the grasp quality energy Egrasp is defined as the minimum over contact sets of the force closure residual (Eq 3).
    The planner assumes a grasp is good if it can resist arbitrary external wrenches; this is a standard but nontrivial modeling choice for the energy landscape.
  • domain assumption Grasp configurations follow a Gibbs distribution with energy E(H,O) as defined in Eq (1)-(2).
    The generative model assumes Boltzmann-type sampling over hand poses conditioned on object geometry, which is a modeling assumption rather than a proven equivalence.
  • domain assumption The camera and lighting model for the tactile sensors is accurate enough for sim-to-real transfer (near-field perspective camera, point light sources, generalized Lambertian reflectance, known elastomer thickness).
    The sensor simulator in Supplementary Sections S2 relies on these optical assumptions to generate training data for the DPS network.
  • domain assumption Object geometry is known in advance for grasp planning and pose estimation.
    Stated in Discussion: 'our current implementation assumes known object geometry'; the adaptive system estimates pose, not shape, of the grasped object.
  • standard math Coulomb friction and friction cone constraints model contact forces at each contact point.
    Standard robotics force closure formulation used in Eq (S6) and the relaxation in Eq (S11).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Embedding high-resolution touch across robotic hands enables adaptive human-like grasping." pith.science (2026). https://pith.science/paper/MOWGTX7E

@misc{pith2026241214482,
  author       = {Pith},
  title        = {Pith review of: Embedding high-resolution touch across robotic hands enables adaptive human-like grasping},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MOWGTX7E}},
  note         = {Machine review of arXiv:2412.14482}
}
read the original abstract

Developing robotic hands that adapt to real-world dynamics remains a fundamental challenge in robotics and machine intelligence. Despite significant advances in replicating human hand kinematics and control algorithms, robotic systems still struggle to match human capabilities in dynamic environments, primarily due to inadequate tactile feedback. To bridge this gap, we present F-TAC Hand, a biomimetic hand featuring high-resolution tactile sensing (0.1mm spatial resolution) across 70% of its surface area. Through optimized hand design, we overcome traditional challenges in integrating high-resolution tactile sensors while preserving the full range of motion. The hand, powered by our generative algorithm that synthesizes human-like hand configurations, demonstrates robust grasping capabilities in dynamic real-world conditions. Extensive evaluation across 600 real-world trials demonstrates that this tactile-embodied system significantly outperforms non-tactile-informed alternatives in complex manipulation tasks (p<0.0001). These results provide empirical evidence for the critical role of rich tactile embodiment in developing advanced robotic intelligence, offering new perspectives on the relationship between physical sensing capabilities and intelligent behavior.

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. Taccel: Scaling Up Vision-based Tactile Robotics via High-performance GPU Simulation

    cs.RO 2025-04 conditional novelty 5.0 of 10

    Taccel combines Incremental Potential Contact and Affine Body Dynamics on GPUs to run parallel vision-based tactile simulations, with measured sim-to-real transfer in object classification and articulated manipulation.

Reference graph

Works this paper leans on

79 extracted references · 77 canonical work pages · cited by 1 Pith paper

  1. [1]

    Artificial Intelligence47(1-3), 139–159 (1991)

    Brooks, R.A.: Intelligence without representation. Artificial Intelligence47(1-3), 139–159 (1991)

    work page 1991

  2. [2]

    Nature571(7766), 18–18 (2019)

    Segal, M.: A more human approach to artificial intelligence. Nature571(7766), 18–18 (2019)

    work page 2019

  3. [3]

    Science 386(6727), 2950 (2024)

    Billard, A.G.: In good hands: A case for improving robotic dexterity. Science 386(6727), 2950 (2024)

    work page 2024

  4. [4]

    In: IEEE International Conference on Robotics and Automation (ICRA) (2011)

    Ma, R.R., Dollar, A.M.: On dexterity and dexterous manipulation. In: IEEE International Conference on Robotics and Automation (ICRA) (2011)

    work page 2011

  5. [5]

    Science 364(6446), 8414 (2019)

    Billard, A., Kragic, D.: Trends and challenges in robot manipulation. Science 364(6446), 8414 (2019)

    2019

  6. [6]

    Science Robotics9(91), 1501 (2024)

    Lepora, N.F.: The future lies in a pair of tactile hands. Science Robotics9(91), 1501 (2024)

    work page 2024

  7. [7]

