Pith. sign in

REVIEW 2 major objections 7 minor 2 cited by

eFlesh: Highly customizable Magnetic Touch Sensing using Cut-Cell Microstructures

T0 review · 2 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A magnetic skin built from 3D-printed microstructures gives robots sub-millimeter touch at hobbyist cost.

desk verdict eFlesh is a genuinely useful, accessible tactile sensor with a clever microstructure-based design, but the cross-instance transfer claim currently rests on one unreported number and needs a revision round, not a rejection. read the letter →

arxiv 2506.09994 v1 pith:2TLI4SK3 submitted 2025-06-11 cs.RO cs.AI

Venkatesh Pattabiraman , Zizhou Huang , Daniele Panozzo , Denis Zorin , Lerrel Pinto , Raunaq Bhirangi This is my paper
classification cs.ROcs.AI
keywords tactilesensingmagneticsensor3Dprintingcut-cellmicrostructuressoftroboticscontactlocalizationslipdetectionvisuotactilepolicylearning
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

eFlesh is a proposal for a tactile sensor that any robotics lab can fabricate from a hobbyist 3D printer, a few neodymium magnets, and a magnetometer board. The paper's central claim is that tiling a 3D-printed part with parameterized cut-cell microstructures makes both the shape and the mechanical response of the sensor designable, while embedded magnets convert surface deformation into measurable magnetic-field changes. If correct, this removes two barriers that have kept tactile sensing out of most manipulation pipelines: cost and geometry-specific fabrication. The paper demonstrates the payoff with sub-millimeter contact localization, accurate normal and shear force estimates, generalized slip detection, and visuotactile policies that clearly beat vision-only baselines on contact-rich tasks.

What carries the argument

The load-bearing object is the tiled cut-cell microstructure lattice, a parameterized family of printable cells whose beam dimensions and cell size set the effective Young's modulus of the TPU part. The design tool trims a voxelized lattice of these cells to any convex CAD shape and adds press-fit pouches for cylindrical neodymium magnets plus a slot for the magnetometer PCB. Deformation of the lattice moves the magnets relative to the Hall sensors, changing the magnetic flux density; raw 15-dimensional field readings are the input to all learned models. The alternating-polarity arrangement of neighboring magnets cancels the dominant far-field magnetic dipole terms, which is the mechanism that suppresses cross-sensor interference.

What would settle it

Measure force-displacement curves for a sweep of cell sizes and beam thicknesses on eFlesh blocks with embedded magnets and compare them to the stiffness predicted by the design tool; if the monotonic trend reverses, or if the variance across three printed instances exceeds the reported coefficient of variation, the customization and cross-instance consistency claims would fail.

Watch

Extended reading notes

Core claim

The central discovery claimed is that a single printed lattice of cut-cell microstructures can act simultaneously as the structural body, the compliance layer, and the magnetic transducer of a tactile sensor. By embedding press-fit magnet pouches in the mid-plane and placing a Hall-effect sensor board beneath, deformation of the microstructure displaces the magnets and changes the measured field. Learned models map the 15-dimensional field signal to contact position (0.5 mm RMSE in-plane, 0.16 mm in depth), normal force (0.27 N), and shear force (0.12 N). A linear classifier on windowed signal statistics detects slip on unseen objects at 95% accuracy, and transformer-based visuotactile policies reach 91% average success on four precision tasks, a 40% improvement over vision-only policies. The paper further claims that arranging magnets in alternating polarities reduces inter-sensor magnetic cross-talk by two orders of magnitude, and that three independently printed instances show less than 5% coefficient of variation in force response.

Load-bearing premise

The design tool assumes the small-strain microstructure stiffness map still predicts sensor behavior when eFlesh is deformed far beyond small strains with magnet pouches embedded; the paper confirms only that the qualitative trend holds, not the quantitative map.

Editorial extensions

If this is right

  • Any convex CAD part can be turned into a tactile sensor with only a hobbyist 3D printer, off-the-shelf magnets, and a magnetometer board, with no manual assembly beyond pausing the print to drop in magnets.
  • The demonstrated resolution (0.5 mm contact localization, 0.27 N normal force, 0.12 N shear force) is sufficient for precise manipulation tasks that need sub-millimeter alignment.
  • Slip can be detected with 95% accuracy on objects never seen during training, using only a linear classifier on windowed magnetometer statistics.
  • Visuotactile policies trained on eFlesh signals succeed 40% more often than vision-only policies and reach 91% average success on USB insertion, plug insertion, card swiping, and whiteboard erasing.
  • Alternating magnet polarities suppress inter-sensor magnetic cross-talk to the equivalent of less than 1 mm of deformation, making multi-sensor configurations practical.

