pith. sign in

arxiv: 2604.17245 · v1 · submitted 2026-04-19 · 💻 cs.RO

MM-Hand: A 21-DOF Multi-modal Modular Dexterous Robotic Hand with Remote Actuation

Zhuoheng Li , Qingquan Lin , Checheng Yu , Qiangyu Chen , Zhiqian Lan , Lutong Zhang , Hongyang Li , Ping Luo This is my paper

Pith reviewed 2026-05-10 06:16 UTC · model grok-4.3

classification 💻 cs.RO
keywords dexterous robotic handremote tendon actuationtendon-drivenmultimodal sensingmodular design21-DOFfingertip force
0
0 comments X

The pith

Remote tendon actuation packs 21 degrees of freedom and 25N fingertip force into a lightweight dexterous hand with rich sensing.

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

The paper develops a robotic hand that moves its drive motors far away using tendon sheaths instead of placing them inside the hand. This choice frees space for twenty-one total degrees of freedom across the fingers and palm, plus multiple sensors and modular parts that can be swapped quickly. The authors measure how sheath length and friction change with motion, then apply that information to design routing paths and a joint controller that works while the arm moves. Experiments confirm that the fingertip still produces twenty-five newtons of force through a one-meter transmission and that the joints follow commands accurately in both fixed and moving conditions. The complete hardware and software are released so others can use the hand for manipulation research.

Core claim

MM-Hand realizes a 21-DOF multi-modal modular dexterous hand through remote tendon-driven actuation with spring-return fingers, quick tendon connectors, and a sensing system that includes joint angle sensors, tactile sensors, motor-side feedback, and in-palm stereo vision. Analysis of tendon-sheath length variation and friction loss informs the routing, motor hub, and closed-loop control design. Experiments establish that the system transmits twenty-five newtons at the fingertip over one meter and maintains command tracking both with a static arm and during arm motion.

What carries the argument

Remote tendon-sheath actuation with spring-return fingers and quantitative analysis of length variation plus friction loss that permits motors to sit outside the hand.

Load-bearing premise

Friction losses and length changes in the tendon sheaths remain predictable and compensable enough for closed-loop control when the arm moves and bends the routing paths.

What would settle it

A test that records fingertip force falling below twenty newtons or joint tracking errors rising sharply after repeated arm motions that flex the one-meter sheaths through many different curvatures.

Figures

Figures reproduced from arXiv: 2604.17245 by Checheng Yu, Hongyang Li, Lutong Zhang, Ping Luo, Qiangyu Chen, Qingquan Lin, Zhiqian Lan, Zhuoheng Li.

Figure 1
Figure 1. Figure 1: Overview of MM-Hand System. designs, offers short transmission paths and high control accuracy [4], [5], [6], [7], but embedding motors and reducers inside the hand increases mass and inertia of the hand, occupies internal volume, and worsens heat dissipation and impact robustness. Compliant transmission uses non-rigid transmission elements to convey force, including tendon-driven, artificial￾muscle, and f… view at source ↗
Figure 2
Figure 2. Figure 2: Hand structure, sensors, and degree of freedom distribution of MM-Hand. A-A denotes [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 5
Figure 5. Figure 5: Software-based tendon pretensioning for spring-return joints. TABLE I: Joint range of motion of the MM-Hand. Joint Thumb Index Middle Ring Little CMC1 rotation 90◦ / / / / CMC2 A-A 90◦ / / / / CMC3 F-E 90◦ / / / / MCP A-A / 80◦ 100◦ 95◦ 85◦ MCP F-E 85◦ 90◦ 90◦ 90◦ 90◦ PIP F-E / 85◦ 85◦ 85◦ 85◦ DIP F-E / 90◦ 90◦ 90◦ 90◦ IP F-E 90◦ / / / / Integrating over the total bending angle ϕ gives T(L) = T(0) exp(µϕ).… view at source ↗
Figure 7
Figure 7. Figure 7: Setup of Sheath-Tendon Friction Measurement. [PITH_FULL_IMAGE:figures/full_fig_p004_7.png] view at source ↗
Figure 6
Figure 6. Figure 6: System-level electronics architecture of MM-Hand, [PITH_FULL_IMAGE:figures/full_fig_p004_6.png] view at source ↗
Figure 9
Figure 9. Figure 9: Friction test results for the four sheath types. (a) Average [PITH_FULL_IMAGE:figures/full_fig_p005_9.png] view at source ↗
Figure 11
Figure 11. Figure 11: Tracking performance of Joint 0 under a 0.5hz [PITH_FULL_IMAGE:figures/full_fig_p005_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Stereo depth estimation from the in-palm cameras. [PITH_FULL_IMAGE:figures/full_fig_p006_12.png] view at source ↗
read the original abstract