    International Journal of Robotics Research (IJRR)3(4), 21–50 (1984)

    Jacobsen, S.C., Wood, J.E., Knutti, D., Biggers, K.B.: The utah/mit dextrous hand: Work in progress. International Journal of Robotics Research (IJRR)3(4), 21–50 (1984)

    work page 1984

  8. [8]

    International Journal of Robotics Research (IJRR)35(1-3), 161–185 (2016)

    Deimel, R., Brock, O.: A novel type of compliant and underactuated robotic hand for dexterous grasping. International Journal of Robotics Research (IJRR)35(1-3), 161–185 (2016)

    work page 2016

Show all 79 references
  1. [9]

    Science Robotics3(25), 3098 (2018)

    Hughes, J., Maiolino, P., Iida, F.: An anthropomorphic soft skeleton hand exploiting conditional models for piano playing. Science Robotics3(25), 3098 (2018)

    2018

  2. [10]

    33 Science Robotics7(68), 4155 (2022)

    De Pascali, C., Naselli, G.A., Palagi, S., Scharff, R.B., Mazzolai, B.: 3D-printed biomimetic artificial muscles using soft actuators that contract and elongate. 33 Science Robotics7(68), 4155 (2022)

    2022

  3. [11]

    Accessed: 2024 (2024)

    ShadowRobot: Shadow Dexterous Hand. Accessed: 2024 (2024). https://www. shadowrobot.com/dexterous-hand-series/

    2024

  4. [12]

    IEEE Robotics and Automation Letters (RA-L)7(2), 4821–4828 (2022)

    Morgan, A.S., Hang, K., Wen, B., Bekris, K., Dollar, A.M.: Complex in-hand manipulation via compliance-enabled finger gaiting and multi-modal planning. IEEE Robotics and Automation Letters (RA-L)7(2), 4821–4828 (2022)

    2022

  5. [13]

    IEEE Robotics and Automation Letters (RA-L) (2024)

    Li, Y., Liu, B., Geng, Y., Li, P., Yang, Y., Zhu, Y., Liu, T., Huang, S.: Grasp multiple objects with one hand. IEEE Robotics and Automation Letters (RA-L) (2024)

    2024

  6. [14]

    International Journal of Robotics Research (IJRR)39(1), 3–20 (2020)

    Andrychowicz, O.M., Baker, B., Chociej, M., Jozefowicz, R., McGrew, B., Pachocki, J., Petron, A., Plappert, M., Powell, G., Ray, A.,et al.: Learning dexterous in- hand manipulation. International Journal of Robotics Research (IJRR)39(1), 3–20 (2020)

    2020

  7. [15]

    In: Proceedings of European Conference on Computer Vision (ECCV) (2022)

    Qin, Y., Wu, Y.-H., Liu, S., Jiang, H., Yang, R., Fu, Y., Wang, X.: Dexmv: Imitation learning for dexterous manipulation from human videos. In: Proceedings of European Conference on Computer Vision (ECCV) (2022)

    2022

  8. [16]

    Science Robotics8(84), 9244 (2023)

    Chen, T., Tippur, M., Wu, S., Kumar, V., Adelson, E., Agrawal, P.: Visual dexterity: In-hand reorientation of novel and complex object shapes. Science Robotics8(84), 9244 (2023)

    2023

  9. [17]

    In: Proceedings of International Conference on Learning Representations (ICLR) (2024) 34

    Ma, Y.J., Liang, W., Wang, G., Huang, D.-A., Bastani, O., Jayaraman, D., Zhu, Y., Fan, L., Anandkumar, A.: Eureka: Human-level reward design via coding large language models. In: Proceedings of International Conference on Learning Representations (ICLR) (2024) 34

    2024

  10. [18]

    Experimental Brain Research53, 277–284 (1984)

    Westling, G., Johansson, R.S.: Factors influencing the force control during precision grip. Experimental Brain Research53, 277–284 (1984)

    1984

  11. [19]

    Nature Reviews Neuroscience10(5), 345–359 (2009)

    Johansson, R.S., Flanagan, J.R.: Coding and use of tactile signals from the fingertips in object manipulation tasks. Nature Reviews Neuroscience10(5), 345–359 (2009)

    2009

  12. [20]

    Brain60(4), 389–443 (1937)

    Penfield, W., Boldrey, E.: Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain60(4), 389–443 (1937)