Reading between the lines

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

  • A natural next step the paper does not take is a full design-space sweep connecting cell parameters to measured sensitivity, which would turn the qualitative stiffness trend into a quantitative lookup table.
  • The alternating-polarity layout points toward dense multi-sensor arrays: neighboring eFlesh cells could be packed tightly if their magnet polarities alternate, and the same cancellation principle should apply to other dipole-based magnetic sensors.
  • If cross-instance consistency holds under wear, the sensor would be a candidate for shared training datasets where data collected on one printed instance trains policies deployed on another; that is an implication beyond the paper's controlled three-instance measurement.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 7 minor

Summary. eFlesh presents a magnetic tactile sensor fabricated from 3D-printed cut-cell microstructures with embedded permanent magnets and a magnetometer PCB. The paper describes an open-source design tool that converts convex CAD shapes into printable lattice structures, and reports characterization results: contact localization RMSE of 0.5 mm, normal force RMSE of 0.27 N, shear force RMSE of 0.12 N, slip detection at 95% accuracy on unseen objects, and visuotactile policies that outperform vision-only baselines. It also proposes alternating magnet polarities to reduce cross-sensor interference and reports raw-signal consistency across independently fabricated instances. The central claims are evaluated against external ground truth with temporally split validation, but the cross-instance transfer of learned models—emphasized as a key advantage—is not reported numerically.

Significance. If the reported performance holds, eFlesh would be a valuable, accessible tactile sensor: it requires only a hobbyist 3D printer, off-the-shelf magnets, and a magnetometer board, and its lattice design allows shape and stiffness customization. The evaluation is generally well-designed: external ground truth (robot proprioception, weighing scale), temporally split validation sets, and held-out objects for slip detection. The open-source tool and data availability statements are strengths. However, the cross-instance generalization claim is not quantified for the learned localization task, and the stiffness-customization mapping is validated only qualitatively for a few parameter points; these gaps currently limit the substantiation of the plug-and-play and 'highly customizable' claims.

major comments (2)
  1. [§2.2, Fig. 3A; §2.7.2] The 0.5 mm contact-localization RMSE is computed on a temporal split of data from a single eFlesh instance. The only evidence for cross-instance generalization is a heatmap (Fig. 3A) with no numeric RMSE, and §2.7.2 reports raw signal-norm consistency (less than 1 N equivalent force, <5% coefficient of variation), which does not by itself imply that a trained localization model transfers. Because plug-and-play interchangeability and data reusability are explicitly cited as key advantages over prior sensors (Introduction, §2.7.2, Discussion), the paper must report the numeric unseen-instance localization RMSE, and ideally transfer metrics for force estimation and slip detection. Without this number, the headline 0.5 mm metric overstates the deployable performance.
  2. [§2.7, Fig. 6] The claim that users can tailor sensor response by specifying relative Young's modulus relies on an extrapolation that is only qualitatively checked. The paper states that the Tozoni et al. microstructure mapping was designed for small-strain regimes while eFlesh operates at large strains with embedded magnet pouches, and that 'the overall trends hold' based on force-compression curves for two variants. The design tool allows arbitrary combinations of cell size, beam dimensions, and spatially graded modulus, but only a coarse stiffness range (uniform E=0.01Ef and a depth-graded 0.002→0.001Ef) is characterized quantitatively. To support the 'highly customizable' and 'adjust sensitivity depending on task' claims, the paper should provide a quantitative validation of the stiffness-to-response map over a few representative design points, or explicitly limit the customizability claim to the demonstrated range.
minor comments (7)
  1. [Abstract, Table 1] The stated '40% improvement' should be expressed as percentage points, not as a relative percentage: Table 1 shows average success rates of 5.0/10 for vision-only and 9.08/10 for vision+eFlesh, which is an improvement of about 41 percentage points, or roughly an 82% relative improvement over the baseline.
  2. [Introduction] The citation 'gel (2023)' is informal and should be replaced with a proper reference to the Gelsight Mini product page.
  3. [§2.2] The notation 'RMSE_x,y' is undefined; please state that it is the root-mean-square error over the in-plane x and y coordinates.
  4. [Fig. 3A] The two heatmaps would be easier to compare if they shared a common color scale and included numeric tick labels for the error range.
  5. [§2.5] The 95% accuracy on unseen objects is reported for a single split; reporting a confidence interval or per-class breakdown would strengthen the claim.
  6. [§4.2.2] The text says the dataset is resampled to 10 Hz and later says it is subsampled by a factor of 5; please clarify the resulting effective rate.
  7. [§2.7] There is a typo: 'sensitivty' should be 'sensitivity'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: headline sensing and policy results are empirical fits to independent ground truth; the only self-citation (Tozoni microstructure framework) is prior published work and is supplemented by the paper's own qualitative validation.

