Morphing MILR: Design and control of a cable-driven limbless robot with rolling joints for maneuvering in complex environments

Daniel I. Goldman; Donoven Dortilus; Galen Tunnicliffe; Matthew Fernandez; Tianyu Wang

REVIEW 3 major objections 5 minor 43 references

Morphing MILR: Design and control of a cable-driven limbless robot with rolling joints for maneuvering in complex environments

T0 review · 3 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read Morphing MILR claims that a single cable-driven limbless robot can switch among lateral undulation, sidewinding, rolling, and twisting by reorienting its bending planes through rolling joints, without giving up the passive body compliance t

desk verdict A working reconfigurable cable-driven limbless robot that runs four gaits and switches among them, but the 'preserves mechanical intelligence' claim is asserted rather than measured, and the sidewinding parameters are internally inconsistent. read the letter →

arxiv 2607.19714 v1 pith:VT3OPV2Q submitted 2026-07-22 cs.RO

Donoven Dortilus , Tianyu Wang , Galen Tunnicliffe , Matthew Fernandez , Daniel I. Goldman This is my paper · ORCID

classification cs.RO

keywords limblessrobotcable-drivenactuationprogrammablecompliancerollingjointssidewindinglateralundulationmorphologicalreconfigurationmechanicalintelligence

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 tries to establish that one reconfigurable cable-driven robot body can produce multiple limbless locomotion modes—lateral undulation, sidewinding, rolling, and twisting—while preserving the 'mechanical intelligence' of passive compliance. Previous compliant cable-driven limbless robots were each locked into a single morphology or bending plane, so they could only handle one kind of terrain. The authors add rolling base joints between modules that actively reorient the plane in which each bending joint moves, effectively remapping the body wave in three dimensions. They show reliable gait generation on open ground, repeated progress through a lattice obstacle field using compliance, and 13-of-15 successful transitions from obstacle-rich to open terrain. If the claim holds, a single inexpensive, low-sensing platform could handle heterogeneous environments that previously required specialized robots or terrain-aware controllers.

What carries the argument

The key object is the rolling base joint: a compact geared transmission—worm, helical, and spur stages with a 1:13 reduction and a non-backdrivable worm gear—mounted between modules so it can rotate the bending plane of the adjacent cable-actuated joint. This joint is what lets the robot remap its body-wave direction and lock the new configuration without continuous power. The second supporting mechanism is the bilateral cable actuation with the generalized compliance variable G, which sets whether a joint is bidirectionally compliant, directionally compliant, or stiff, and is used to keep the body adaptable during obstacle contact and gait transitions.

What would settle it

Run the same lattice course with the rolling joints locked flat (undulator mode) and compare progress, jamming rate, and power draw against a purpose-built single-morphology undulator of comparable mass and length; if Morphing MILR jams more often or moves slower at identical gait parameters, the claim that compliance robustness survives the added rolling-joint complexity would be undercut.

Watch

Extended reading notes

Core claim

The central claim is that actively reorienting the bending planes of a cable-driven limbless robot via rolling joints does not sacrifice the robustness that programmable passive compliance provides. Instead, by commanding a fixed pattern of rolling-joint angles—all zero for undulation, alternating orthogonal for sidewinding and rolling—the same modules synthesize traveling body waves in different planes. A pure twisting gait is generated by rolling joints alone with all bending joints held straight. The mechanism for preserving compliance is the generalized compliance variable G from the bilateral cable actuation, which makes each joint bidirectionally stiff, directionally compliant, or full

Load-bearing premise

The claim rests on the assumption that the robustness benefits of programmable passive compliance from prior cable-driven limbless robots are preserved after adding rolling joints, gears, and extra mass; the paper provides supportive but not baseline-matched evidence.

Editorial extensions

If this is right

A single limbless platform can traverse heterogeneous terrain—such as moving from a dense lattice into open ground—by switching gaits mid-run without relying on terrain mapping or high-gain feedback.
Because the rolling joints lock mechanically, the robot can hold sidewinding or rolling configurations with zero continuous motor power, reducing energy cost during long traversals.
The modular control structure, which separates morphology reconfiguration from wave generation, can be extended to new three-dimensional motion patterns beyond the four demonstrated gaits.
The demonstrated transition strategy—return to a straight home pose, reorient rolling bases, then resume undulation—provides a simple open-loop recipe for multi-mode locomotion in confined spaces.
The platform offers a testbed for studying how body morphology changes interact with body-terrain dynamics, since the same body can be switched between planar and three-dimensional configurations.