    1937

  13. [21]

    Science 204(4392), 521–523 (1979)

    Kaas, J.H., Nelson, R.J., Sur, M., Lin, C.-S., Merzenich, M.M.: Multiple represen- tations of the body within the primary somatosensory cortex of primates. Science 204(4392), 521–523 (1979)

    1979

  14. [22]

    Annales de chirurgie de la main: organe officiel des societes de chirurgie de la main 5(1), 67–73 (1986)

    Kapandji, A.: Clinical test of apposition and counter-apposition of the thumb. Annales de chirurgie de la main: organe officiel des societes de chirurgie de la main 5(1), 67–73 (1986)

    1986

  15. [23]

    IEEE Transactions on Human-Machine Systems 46(1), 66–77 (2015)

    Feix, T., Romero, J., Schmiedmayer, H.-B., Dollar, A.M., Kragic, D.: The grasp taxonomy of human grasp types. IEEE Transactions on Human-Machine Systems 46(1), 66–77 (2015)

    2015

  16. [24]

    IEEE Transactions on Robotics (T-RO) (2023)

    Yao, K., Billard, A.: Exploiting kinematic redundancy for robotic grasping of multiple objects. IEEE Transactions on Robotics (T-RO) (2023)

    2023

  17. [25]

    International Journal of Robotics Research (IJRR) (2021)

    She, Y., Wang, S., Dong, S., Sunil, N., Rodriguez, A., Adelson, E.: Cable manipu- lation with a tactile-reactive gripper. International Journal of Robotics Research (IJRR) (2021)

    2021

  18. [26]

    International Journal of Robotics Research (IJRR)43(7), 1024–1055 (2024) 35

    Lloyd, J., Lepora, N.F.: Pose-and-shear-based tactile servoing. International Journal of Robotics Research (IJRR)43(7), 1024–1055 (2024) 35

    2024

  19. [27]

    IEEE Transactions on Robotics (T-RO) (2024)

    Zhao, Z., Li, Y., Li, W., Qi, Z., Ruan, L., Zhu, Y., Althoefer, K.: Tac-Man: Tactile- informed prior-free manipulation of articulated objects. IEEE Transactions on Robotics (T-RO) (2024)

    2024

  20. [28]

    Mind59(236), 433–460 (1950)

    Turing, A.M.: Computing machinery and intelligence. Mind59(236), 433–460 (1950)

    1950

  21. [29]

    American Association for the Advancement of Science (2024)

    Mitchell, M.: Debates on the nature of artificial general intelligence. American Association for the Advancement of Science (2024)

    2024

  22. [30]

    Journal of Biomechanics34(11), 1491–1500 (2001)

    Sancho-Bru, J., Perez-Gonzalez, A., Vergara-Monedero, M., Giurintano, D.: A 3-D dynamic model of human finger for studying free movements. Journal of Biomechanics34(11), 1491–1500 (2001)

    2001

  23. [31]

    Human Neurobiology3(1), 3–14 (1984)

    Vallbo, A.B., Johansson, R.S.,et al.: Properties of cutaneous mechanoreceptors in the human hand related to touch sensation. Human Neurobiology3(1), 3–14 (1984)

    1984

  24. [32]

    Optical Engineering19(1), 139–144 (1980)

    Woodham, R.J.: Photometric method for determining surface orientation from multiple images. Optical Engineering19(1), 139–144 (1980)

    1980

  25. [33]

    Sensors17(12), 2762 (2017)

    Yuan, W., Dong, S., Adelson, E.H.: Gelsight: High-resolution robot tactile sensors for estimating geometry and force. Sensors17(12), 2762 (2017)

    2017

  26. [34]

    Siciliano, B., Khatib, O., Kr¨ oger, T.: Springer Handbook of Robotics vol. 200. Springer, Berlin Heidelberg (2008)

    2008

  27. [35]

    Science Robotics5(48), 7710 (2020)

    Ichnowski, J., Avigal, Y., Satish, V., Goldberg, K.: Deep learning can accelerate grasp-optimized motion planning. Science Robotics5(48), 7710 (2020)

    2020

  28. [36]

    IEEE Robotics and Automation Letters (RA-L)7(1), 470–477 (2021)

    Liu, T., Liu, Z., Jiao, Z., Zhu, Y., Zhu, S.-C.: Synthesizing diverse and physically stable grasps with arbitrary hand structures using differentiable force closure 36 estimator. IEEE Robotics and Automation Letters (RA-L)7(1), 470–477 (2021)