full rationale

The paper's headline metrics—contact localization RMSE of 0.5 mm, normal force RMSE of 0.27 N, shear force RMSE of 0.12 N, slip detection accuracy of 95% on unseen objects, and 91% policy success—are produced by training models on raw magnetometer signals with ground-truth labels from robot proprioception, a weighing scale, human video annotation, or task success. These are empirical fits evaluated on temporally separated or held-out data, not quantities identical to the model inputs by construction. The localization experiment explicitly uses the first four indentation passes for training and the fifth pass for validation, and the force experiments adopt the same temporal-split protocol; this is genuine held-out prediction rather than a fitted input renamed as a prediction. The policy learning comparison isolates the tactile modality by using the same transformer architecture for vision-only and visuotactile policies, so the 40% improvement is not built into the inputs. The slip-detection classifier is tested on a separate set of 20 unseen objects, providing external generalization evidence. Two caveats are noted but they are not circularity. First, cross-instance localization transfer is shown only as a heatmap in Fig. 3A without a reported RMSE, and the cross-instance consistency evidence in Sec. 2.7.2 is raw signal-norm agreement rather than directly measured learned-model transfer; this is a missing-quantification/correctness concern, not a reduction of the claim to its inputs. Second, the design tool relies on the cut-cell microstructure stiffness mapping from Tozoni et al., a prior publication with overlapping authors. The paper candidly states in Sec. 2.7 that this framework 'was originally designed for small-strain regimes' while 'eFlesh sensors operate in large-strain regimes and contain inhomogeneities introduced by embedded magnet pouches,' and it supports the retained trend with its own force-displacement curves in Fig. 6B. Thus the self-citation is load-bearing for the convenient design parametrization but is not the sole evidence for the sensor's mechanical-response claims, and no equation or fitted parameter was found to be equivalent to the claimed result by definition. No circular step satisfies the evidentiary bar of quoting a specific equation-level reduction or fitted-input-as-prediction identity.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The paper contributes an empirical system, not a new physical theory. Its claims rest on hand-chosen design parameters, standard magnetostatics, and the untested extension of a prior microstructure model to large strains. No new particles, forces, or abstract entities are postulated.

free parameters (3)
  • Layer stiffness ratios (E/Ef) = 0.001, 0.0015, 0.002 (graded top-to-bottom)
    Hand-chosen design values that set sensor stiffness and sensitivity; central to the customizability claim, not fitted to a target response.
  • Microstructure cell size = 8 mm cube
    Chosen for print resolution and desired compliance; affects spatial resolution and force range, and is a user-specified input to the design tool.
  • Magnet geometry and placement = N52, 9.525 mm diameter x 3.175 mm thick, four pouches
    Off-the-shelf components whose size, number, and position shape the magnetic signal; treated as user-specified inputs in the design workflow.
assumptions (5)
  • domain assumption Tozoni et al. cut-cell microstructure stiffness model remains valid in large-strain regimes with embedded magnet pouches
    Stated in Section 2.7 as a limitation; the design tool uses Young's modulus mapping without quantitative validation in eFlesh's operating regime.
  • domain assumption Hall sensor readings change smoothly and repeatably with magnet displacement under deformation
    Required for all learned models; supported by the three-instance repeatability experiment in Fig. 7C, but not by an analytic transduction model.
  • domain assumption Weighing scale and robot proprioception provide accurate ground truth
    Used to label force and contact location training data in Sections 2.2, 2.3, and 2.4.
  • standard math Alternating magnet polarities reduce far-field interference via dipole cancellation
    Standard magnetostatics; the paper demonstrates a two-orders-of-magnitude reduction in stray-field intensity in Fig. 7D.
  • domain assumption TPU filament mechanical properties are consistent across prints
    Cross-instance consistency claims require print-to-print material stability; only three instances are tested in Section 2.7.2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of eFlesh: Highly customizable Magnetic Touch Sensing using Cut-Cell Microstructures." pith.science (2026). https://pith.science/paper/2TLI4SK3

@misc{pith2026250609994,
  author       = {Pith},
  title        = {Pith review of: eFlesh: Highly customizable Magnetic Touch Sensing using Cut-Cell Microstructures},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2TLI4SK3}},
  note         = {Machine review of arXiv:2506.09994}
}
abstract

If human experience is any guide, operating effectively in unstructured environments -- like homes and offices -- requires robots to sense the forces during physical interaction. Yet, the lack of a versatile, accessible, and easily customizable tactile sensor has led to fragmented, sensor-specific solutions in robotic manipulation -- and in many cases, to force-unaware, sensorless approaches. With eFlesh, we bridge this gap by introducing a magnetic tactile sensor that is low-cost, easy to fabricate, and highly customizable. Building an eFlesh sensor requires only four components: a hobbyist 3D printer, off-the-shelf magnets (<$5), a CAD model of the desired shape, and a magnetometer circuit board. The sensor is constructed from tiled, parameterized microstructures, which allow for tuning the sensor's geometry and its mechanical response. We provide an open-source design tool that converts convex OBJ/STL files into 3D-printable STLs for fabrication. This modular design framework enables users to create application-specific sensors, and to adjust sensitivity depending on the task. Our sensor characterization experiments demonstrate the capabilities of eFlesh: contact localization RMSE of 0.5 mm, and force prediction RMSE of 0.27 N for normal force and 0.12 N for shear force. We also present a learned slip detection model that generalizes to unseen objects with 95% accuracy, and visuotactile control policies that improve manipulation performance by 40% over vision-only baselines -- achieving 91% average success rate for four precise tasks that require sub-mm accuracy for successful completion. All design files, code and the CAD-to-eFlesh STL conversion tool are open-sourced and available on https://e-flesh.com.