Reading between the lines

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

The same rolling-joint mechanism could support gaits the paper does not test, such as helical rolling or climbing, by commanding non-orthogonal or time-varying rolling-angle sequences; the paper's 'free form' configuration suggests the hardware already allows this.
If compliance benefits are genuinely preserved despite the added mass and friction of gears and rolling joints, the design philosophy could transfer to untethered, softer-bodied robots; the tether and external computer currently limit field deployment.
A matched head-to-head test against purpose-built single-gait robots from earlier work would show whether unification comes with an efficiency cost in cost of transport or body-lengths-per-cycle—the authors explicitly defer this comparison.
The tail jamming observed in transition trials suggests the bottleneck is not the individual gaits but the reconfiguration maneuver itself, so compliance shaping specifically during the transition phase may be the most direct avenue for improvement.

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.

Desk Editor's Note

read the letter

Short version: this is a solid hardware demonstration of a reconfigurable cable-driven limbless robot that runs four gaits and switches among them, but the central claim that it preserves mechanical intelligence is asserted rather than measured, and the sidewinding numbers are internally inconsistent. I'd send it out, but it needs revision.

The genuinely new piece is hardware: bilateral cable actuation plus actively reorientable rolling joints with a worm gear that holds configuration without power. Putting four gaits—undulation, sidewinding, rolling, twisting—on one compliant platform and showing transitions between them is a real step beyond the fixed-morphology MILR line. The design details (slip rings, gear reduction, compliance modulation during transitions) are useful. The paper is honest about limits, and the conclusion explicitly defers quantitative comparison to future work.

Soft spots, in order. First, the sidewinding parameter inconsistency is real and needs fixing: Section III.B.2 says AH should be lower than AV, but IV.A reports AH=60° and AV=30°; Fig.10 lists AV=60°, AH=30°. The two experimental sections can't both be describing the same command, and a reader cannot reproduce the gait as reported. This is a reporting error, not necessarily a strike against the platform, but it has to be resolved.

Second, the preservation claim. The paper says the platform preserves mechanically intelligent, compliance-based robustness, but the evidence is 20 lattice trials and a 13/15 transition success count with no baseline on the original MILR. Since the whole point is doing what [34],[35] do plus morphology, you need paired trials or at least a stated performance bound. The authors acknowledge this; it's a missing experiment, not a false claim.

Third, the trial evidence is thin and qualitative—five pass/fail trials per gait, video-based. That's acceptable for a platform demonstration but gives no error bars, no speed, no cost of transport. If the paper is to be more than a demo, it should report at least body lengths per cycle for each gait.

Overall, the central argument holds up at the level it is made: the robot does what it says, the design is reproducible, and the gaits are sensible. The incompleteness is in the evaluation, not the mechanism. I'd send it to peer review, expecting a major-revision recommendation focusing on the parameter inconsistency and a baseline or explicit scoping.

For whom: robotics conference or journal audience interested in limbless locomotion, compliant actuation, and morphology reconfiguration. Worth citing as a platform paper.

Referee Report

3 major / 5 minor

Summary. The paper presents Morphing MILR, a six-module cable-driven limbless robot in which each module combines antagonistic cable-driven bending with a motorized rolling base that reorients the bending plane. The authors claim that this architecture unifies lateral undulation, sidewinding, rolling, and twisting in a single platform while retaining the programmable passive compliance that enabled robust, sensing-light locomotion in prior MILR designs. They report obstacle-free gait trials, lattice experiments with varying compliance parameter G, and 15 mixed-environment transition trials (13 successes), concluding that morphing does not sacrifice the mechanically intelligent benefits of compliance.

Significance. If the central claim holds, the platform is a meaningful advance over gait-specialized compliant robots: it would allow a single hardware morphology to switch locomotion modes in cluttered terrain without terrain sensing or high-bandwidth feedback, with clear applications in search, rescue, and inspection. Strengths include the detailed mechanical design, the simple open-loop gait templates, and the explicit use of the prior MILR framework for compliance control. However, the paper's own evidence does not yet establish the key preservation claim, and one reported gait parameter set is internally inconsistent. The result is promising but needs additional quantitative support before the abstract's claim is fully substantiated.

major comments (3)