    2021

  29. [37]

    Quarterly of Applied Mathematics77(2), 269–321 (2019)

    Hill, M., Nijkamp, E., Zhu, S.-C.: Building a telescope to look into high-dimensional image spaces. Quarterly of Applied Mathematics77(2), 269–321 (2019)

    2019

  30. [38]

    In: Robotics: Science and Systems (RSS) (2017)

    Mahler, J., Liang, J., Niyaz, S., Laskey, M., Doan, R., Liu, X., Ojea, J.A., Goldberg, K.: Dex-net 2.0: Deep learning to plan robust grasps with synthetic point clouds and analytic grasp metrics. In: Robotics: Science and Systems (RSS) (2017)

    2017

  31. [39]

    In: IEEE International Conference on Robotics and Automation (ICRA) (2023)

    Li, P., Liu, T., Li, Y., Geng, Y., Zhu, Y., Yang, Y., Huang, S.: Gendexgrasp: Generalizable dexterous grasping. In: IEEE International Conference on Robotics and Automation (ICRA) (2023)

    2023

  32. [40]

    Sensors and Actuators A: physical167(2), 171–187 (2011)

    Yousef, H., Boukallel, M., Althoefer, K.: Tactile sensing for dexterous in-hand manipulation in robotics—a review. Sensors and Actuators A: physical167(2), 171–187 (2011)

    2011

  33. [41]

    In: International Conference on Biomedical Robotics and Biomechatronics (BioRob) (2012)

    Fishel, J.A., Loeb, G.E.: Sensing tactile microvibrations with the bio- tac—comparison with human sensitivity. In: International Conference on Biomedical Robotics and Biomechatronics (BioRob) (2012)

    2012

  34. [42]

    Nature Communications5(1), 1–11 (2014)

    Kim, J., Lee, M., Shim, H.J., Ghaffari, R., Cho, H.R., Son, D., Jung, Y.H., Soh, M., Choi, C., Jung, S.,et al.: Stretchable silicon nanoribbon electronics for skin prosthesis. Nature Communications5(1), 1–11 (2014)

    2014

  35. [43]

    Autonomous Robots39, 25–41 (2015)

    Liu, H., Nguyen, K.C., Perdereau, V., Bimbo, J., Back, J., Godden, M., Seneviratne, L.D., Althoefer, K.: Finger contact sensing and the application in dexterous hand manipulation. Autonomous Robots39, 25–41 (2015)

    2015

  36. [44]

    Soft Robotics5(2), 216–227 (2018)

    Ward-Cherrier, B., Pestell, N., Cramphorn, L., Winstone, B., Giannaccini, M.E., Rossiter, J., Lepora, N.F.: The tactip family: Soft optical tactile sensors with 37 3d-printed biomimetic morphologies. Soft Robotics5(2), 216–227 (2018)

    2018

  37. [45]

    Sensors21(4), 4300–4309 (2020)

    Li, W., Alomainy, A., Vitanov, I., Noh, Y., Qi, P., Althoefer, K.: F-touch sensor: Concurrent geometry perception and multi-axis force measurement. Sensors21(4), 4300–4309 (2020)

    2020

  38. [46]

    Nature Machine Intelligence4(2), 135–145 (2022)

    Sun, H., Kuchenbecker, K.J., Martius, G.: A soft thumb-sized vision-based sensor with accurate all-round force perception. Nature Machine Intelligence4(2), 135–145 (2022)

    2022

  39. [47]

    IEEE Robotics and Automation Letters (RA-L)8(8), 5148–5155 (2023)

    Li, W., Wang, M., Li, J., Su, Y., Jha, D.K., Qian, X., Althoefer, K., Liu, H.: L 3 f-touch: A wireless gelsight with decoupled tactile and three-axis force sensing. IEEE Robotics and Automation Letters (RA-L)8(8), 5148–5155 (2023)

    2023

  40. [48]

    IEEE Robotics and Automation Letters (RA-L)9(12), 11170–11177 (2024)

    Li, W., Zhao, Z., Cui, L., Zhang, W., Liu, H., Li, L.-A., Zhu, Y.: Minitac: An ultra- compact 8 mm vision-based tactile sensor for enhanced palpation in robot-assisted minimally invasive surgery. IEEE Robotics and Automation Letters (RA-L)9(12), 11170–11177 (2024)