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

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

  1. FELT: Generating Tactile Signals from Vision for Visuo-Tactile Manipulation

    cs.RO 2026-07 conditional novelty 6.0 of 10

    FELT predicts finger pressure maps from RGB images and uses them or their learned features to improve manipulation policies without real tactile sensors at deployment.

  2. 3D Cal: An Open-Source Software Library for Depth Reconstruction on Vision-Based Tactile Sensors

    cs.RO 2025-11 conditional novelty 6.0 of 10

    3D Cal repurposes a 3D printer as an automated calibration rig and trains a lightweight CNN, TouchNet, to reconstruct depth maps for DIGIT and GelSight Mini.

Reference graph

Works this paper leans on

66 extracted references · 51 canonical work pages · cited by 2 Pith papers

  1. [1]

    On a new action of the magnet on electric currents

    Edwin H Hall et al. On a new action of the magnet on electric currents. American Journal of Mathematics, 2 0 (3): 0 287--292, 1879

    work page

  2. [2]

    Biomimetic tactile sensor array

    Nicholas Wettels, Veronica J Santos, Roland S Johansson, and Gerald E Loeb. Biomimetic tactile sensor array. Advanced robotics, 22 0 (8): 0 829--849, 2008

    work page 2008

  3. [3]

    Flexible tactile sensing based on piezoresistive composites: A review

    Stefano Stassi, Valentina Cauda, Giancarlo Canavese, and Candido Fabrizio Pirri. Flexible tactile sensing based on piezoresistive composites: A review. Sensors, 14 0 (3): 0 5296--5332, 2014

    work page 2014

  4. [4]

    Elastic textures for additive fabrication

    Julian Panetta, Qingnan Zhou, Luigi Malomo, Nico Pietroni, Paolo Cignoni, and Denis Zorin. Elastic textures for additive fabrication. ACM Trans. Graph., 34 0 (4), July 2015. ISSN 0730-0301. doi:10.1145/2766937. URL https://doi.org/10.1145/2766937

    doi:10.1145/2766937 2015

  5. [5]

    Engeberg, and Jae-Won Choi

    Morteza Vatani, Yanfeng Lu, Erik D. Engeberg, and Jae-Won Choi. Combined 3d printing technologies and material for fabrication of tactile sensors. International Journal of Precision Engineering and Manufacturing, 16: 0 1375--1383, 2015. URL https://api.semanticscholar.org/CorpusID:137346890

    work page 2015

  6. [6]

    Stretch not flex: programmable rubber keyboard

    Daniel Xu, Andreas Tairych, and Iain Alexander Anderson. Stretch not flex: programmable rubber keyboard. Smart Materials and Structures, 25, 2015. URL https://api.semanticscholar.org/CorpusID:111206581

    work page 2015

  7. [7]

    Robust material classification with a tactile skin using deep learning

    Shiv S Baishya and Berthold B \"a uml. Robust material classification with a tactile skin using deep learning. In 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 8--15. IEEE, 2016

    work page 2016

  8. [8]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770--778, 2016. doi:10.1109/CVPR.2016.90

    doi:10.1109/cvpr.2016.90 2016

Show all 66 references
  1. [9]

    McAlpine

    Shuang-Zhuang Guo, Kaiyan Qiu, Fanben Meng, Sung Hyun Park, and Michael C. McAlpine. 3d printed stretchable tactile sensors. Advanced Materials, 29 0 (27): 0 1701218, 2017. doi:https://doi.org/10.1002/adma.201701218. URL https://advanced.onlinelibrary.wiley.com/doi/abs/10.1002...

    2017 doi

  2. [10]

    Worst-case stress relief for microstructures

    Julian Panetta, Abtin Rahimian, and Denis Zorin. Worst-case stress relief for microstructures. ACM Trans. Graph., 36 0 (4): 0 122:1--122:16, July 2017. ISSN 0730-0301. doi:10.1145/3072959.3073649. URL http://doi.acm.org/10.1145/3072959.3073649

    2017

  3. [11]

    Gelsight: High-resolution robot tactile sensors for estimating geometry and force

    Wenzhen Yuan, Siyuan Dong, and Edward H Adelson. Gelsight: High-resolution robot tactile sensors for estimating geometry and force. Sensors, 17 0 (12): 0 2762, 2017

    2017

  4. [12]

    Two-scale topology optimization with microstructures

    Bo Zhu, M\' e lina Skouras, Desai Chen, and Wojciech Matusik. Two-scale topology optimization with microstructures. ACM Trans. Graph., 36 0 (4), July 2017. ISSN 0730-0301. doi:10.1145/3072959.3095815. URL https://doi.org/10.1145/3072959.3095815

    2017

  5. [13]

    Adelson, and Sergey Levine

    Roberto Calandra, Andrew Owens, Dinesh Jayaraman, Justin Lin, Wenzhen Yuan, Jitendra Malik, Edward H. Adelson, and Sergey Levine. More than a feeling: Learning to grasp and regrasp using vision and touch. IEEE Robotics and Automation Letters, 3: 0 3300--3307, 2018. URL https:/...