High-DOF dexterous hands require compact actuation, rich sensing, and reliable thermal behavior, but conventional designs often occupy valuable in-hand space, increase end-effector mass, and suffer from heat accumulation near the hand. Remote tendon-driven actuation offers an alternative by relocating motors to the robot base or an external motor hub, thereby freeing the fingers and palm for additional degrees of freedom, sensing modules, and maintainable mechanical structures. This paper presents MM-Hand, a 21-DOF Multimodal Modular dexterous hand based on remote tendon-driven actuation. The hand integrates spring-return tendon-driven fingers, modular 3D-printed finger and palm structures, quick tendon connectors for maintenance, and a multimodal sensing system including joint angle sensors, tactile sensors, motor-side feedback, and in-palm stereo vision. We further analyze tendon-sheath length variation and friction loss to guide the design of the routing, motor hub, and closed-loop joint control. Experiments validate the transmission, output force, sensing, and control capability of the system. The fingertip force reaches 25N under a 1m remote sheath transmission, demonstrating practical load capacity despite long-distance tendon routing. Closed-loop joint-level experiments further evaluate command tracking with a static arm and during arm motion. These results show that MM-Hand provides a lightweight, sensor-rich, and maintainable hardware platform for dexterous manipulation research. To support the community, all hardware designs and software frameworks are made fully open-source at https://mmlab.hk/research/MM-Hand.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript presents MM-Hand, a 21-DOF multi-modal modular dexterous robotic hand using remote tendon-driven actuation with 1 m sheaths to relocate motors away from the hand. It describes spring-return fingers, modular 3D-printed structures, quick tendon connectors, and a multimodal sensing suite (joint angles, tactile, motor feedback, in-palm stereo vision). The authors analyze tendon-sheath length variation and friction loss to guide routing, motor hub, and closed-loop control design, then report experiments claiming 25 N fingertip force and successful joint-level command tracking under both static and moving-arm conditions. All hardware designs and software are released open-source.

Significance. If the performance claims are substantiated with quantitative bounds, the work supplies a lightweight, maintainable, sensor-rich open-source platform that directly addresses mass, heat, and space limitations of conventional in-hand actuation for dexterous manipulation research.

major comments (2)
  1. [Abstract and Experiments] Abstract and Experiments section: the claim of reliable closed-loop joint control during arm motion rests on the assertion that length variation and friction remain manageable, yet no measured transmission efficiency, friction loss percentage as a function of sheath curvature or velocity, or position-error statistics attributable to length change are provided; without these data the predictability of the controller outside the specific test trajectories cannot be verified.
  2. [Experiments] Experiments section: the reported 25 N fingertip force under 1 m remote transmission lacks accompanying details on measurement protocol, number of trials, error bars, or comparison to direct-drive baselines, which is load-bearing for the central claim of practical load capacity.
minor comments (1)
  1. [Abstract] The abstract states that closed-loop tracking was tested with both static and moving arm but does not specify the number of trials, trajectory types, or quantitative tracking metrics (e.g., RMSE values).

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback on our manuscript. We address each major comment below and have incorporated revisions to strengthen the presentation of our experimental results and analysis.

read point-by-point responses

  1. Referee: [Abstract and Experiments] Abstract and Experiments section: the claim of reliable closed-loop joint control during arm motion rests on the assertion that length variation and friction remain manageable, yet no measured transmission efficiency, friction loss percentage as a function of sheath curvature or velocity, or position-error statistics attributable to length change are provided; without these data the predictability of the controller outside the specific test trajectories cannot be verified.

    Authors: We appreciate this observation. Our analysis of tendon-sheath length variation and friction loss was used to inform the routing design and controller parameters, and the closed-loop experiments demonstrated reliable tracking under both static and dynamic arm conditions. However, we agree that explicit quantitative measurements of transmission efficiency versus curvature/velocity and associated position-error statistics would better substantiate generalizability. In the revised manuscript we will add these data from additional characterization experiments, including efficiency curves and error statistics broken down by trajectory type. revision: yes

  2. Referee: [Experiments] Experiments section: the reported 25 N fingertip force under 1 m remote transmission lacks accompanying details on measurement protocol, number of trials, error bars, or comparison to direct-drive baselines, which is load-bearing for the central claim of practical load capacity.