    2024

  41. [49]

    In: Proceedings of the Annual Conference on Computer Graphics and Interactive Techniques (1996)

    Curless, B., Levoy, M.: A volumetric method for building complex models from range images. In: Proceedings of the Annual Conference on Computer Graphics and Interactive Techniques (1996)

    1996

  42. [50]

    In: Proceedings of Advances in Neural Information Processing Systems (NeurIPS) (2021)

    Wang, P., Liu, L., Liu, Y., Theobalt, C., Komura, T., Wang, W.: Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction. In: Proceedings of Advances in Neural Information Processing Systems (NeurIPS) (2021)

    2021

  43. [51]

    Psychological Review63(2), 129 (1956) 38

    Simon, H.A.: Rational choice and the structure of the environment. Psychological Review63(2), 129 (1956) 38

    1956

  44. [52]

    Routledge, London (1962)

    Merleau-Ponty, M.: Phenomenology of Perception. Routledge, London (1962)

    1962

  45. [53]

    MIT Press, Cambridge (2017)

    Varela, F.J., Thompson, E., Rosch, E.: The Embodied Mind, Revised Edition: Cognitive Science and Human Experience. MIT Press, Cambridge (2017)

    2017

  46. [54]

    Science383(6682), 504–511 (2024)

    Vong, W.K., Wang, W., Orhan, A.E., Lake, B.M.: Grounded language acquisition through the eyes and ears of a single child. Science383(6682), 504–511 (2024)

    2024

  47. [55]

    Zenodo (2025)

    Zhao, Z., Li, W., Li, Y., Liu, T., Li, B., Wang, M., Du, K., Liu, H., Zhu, Y., Wang, Q., Althoefer, K., Zhu, S.-C.: Embedding high-resolution touch across robotic hands enables adaptive human-like grasping: data and codes. Zenodo (2025). https://doi.org/10.5281/zenodo.15193164

    2025 doi

  48. [56]

    Accessed: 2024 (2024)

    Robotics, S.: EZGripper. Accessed: 2024 (2024). https://sakerobotics.com/

    2024

  49. [57]

    Accessed: 2024 (2024)

    Technology, B.: Barrett BH8-282 3-fingered Gripper. Accessed: 2024 (2024). https: //advanced.barrett.com/barretthand

    2024

  50. [58]

    Accessed: 2024 (2024)

    Robotics, W.: Allegro Hand. Accessed: 2024 (2024). https://www.wonikrobotics. com/research-robot-hand

    2024

  51. [59]

    Transactions on Pattern Analysis and Machine Intelligence (TPAMI)22(11), 1330–1334 (2002)

    Zhang, Z.: A flexible new technique for camera calibration. Transactions on Pattern Analysis and Machine Intelligence (TPAMI)22(11), 1330–1334 (2002)

    2002

  52. [60]

    Nature Communications15(1), 9513 (2024)

    Jiang, C., Xu, W., Li, Y., Yu, Z., Wang, L., Hu, X., Xie, Z., Liu, Q., Yang, B., Wang, X.,et al.: Capturing forceful interaction with deformable objects using a deep learning-powered stretchable tactile array. Nature Communications15(1), 9513 (2024)

    2024

  53. [61]

    Cambridge university press, Cambridge, UK (2003) 39

    Hartley, R., Zisserman, A.: Multiple View Geometry in Computer Vision. Cambridge university press, Cambridge, UK (2003) 39

    2003

  54. [62]

    In: Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), vol

    Mitsunaga, T., Nayar, S.K.: Radiometric self calibration. In: Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), vol. 1, pp. 374–380 (1999)

    1999

  55. [63]

    International Journal of Computer Vision (IJCV)81, 155–166 (2009)

    Lepetit, V., Moreno-Noguer, F., Fua, P.: Epnp: An accurate O(n) solution to the pnp problem. International Journal of Computer Vision (IJCV)81, 155–166 (2009)

    2009

  56. [64]

    Transactions on Pattern Analysis and Machine Intelligence (TPAMI)25(3), 290–300 (2003)

    Sato, I., Sato, Y., Ikeuchi, K.: Illumination from shadows. Transactions on Pattern Analysis and Machine Intelligence (TPAMI)25(3), 290–300 (2003)

    2003

  57. [65]