    2018

  6. [14]

    Slip detection with combined tactile and visual information

    Jianhua Li, Siyuan Dong, and Edward Adelson. Slip detection with combined tactile and visual information. In 2018 IEEE International Conference on Robotics and Automation (ICRA), pages 7772--7777. IEEE, 2018 a

    2018

  7. [15]

    3d printed stretchable capacitive sensors for highly sensitive tactile and electrochemical sensing

    Kai Li, Hong Wei, Wenguang Liu, Hong Meng, Pei Zhang, and Chaoyi Yan. 3d printed stretchable capacitive sensors for highly sensitive tactile and electrochemical sensing. Nanotechnology, 29, 2018 b . URL https://api.semanticscholar.org/CorpusID:46827322

    2018

  8. [16]

    A new silicone structure for uskin—a soft, distributed, digital 3-axis skin sensor and its integration on the humanoid robot icub

    Tito Pradhono Tomo, Massimo Regoli, Alexander Schmitz, Lorenzo Natale, Harris Kristanto, Sophon Somlor, Lorenzo Jamone, Giorgio Metta, and Shigeki Sugano. A new silicone structure for uskin—a soft, distributed, digital 3-axis skin sensor and its integration on the humanoid rob...

    2018

  9. [17]

    A comprehensive realization of robot skin: Sensors, sensing, control, and applications

    Gordon Cheng, Emmanuel Dean-Leon, Florian Bergner, Julio Rogelio Guadarrama Olvera, Quentin Leboutet, and Philipp Mittendorfer. A comprehensive realization of robot skin: Sensors, sensing, control, and applications. Proceedings of the IEEE, PP: 0 1--18, 08 2019. doi:10.1109/JP...

    2019

  10. [18]

    Klatzky, and Helge J

    Sascha Fleer, Alexandra Moringen, Roberta L. Klatzky, and Helge J. Ritter. Learning efficient haptic shape exploration with a rigid tactile sensor array. PLoS ONE, 15, 2019. URL https://api.semanticscholar.org/CorpusID:67770678

    2019

  11. [19]

    Deformation capture via soft and stretchable sensor arrays

    Oliver Glauser, Daniele Panozzo, Otmar Hilliges, and Olga Sorkine-Hornung. Deformation capture via soft and stretchable sensor arrays. ACM Transactions on Graphics (TOG), 38 0 (2): 0 1--16, 2019

    2019

  12. [20]

    Soft magnetic skin for continuous deformation sensing

    Tess Hellebrekers, Oliver Kroemer, and Carmel Majidi. Soft magnetic skin for continuous deformation sensing. Advanced Intelligent Systems, 1 0 (4): 0 1900025, 2019

    2019

  13. [21]

    Learning to identify object instances by touch: Tactile recognition via multimodal matching

    Justin Lin, Roberto Calandra, and Sergey Levine. Learning to identify object instances by touch: Tactile recognition via multimodal matching. In 2019 International Conference on Robotics and Automation (ICRA), pages 3644--3650. IEEE, 2019

    2019

  14. [22]

    Design considerations for 3d printed, soft, multimaterial resistive sensors for soft robotics

    Benjamin Shih, Caleb Christianson, Kyle Gillespie, Sebastian Lee, Jason Mayeda, Zhaoyuan Huo, and Michael T Tolley. Design considerations for 3d printed, soft, multimaterial resistive sensors for soft robotics. Frontiers in Robotics and AI, 6: 0 30, 2019

    2019

  15. [23]

    Learning the signatures of the human grasp using a scalable tactile glove

    Subramanian Sundaram, Petr Kellnhofer, Yunzhu Li, Jun-Yan Zhu, Antonio Torralba, and Wojciech Matusik. Learning the signatures of the human grasp using a scalable tactile glove. Nature, 569 0 (7758): 0 698--702, 2019

    2019

  16. [24]

    Heterogeneous sensing in a multifunctional soft sensor for human-robot interfaces

    Taekyoung Kim, Sudong Lee, Taehwa Hong, Gyowook Shin, Taehwan Kim, and Yong-Lae Park. Heterogeneous sensing in a multifunctional soft sensor for human-robot interfaces. Science robotics, 5 0 (49): 0 eabc6878, 2020

    2020

  17. [25]

    Digit: A novel design for a low-cost compact high-resolution tactile sensor with application to in-hand manipulation

    Mike Lambeta, Po-Wei Chou, Stephen Tian, Brian Yang, Benjamin Maloon, Victoria Rose Most, Dave Stroud, Raymond Santos, Ahmad Byagowi, Gregg Kammerer, et al. Digit: A novel design for a low-cost compact high-resolution tactile sensor with application to in-hand manipulation. IE...