    Authors: We agree that the force results require more supporting detail to be fully convincing. The 25 N value was obtained with a calibrated load cell at the fingertip under quasi-static conditions with the 1 m sheath routing; the value represents the mean across repeated trials. In the revised Experiments section we will include the complete measurement protocol, number of trials, standard deviations, and a direct comparison against an equivalent direct-drive configuration to quantify the transmission penalty. revision: yes

Circularity Check

0 steps flagged

No circularity: hardware description and experimental reporting with independent validation.

full rationale

The manuscript is a design-and-experiment paper. It describes a 21-DOF tendon-driven hand, states that tendon-sheath length variation and friction were analyzed to inform routing and control choices, and reports measured fingertip force (25 N) plus closed-loop tracking results under static and moving-arm conditions. No equations, fitted parameters, or predictions are presented that reduce by construction to the inputs; the force and tracking claims rest on direct experimental measurement rather than any self-referential derivation or self-citation chain. The analysis of friction and length variation is used only to guide design decisions and is not invoked as a uniqueness theorem or load-bearing premise that collapses into prior self-work.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Engineering design paper with no mathematical free parameters, axioms, or invented physical entities; design decisions such as sheath routing are guided by analysis of length variation and friction but are not fitted constants or new postulates.

pith-pipeline@v0.9.0 · 5604 in / 1258 out tokens · 45120 ms · 2026-05-10T06:16:27.018224+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

  1. [1]

    L. A. Jones and S. J. Lederman,Human Hand Function. New York: Oxford University Press, 2006. [Online]. Available: https: //doi.org/10.1093/acprof:oso/9780195173154.001.0001

    work page doi:10.1093/acprof:oso/9780195173154.001.0001 2006

  2. [2]

    Learning dexterous in-hand manipulation,

    OpenAI, I. Akkaya, M. Andrychowicz, M. Chociej, M. Litwin, B. McGrew, A. Petron, A. Paino, M. Plappert, G. Powellet al., “Learning dexterous in-hand manipulation,”The International Journal of Robotics Research, vol. 39, no. 1, pp. 3–20, 2020. [Online]. Available: https://doi.org/10.1177/0278364919887447

    work page doi:10.1177/0278364919887447 2020

  3. [3]

    Trends and challenges in robot manipulation,

    A. Billard and D. Kragic, “Trends and challenges in robot manipulation,” Science, vol. 364, no. 6446, p. eaat8414, 2019. [Online]. Available: https://doi.org/10.1126/science.aat8414

    work page doi:10.1126/science.aat8414 2019

  4. [4]

    Kitech-hand: A highly dexterous and modularized robotic hand,

    D.-H. Lee, J.-H. Park, S.-W. Park, M.-H. Baeg, and J.-H. Bae, “Kitech-hand: A highly dexterous and modularized robotic hand,” IEEE/ASME Transactions on Mechatronics, vol. 22, no. 2, pp. 876–887,

    work page

  5. [5]

    Available: https://doi.org/10.1109/TMECH.2016.2634602

    [Online]. Available: https://doi.org/10.1109/TMECH.2016.2634602

    work page doi:10.1109/tmech.2016.2634602 2016

  6. [6]

    LEAP Hand: Low- cost, efficient, and anthropomorphic hand for robot learning,

    K. Shaw, A. Agarwal, and D. Pathak, “LEAP Hand: Low- cost, efficient, and anthropomorphic hand for robot learning,” in Robotics: Science and Systems (RSS), 2023. [Online]. Available: https://doi.org/10.15607/rss.2023.xix.089

    work page doi:10.15607/rss.2023.xix.089 2023

  7. [7]

    [Online]

    Sharpa, “Wave,” 2025. [Online]. Available: https://www.sharpa.com/ pages/wave

    work page 2025

  8. [8]

    [Online]

    PsiBot, “Psi-h1,” 2025. [Online]. Available: https://www.psibot.ai/en/ products/product_psibot-h1/

    work page 2025

  9. [9]

    Tendon driven robotic hands: A review,

    E. Nazma and S. Mohd, “Tendon driven robotic hands: A review,” International Journal of Mechanical Engineering and Robotics Research, vol. 1, no. 3, pp. 350–357, October 2012. [Online]. Available: https://www.ijmerr.com/uploadfile/2015/0409/20150409024829230.pdf

    work page 2012

  10. [10]

    Soft robotic grippers,

    J. Shintake, V . Cacucciolo, D. Floreano, and H. Shea, “Soft robotic grippers,”Advanced Materials, vol. 30, no. 29, p. 1707035, 2018. [Online]. Available: https://doi.org/10.1002/adma.201707035

    work page doi:10.1002/adma.201707035 2018

  11. [11]

    A dexterous and compliant (DexCo) hand based on soft hydraulic actuation for human-inspired fine in-hand manipulation,

    J. Zhou, J. Huang, Q. Dou, P. Abbeel, and Y . Liu, “A dexterous and compliant (DexCo) hand based on soft hydraulic actuation for human-inspired fine in-hand manipulation,”IEEE Transactions on Robotics, vol. 41, pp. 666–686, 2025. [Online]. Available: https://doi.org/10.1109/TRO.2024.3508932

    work page doi:10.1109/tro.2024.3508932 2025

  12. [12]