[IV.B, IV.C, and Conclusion] The central claim that Morphing MILR 'preserved the mechanically intelligent benefits of bilateral cable actuation' (IV.B) and 'without losing the mechanically intelligent benefits of compliance' (IV.C) is not supported by any matched comparison against the original MILR in [34] or the sidewinding platform in [35]. The lattice test defines success only as reaching the end, and the transition trial reports 13/15 successes; no speed, cost of transport, body lengths per cycle, or failure-mode data are provided for either platform. The added 0.25 kg/module, gear friction, and non-backdrivable worm gear (II.B) could plausibly alter the body-terrain interaction that underlies the claimed robustness. The Conclusion explicitly defers this comparison ('future work will gather more quantitative data to compare...'), making the preservation claim currently unverified. A matched-condition baseline (
[III.B.2, IV.A, Fig. 10] The sidewinding parameters are internally inconsistent. Section III.B.2 states that 'A_H [is] kept lower than A_V', but Section IV.A reports 'sidewinding parameters utilized were A_V = 30°, A_H = 60°', the opposite inequality. Figure 10 lists 'AV = 60, AH = 30', which is consistent with III.B.2 but contradicts IV.A. Since Eq. (6) assigns the two amplitudes to odd/even joints, swapping the values changes the waveform and the resulting contact pattern. This makes the sidewinding demonstration non-reproducible as reported and must be corrected.
[IV.B and IV.C] The experimental reporting is too coarse to support the reliability claims. The lattice section says 'varying G values over 20 trials' and 'majority of successful runs occurred with G=1', but it does not state how many trials were run at each G value, the actual G values tested, or the success count per condition. Similarly, the environment-transition section reports 13/15 successes with tail jamming as the failure cause but provides no per-trial detail, no confidence intervals, and no statistical treatment. Given that the paper's contribution is robustness through morphology and compliance, raw per-condition data and effect sizes (or at least a table of all trials) are needed for the 15 trials and the 20 lattice runs.

minor comments (5)

[Fig. 10] Typo: 'sidwinding' should be 'sidewinding'.
[Eq. (9)] The symbol ω is used both as a scalar angular velocity and as a vector in the recurrence Ω_i = ω e_i + Ω_(i−1). Please distinguish scalar and vector notation, and define Ω_0.
[III.B.1] The definition of β_i as 'relative to the previous module' appears only in the lateral undulation subsection; it would help to state this once for all gait definitions, since it matters for interpreting Eq. (5).
[Table I] Minor formatting: 'Power' row should read '12 V, 1 A' with a consistent space; the 'Communication' and 'Sensing' entries could use units or a brief clarification.
[II.B] The 'overall gear reduction' is reported as 1:13 with a safety factor, but the number of stages and individual ratios are not given. Adding a small table or schematic of the transmission would aid reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the cited prior-work equations and standard gait templates are used as stated, and no fitted quantity is presented as a prediction.

full rationale

Morphing MILR's contribution is hardware integration and gait demonstrations, not a new first-principles derivation. The cable-length and compliance equations (Eqs. 1 and 2) are explicitly attributed to the authors' prior work [34] rather than re-derived or disguised as new predictions; [34] is a peer-reviewed, externally falsifiable experimental study, so citing it is legitimate support rather than circular self-citation. The serpenoid and orthogonal-wave gaits (Eqs. 4, 6, 7) are standard templates from the limbless-locomotion literature, and the gait parameters are hand-tuned experimental choices, not fitted quantities later relabeled as predictions. The lattice and transition trials report empirical success counts (20 trials and 13/15, respectively) against defined success criteria; whether these results adequately prove that mechanical intelligence is 'preserved' relative to [34]/[35] is a baseline/completeness concern, not a circularity. The internal inconsistency between the sidewinding amplitudes stated in IV.A (AV=30, AH=60) and the guidance in III.B.2 and Fig.10 (AH lower than AV; AV=60, AH=30) is a reproducibility/correctness defect, not a circular derivation. No step in the paper reduces a claimed result to its own inputs by construction.

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

The paper's contribution is integration, so the ledger is dominated by inherited models and hand-tuned gait parameters. No new physical entities are introduced. The empirical weight rests on small numbers of pass/fail trials and designer-selected G values, which is the most honest measure of what the paper adds beyond prior mechanically intelligent limbless robots.

free parameters (7)