    2020

  18. [26]

    Electronic skins and machine learning for intelligent soft robots

    Benjamin Shih, Dylan Shah, Jinxing Li, Thomas G Thuruthel, Yong-Lae Park, Fumiya Iida, Zhenan Bao, Rebecca Kramer-Bottiglio, and Michael T Tolley. Electronic skins and machine learning for intelligent soft robots. Science Robotics, 5 0 (41): 0 eaaz9239, 2020

    2020

  19. [27]

    Capacitivo: Contact-based object recognition on interactive fabrics using capacitive sensing

    Te-Yen Wu, Lu Tan, Yuji Zhang, Teddy Seyed, and Xing-Dong Yang. Capacitivo: Contact-based object recognition on interactive fabrics using capacitive sensing. In Proceedings of the 33rd annual acm symposium on user interface software and technology, pages 649--661, 2020

    2020

  20. [28]

    McAlpine

    Zhijie Zhu, Hyun Soo Park, and Michael C. McAlpine. 3d printed deformable sensors. Science Advances, 6 0 (25): 0 eaba5575, 2020. doi:10.1126/sciadv.aba5575. URL https://www.science.org/doi/abs/10.1126/sciadv.aba5575

    2020 doi

  21. [29]

    Reskin: versatile, replaceable, lasting tactile skins

    Raunaq Bhirangi, Tess Hellebrekers, Carmel Majidi, and Abhinav Gupta. Reskin: versatile, replaceable, lasting tactile skins. In 5th Annual Conference on Robot Learning, 2021. URL https://openreview.net/forum?id=87_OJU4sw3V

    2021

  22. [30]

    A fully 3d‐printed wearable piezoresistive strain and tactile sensing array for robot hand

    Zhen Pei, Qiang Zhang, Kun Yang, Zhongyun Yuan, Wendong Zhang, and Shengbo Sang. A fully 3d‐printed wearable piezoresistive strain and tactile sensing array for robot hand. Advanced Materials Technologies, 6, 2021. URL https://api.semanticscholar.org/CorpusID:236420649

    2021

  23. [31]

    Gelsight wedge: Measuring high-resolution 3d contact geometry with a compact robot finger

    Shaoxiong Wang, Yu She, Branden Romero, and Edward Adelson. Gelsight wedge: Measuring high-resolution 3d contact geometry with a compact robot finger. In 2021 IEEE International Conference on Robotics and Automation (ICRA), pages 6468--6475. IEEE, 2021

    2021

  24. [32]

    Soft magnetic skin for super-resolution tactile sensing with force self-decoupling

    Youcan Yan, Zhe Hu, Zhengbao Yang, Wenzhen Yuan, Chaoyang Song, Jia Pan, and Yajing Shen. Soft magnetic skin for super-resolution tactile sensing with force self-decoupling. Science Robotics, 6, 2021. URL https://api.semanticscholar.org/CorpusID:232040651

    2021

  25. [33]

    Three-dimensional printing of tactile sensors for soft robotics

    Xinran Zhou and Pooi See Lee. Three-dimensional printing of tactile sensors for soft robotics. MRS Bulletin, 46: 0 330 -- 336, 2021. URL https://api.semanticscholar.org/CorpusID:234861306

    2021

  26. [34]

    Coupling vision and proprioception for navigation of legged robots

    Zipeng Fu, Ashish Kumar, Ananye Agarwal, Haozhi Qi, Jitendra Malik, and Deepak Pathak. Coupling vision and proprioception for navigation of legged robots. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 17273--17283, June 2022

    2022

  27. [35]

    3d-printed soft sensors for adaptive sensing with online and offline tunable stiffness

    Liang He, Nicolas Herzig, Thrishantha Nanayakkara, and Perla Maiolino. 3d-printed soft sensors for adaptive sensing with online and offline tunable stiffness. Soft robotics, 2022. URL https://api.semanticscholar.org/CorpusID:247677615

    2022

  28. [36]

    Théo Le Signor, Nicolas Dupré, and Gael F. Close. A gradiometric magnetic force sensor immune to stray magnetic fields for robotic hands and grippers. IEEE Robotics and Automation Letters, 7 0 (2): 0 3070--3076, 2022. doi:10.1109/LRA.2022.3146507

    2022

  29. [37]

    Pannen, Steffen Puhlmann, and Oliver Brock

    Tessa J. Pannen, Steffen Puhlmann, and Oliver Brock. A low-cost, easy-to-manufacture, flexible, multi-taxel tactile sensor and its application to in-hand object recognition. In 2022 International Conference on Robotics and Automation (ICRA), pages 10939--10944, 2022. doi:10.11...

    2022

  30. [38]

    Gelslim 3.0: High-resolution measurement of shape, force and slip in a compact tactile-sensing finger

    Ian H Taylor, Siyuan Dong, and Alberto Rodriguez. Gelslim 3.0: High-resolution measurement of shape, force and slip in a compact tactile-sensing finger. In 2022 International Conference on Robotics and Automation (ICRA), pages 10781--10787. IEEE, 2022

    2022

  31. [39]

    Yao, Jiaying Mo, Songyue Chen, Yu Xie, Sheng lin Ma, Rui Chen, Tao Luo, Weisong Ling, Lifeng Qin, Zuankai Wang, and Wei Zhou

    Jinhui Zhang, H. Yao, Jiaying Mo, Songyue Chen, Yu Xie, Sheng lin Ma, Rui Chen, Tao Luo, Weisong Ling, Lifeng Qin, Zuankai Wang, and Wei Zhou. Finger-inspired rigid-soft hybrid tactile sensor with superior sensitivity at high frequency. Nature Communications, 13, 2022. URL htt...