    Science267(5201), 1153–1156 (1995)

    Nayar, S.K., Oren, M.: Visual appearance of matte surfaces. Science267(5201), 1153–1156 (1995)

    1995

  58. [66]

    ACM Transactions on Graphics (TOG)26(3) (2007)

    Katz, S., Tal, A., Basri, R.: Direct visibility of point sets. ACM Transactions on Graphics (TOG)26(3) (2007)

    2007

  59. [67]

    IEEE Robotics and Automation Letters (RA-L)7(2), 2361–2368 (2022)

    Si, Z., Yuan, W.: Taxim: An example-based simulation model for gelsight tactile sensors. IEEE Robotics and Automation Letters (RA-L)7(2), 2361–2368 (2022)

    2022

  60. [68]

    IEEE Robotics and Automation Letters (RA-L)6(2), 4177–4184 (2021)

    Gomes, D.F., Paoletti, P., Luo, S.: Generation of gelsight tactile images for sim2real learning. IEEE Robotics and Automation Letters (RA-L)6(2), 4177–4184 (2021)

    2021

  61. [69]

    In: Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR) (2019)

    Brahmbhatt, S., Ham, C., Kemp, C.C., Hays, J.: Contactdb: Analyzing and predicting grasp contact via thermal imaging. In: Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR) (2019)

    2019

  62. [70]

    Robotics and Autonomous Systems60(3), 326–336 (2012)

    Sahbani, A., El-Khoury, S., Bidaud, P.: An overview of 3D object grasp synthesis algorithms. Robotics and Autonomous Systems60(3), 326–336 (2012)

    2012

  63. [71]

    International Journal of Robotics Research (IJRR)31(7), 886–900 (2012) 40

    Rodriguez, A., Mason, M.T., Ferry, S.: From caging to grasping. International Journal of Robotics Research (IJRR)31(7), 886–900 (2012) 40

    2012

  64. [72]

    Robotics and Autonomous Systems60(3), 337–346 (2012)

    Prattichizzo, D., Malvezzi, M., Gabiccini, M., Bicchi, A.: On the manipulabil- ity ellipsoids of underactuated robotic hands with compliance. Robotics and Autonomous Systems60(3), 337–346 (2012)

    2012

  65. [73]

    Springer Handbook of Robotics, 955–988 (2016)

    Prattichizzo, D., Trinkle, J.C.: Grasping. Springer Handbook of Robotics, 955–988 (2016)

    2016

  66. [74]

    CRC press, Boca Raton (2017)

    Murray, R.M., Li, Z., Sastry, S.S.: A Mathematical Introduction to Robotic Manipulation. CRC press, Boca Raton (2017)

    2017

  67. [75]

    Science Robotics7(67), 0495 (2022)

    Yu, Y., Li, J., Solomon, S.A., Min, J., Tu, J., Guo, W., Xu, C., Song, Y., Gao, W.: All-printed soft human-machine interface for robotic physicochemical sensing. Science Robotics7(67), 0495 (2022)

    2022

  68. [76]

    Nature Electronics4(3), 193–201 (2021)

    Luo, Y., Li, Y., Sharma, P., Shou, W., Wu, K., Foshey, M., Li, B., Palacios, T., Torralba, A., Matusik, W.: Learning human–environment interactions using conformal tactile textiles. Nature Electronics4(3), 193–201 (2021)

    2021

  69. [77]

    Science Robotics5(49), 8134 (2020)

    Li, G., Liu, S., Wang, L., Zhu, R.: Skin-inspired quadruple tactile sensors integrated on a robot hand enable object recognition. Science Robotics5(49), 8134 (2020)

    2020

  70. [78]

    Science Robotics4(32), 2198 (2019)

    Lee, W.W., Tan, Y.J., Yao, H., Li, S., See, H.H., Hon, M., Ng, K.A., Xiong, B., Ho, J.S., Tee, B.C.: A neuro-inspired artificial peripheral nervous system for scalable electronic skins. Science Robotics4(32), 2198 (2019)

    2019

  71. [79]

    Sundaram, S., Kellnhofer, P., Li, Y., Zhu, J.-Y., Torralba, A., Matusik, W.: Learning the signatures of the human grasp using a scalable tactile glove. Nature 569(7758), 698–702 (2019) 41 Supplementary information Embedding high-resolution touch across robotic hands enables ad...

    2019

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.