    Sanctuary ai technology,

    Sanctuary AI, “Sanctuary ai technology,” 2025. [Online]. Available: https://www.sanctuary.ai/technology

    work page 2025

  13. [13]

    Designing Anthropomorphic Soft Hands Through Interaction

    Y . Toshimitsu, B. Forrai, B. G. Cangan, U. Steger, M. Knecht, S. Weirich, and R. K. Katzschmann, “Getting the ball rolling: Learning a dexterous policy for a biomimetic tendon-driven hand with rolling contact joints,” in2023 IEEE-RAS 22nd International Conference on Humanoid Robots (Humanoids), 2023, pp. 1–7. [Online]. Available: https://doi.org/10.1109/...

    work page doi:10.1109/humanoids57100.2023.10375231 2023

  14. [14]

    Development of the bioinspired tendon-driven dexhand 021 with proprioceptive compliance control,

    J. Yuan, H. Zhu, J. Dai, and S. Yi, “Development of the bioinspired tendon-driven dexhand 021 with proprioceptive compliance control,”arXiv preprint arXiv:2511.03481, 2025. [Online]. Available: https://arxiv.org/abs/2511.03481

    work page arXiv 2025

  15. [15]

    Christoph, Maximilian Eberlein, Filippos Katsimalis, Arturo Roberti, Aristotelis Sympetheros, Michel R

    C. C. Christoph, M. Eberlein, F. Katsimalis, A. Roberti, A. Sympetheros, M. R. V ogt, D. Liconti, C. Yang, B. G. Cangan, R. J. Hinchet, and R. K. Katzschmann, “Orca: An open-source, reliable, cost- effective, anthropomorphic robotic hand for uninterrupted dexterous task learning,”arXiv preprint arXiv:2504.04259, 2025. [Online]. Available: https://arxiv.or...

    work page arXiv 2025

  16. [16]

    RUKA: Rethinking the design of humanoid hands with learning,

    A. Zorin, I. Guzey, B. Yan, A. Iyer, L. Kondrich, N. X. Bhattasali, and L. Pinto, “RUKA: Rethinking the design of humanoid hands with learning,”arXiv preprint arXiv:2504.13165, 2025. [Online]. Available: https://arxiv.org/abs/2504.13165

    work page arXiv 2025

  17. [17]

    Pantheon hand,

    Stella Robot, “Pantheon hand,” 2025. [Online]. Available: https: //www.stella-robot.com/en/h-col-108.html

    work page 2025

  18. [18]

    Antagonistic bowden-cable actuation of a lightweight robotic hand: Toward dexterous manipulation for payload constrained humanoids,

    S. Min, H. Kim, and D. H. Shim, “Antagonistic bowden-cable actuation of a lightweight robotic hand: Toward dexterous manipulation for payload constrained humanoids,”arXiv preprint arXiv:2512.24657, 2025. [Online]. Available: https://arxiv.org/abs/2512.24657

    work page arXiv 2025

  19. [19]

    Development of low-inertia high-stiffness manipulator LIMS2 for high-speed manipulation of foldable objects,

    H. Song, Y .-S. Kim, J. Yoon, S.-H. Yun, J. Seo, and Y .- J. Kim, “Development of low-inertia high-stiffness manipulator LIMS2 for high-speed manipulation of foldable objects,” in2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2018, pp. 4145–4151. [Online]. Available: https://doi.org/10.1109/IROS.2018.8594005

    work page doi:10.1109/iros.2018.8594005 2018

  20. [20]

    Astribot S1,

    Astribot, “Astribot S1,” 2025. [Online]. Available: https://www.astribot. com/en/product/

    work page 2025

  21. [21]

    NEO Home Robot,

    1X, “NEO Home Robot,” 2025. [Online]. Available: https://www.1x. tech/neo

    work page 2025

  22. [22]

    The mechanics of friction in rope rescue,

    S. W. Attaway, “The mechanics of friction in rope rescue,” in Proceedings of the International Technical Rescue Symposium (ITRS ’99), Fort Collins, CO, USA, 1999. [Online]. Available: https://web.mit. edu/sp255/www/reference_vault/Friction%20Large%20Format.pdf

    work page 1999

  23. [23]

    Lipson, Z

    L. Lipson, Z. Teed, and J. Deng, “Raft-stereo: Multilevel recurrent field transforms for stereo matching,” inInternational Conference on 3D Vision (3DV), 2021. [Online]. Available: https://doi.org/10.1109/ 3dv53792.2021.00032

    work page arXiv 2021