    2022

  32. [40]

    https://www.gelsight.com/gelsightmini/, 2023

    Gelsight mini. https://www.gelsight.com/gelsightmini/, 2023

    2023

  33. [41]

    All the feels: A dexterous hand with large-area tactile sensing

    Raunaq Bhirangi, Abigail DeFranco, Jacob Adkins, Carmel Majidi, Abhinav Gupta, Tess Hellebrekers, and Vikash Kumar. All the feels: A dexterous hand with large-area tactile sensing. IEEE Robotics and Automation Letters, 2023

    2023

  34. [42]

    Carbon design engine, 2023

    Carbon3D. Carbon design engine, 2023. URL https://www.carbon3d.com/products/carbon-design-engine. Carbon Design Engine

    2023

  35. [43]

    Liu, Leonardo Zamora Yañez, and Edward H

    Sandra Q. Liu, Leonardo Zamora Yañez, and Edward H. Adelson. Gelsight endoflex: A soft endoskeleton hand with continuous high-resolution tactile sensing. In 2023 IEEE International Conference on Soft Robotics (RoboSoft), pages 1--6, 2023. doi:10.1109/RoboSoft55895.2023.10122053

    2023

  36. [44]

    ntop platform, 2023

    nTopology. ntop platform, 2023. URL https://ntopology.com/. Engineering software built for digital manufacturing

    2023

  37. [45]

    On bringing robots home

    Nur Muhammad Mahi Shafiullah, Anant Rai, Haritheja Etukuru, Yiqian Liu, Ishan Misra, Soumith Chintala, and Lerrel Pinto. On bringing robots home. arXiv preprint arXiv:2311.16098, 2023

    2023 arXiv

  38. [46]

    Learning fine-grained bimanual manipulation with low-cost hardware

    Tony Z Zhao, Vikash Kumar, Sergey Levine, and Chelsea Finn. Learning fine-grained bimanual manipulation with low-cost hardware. arXiv preprint arXiv:2304.13705, 2023

    2023 arXiv

  39. [48]

    Anyskin: Plug-and-play skin sensing for robotic touch

    Raunaq Bhirangi, Venkatesh Pattabiraman, Enes Erciyes, Yifeng Cao, Tess Hellebrekers, and Lerrel Pinto. Anyskin: Plug-and-play skin sensing for robotic touch. arXiv preprint arXiv:2409.08276, 2024 a

    2024 arXiv

  40. [49]

    Hierarchical state space models for continuous sequence-to-sequence modeling

    Raunaq Bhirangi, Chenyu Wang, Venkatesh Pattabiraman, Carmel Majidi, Abhinav Gupta, Tess Hellebrekers, and Lerrel Pinto. Hierarchical state space models for continuous sequence-to-sequence modeling. In Forty-first International Conference on Machine Learning, 2024 b

    2024

  41. [50]

    Diffusion policy: Visuomotor policy learning via action diffusion

    Cheng Chi, Zhenjia Xu, Siyuan Feng, Eric Cousineau, Yilun Du, Benjamin Burchfiel, Russ Tedrake, and Shuran Song. Diffusion policy: Visuomotor policy learning via action diffusion. The International Journal of Robotics Research, 2024 a

    2024

  42. [51]

    Universal manipulation interface: In-the-wild robot teaching without in-the-wild robots

    Cheng Chi, Zhenjia Xu, Chuer Pan, Eric Cousineau, Benjamin Burchfiel, Siyuan Feng, Russ Tedrake, and Shuran Song. Universal manipulation interface: In-the-wild robot teaching without in-the-wild robots. In Proceedings of Robotics: Science and Systems (RSS), 2024 b

    2024

  43. [53]

    Robot utility models: General policies for zero-shot deployment in new environments

    Haritheja Etukuru, Norihito Naka, Zijin Hu, Seungjae Lee, Julian Mehu, Aaron Edsinger, Chris Paxton, Soumith Chintala, Lerrel Pinto, and Nur Muhammad Mahi Shafiullah. Robot utility models: General policies for zero-shot deployment in new environments. arXiv preprint arXiv:2409...

    2024 arXiv

  44. [54]

    BAKU : An efficient transformer for multi-task policy learning

    Siddhant Haldar, Zhuoran Peng, and Lerrel Pinto. BAKU : An efficient transformer for multi-task policy learning. In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. URL https://openreview.net/forum?id=uFXGsiYkkX

    2024

  45. [55]

    3d-vitac: Learning fine-grained manipulation with visuo-tactile sensing

    Binghao Huang, Yixuan Wang, Xinyi Yang, Yiyue Luo, and Yunzhu Li. 3d-vitac: Learning fine-grained manipulation with visuo-tactile sensing. In 8th Annual Conference on Robot Learning, nov 2024

    2024

  46. [56]

    OPEN TEACH : A versatile teleoperation system for robotic manipulation

    Aadhithya Iyer, Zhuoran Peng, Yinlong Dai, Irmak Guzey, Siddhant Haldar, Soumith Chintala, and Lerrel Pinto. OPEN TEACH : A versatile teleoperation system for robotic manipulation. In 8th Annual Conference on Robot Learning, 2024. URL https://openreview.net/forum?id=cvAIaS6V2I

    2024

  47. [57]

    Droid: A large-scale in-the-wild robot manipulation dataset

    Alexander Khazatsky, Karl Pertsch, Suraj Nair, Ashwin Balakrishna, Sudeep Dasari, Siddharth Karamcheti, Soroush Nasiriany, Mohan Kumar Srirama, Lawrence Yunliang Chen, Kirsty Ellis, et al. Droid: A large-scale in-the-wild robot manipulation dataset. In Robotics: Science and Sy...

    2024

  48. [58]

    Open x-embodiment: Robotic learning datasets and rt-x models : Open x-embodiment collaboration0

    Abby O’Neill, Abdul Rehman, Abhiram Maddukuri, Abhishek Gupta, Abhishek Padalkar, Abraham Lee, Acorn Pooley, Agrim Gupta, Ajay Mandlekar, Ajinkya Jain, et al. Open x-embodiment: Robotic learning datasets and rt-x models : Open x-embodiment collaboration0. In 2024 IEEE Internat...

    2024

  49. [59]

    Learning precise, contact-rich manipulation through uncalibrated tactile skins

    Venkatesh Pattabiraman, Yifeng Cao, Siddhant Haldar, Lerrel Pinto, and Raunaq Bhirangi. Learning precise, contact-rich manipulation through uncalibrated tactile skins. arXiv preprint arXiv:2410.17246, 2024

    2024 arXiv

  50. [60]

    Neuralfeels with neural fields: Visuotactile perception for in-hand manipulation

    Sudharshan Suresh, Haozhi Qi, Tingfan Wu, Taosha Fan, Luis Pineda, Mike Lambeta, Jitendra Malik, Mrinal Kalakrishnan, Roberto Calandra, Michael Kaess, Joseph Ortiz, and Mustafa Mukadam. Neuralfeels with neural fields: Visuotactile perception for in-hand manipulation. Science R...

    2024 doi

  51. [61]

    Cut‐cell microstructures for two‐scale structural optimization

    Davi Tozoni, Zizhou Huang, Daniele Panozzo, and Denis Zorin. Cut‐cell microstructures for two‐scale structural optimization. Computer Graphics Forum, 43, 07 2024. doi:10.1111/cgf.15139

    2024 doi

  52. [62]

    Decomposing the generalization gap in imitation learning for visual robotic manipulation

    Annie Xie, Lisa Lee, Ted Xiao, and Chelsea Finn. Decomposing the generalization gap in imitation learning for visual robotic manipulation. In 2024 IEEE International Conference on Robotics and Automation (ICRA), pages 3153--3160. IEEE, 2024

    2024

  53. [63]

    Cushsense: Soft, stretchable, and comfortable tactile-sensing skin for physical human-robot interaction

    Boxin Xu, Luoyan Zhong, Grace Zhang, Xiaoyu Liang, Diego Virtue, Rishabh Madan, and Tapomayukh Bhattacharjee. Cushsense: Soft, stretchable, and comfortable tactile-sensing skin for physical human-robot interaction. 2024 IEEE International Conference on Robotics and Automation ...

    2024

  54. [64]

    Pikul, Mark Yim, and Tess Lee Hellebrekers

    Jessica Yin, Haozhi Qi, Jitendra Malik, James H. Pikul, Mark Yim, and Tess Lee Hellebrekers. Learning in-hand translation using tactile skin with shear and normal force sensing. CoRR, abs/2407.07885, 2024. URL https://doi.org/10.48550/arXiv.2407.07885

    2024 doi arXiv

  55. [65]

    Robot synesthesia: In-hand manipulation with visuotactile sensing

    Ying Yuan, Haichuan Che, Yuzhe Qin, Binghao Huang, Zhao-Heng Yin, Kang-Won Lee, Yi Wu, Soo-Chul Lim, and Xiaolong Wang. Robot synesthesia: In-hand manipulation with visuotactile sensing. In 2024 IEEE International Conference on Robotics and Automation (ICRA), pages 6558--6565....

    2024

  56. [66]

    Transferable tactile transformers for representation learning across diverse sensors and tasks

    Jialiang Zhao, Yuxiang Ma, Lirui Wang, and Edward Adelson. Transferable tactile transformers for representation learning across diverse sensors and tasks. In 8th Annual Conference on Robot Learning, 2024. URL https://openreview.net/forum?id=KXsropnmNI

    2024

  57. [67]

    Data scaling laws in imitation learning for robotic manipulation

    Fanqi Lin, Yingdong Hu, Pingyue Sheng, Chuan Wen, Jiacheng You, and Yang Gao. Data scaling laws in imitation learning for robotic manipulation. In The Thirteenth International Conference on Learning Representations, 2025. URL https://openreview.net/forum?id=pISLZG7ktL

    2025

  58. [68]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

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