Undulation/rolling wave amplitude A = 60° (rolling); 70° (undulator as listed in Fig. 10)
Hand-chosen curvature amplitude used in all experiments; not derived from terrain or a dynamics model.
Sidewinding amplitudes AH and AV = Reported inconsistently: IV.A says AV=30°, AH=60°; III.B.2 and Fig. 10 say AV=60°, AH=30°
Manual gait parameters; the inconsistency makes the executed trial settings ambiguous.
Spatial frequency ξ = 1.1–1.2 cycles per body
Chosen per gait; controls the number of body-wave periods and is not optimized or derived.
Temporal frequency ω = 2 Hz (lattice/undulation), 4 Hz (sidewinding/rolling)
Hand-set wave propagation speed; not derived from any mechanical or terrain model.
Compliance variable G = 1 for lattice and transitions; 0 for obstacle-free tests
Experimenter-selected. The paper reports that most lattice successes occur at G=1, so this value functions as a tuned control setting rather than an independent prediction.
Rolling base gear ratio = 1:13
Chosen from the maximum torque needed to lift two adjacent modules plus a safety factor; a design parameter rather than a fitted value.
Cable slack parameter l0 = Not specified (inherited from [34])
Taken from the prior MILR paper without re-derivation or re-measurement for this platform.

assumptions (6)

domain assumption The serpenoid traveling-wave template (Hirose 1993) produces effective lateral undulation for this robot.
Used in Eq. (4) without re-validation against the new platform's rolling joints.
domain assumption Sidewinding is well approximated by the superposition of two orthogonal body waves (Astley et al. 2015; Marvi et al. 2014).
Used in Eqs. (5)-(6); no dynamics model is provided for the new platform.
domain assumption The cable-length kinematic model in Eqs. (1)-(2) from [34] remains valid for this platform's bending joints.
The modules are dimensionally similar to [34], but the added rolling joints alter structural stiffness; the model is imported as-is.
domain assumption Rolling-base orientations can be treated as static during each gait, decoupling roll dynamics from bending-wave dynamics.
Section III.A states the rolling joints are mainly static during normal operation; this is an implicit quasi-static assumption supported only by the pass/fail trials.
domain assumption The worm-gear transmission is non-backdrivable and holds configuration without continuous power.
Section II.B asserts this; no endurance or load-holding data are provided.
domain assumption Open-loop joint commands plus passive compliance are sufficient for effective locomotion in the tested terrains without terrain sensing.
This is the core of the mechanical-intelligence approach inherited from [34]; it is tested in the lattice and transition trials but not modeled.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Morphing MILR: Design and control of a cable-driven limbless robot with rolling joints for maneuvering in complex environments." pith.science (2026). https://pith.science/paper/VT3OPV2Q

@misc{pith2026260719714,
  author       = {Pith},
  title        = {Pith review of: Morphing MILR: Design and control of a cable-driven limbless robot with rolling joints for maneuvering in complex environments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VT3OPV2Q}},
  note         = {Machine review of arXiv:2607.19714}
}

read the original abstract

Limbless robots offer exceptional mobility in confined and cluttered environments due to their slender bodies and their ability to exploit body-terrain interactions. Recent designs incorporating compliance demonstrate robust locomotion without complex sensing or control; however, these systems typically rely on fixed body configurations, with each morphology specialized for a single locomotion mode or environment. This raises a key challenge: how can a single limbless robot achieve versatile locomotion while preserving the robustness of compliance-mediated locomotion? To address this challenge, we present a cable-driven limbless robot that reconfigures body morphology and compliance to enable diverse locomotion modes. Distributed cable actuation generates traveling body waves, while programmable passive compliance enables robust contact-rich locomotion without terrain knowledge or high-bandwidth feedback. Rolling joints reorient bending planes along the body, enabling rapid reconfiguration and smooth transitions between locomotion styles, and incorporate geared locking to maintain configuration without continuous power. By combining programmable bending compliance and morphology control, the platform achieves lateral undulation, sidewinding, rolling, and twisting within a single system. Experiments demonstrate reliable gait generation, traversal in obstacle-rich environments, and transitions between modes, establishing a versatile limbless platform for navigating complex environments with applications in search and rescue, environmental monitoring, and inspection.

Figures

Figures reproduced from arXiv: 2607.19714 by the authors.

**Figure 2.** Overview of full design (A) broken down into bilateral cable [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 5.** Transitioning from lateral undulation to sidewinding via [PITH_FULL_IMAGE:figures/full_fig_p003_5.png] view at source ↗

Figures from the paper (3 more)

**Figure 7.** Figure 7: Sidewinding, orthogonal serpenoid waves at differing am [PITH_FULL_IMAGE:figures/full_fig_p004_7.png]

**Figure 9.** Figure 9: Twisting, rotating modules at different angular velocities to [PITH_FULL_IMAGE:figures/full_fig_p005_9.png]

**Figure 10.** Figure 10: Morphing MILR transitions from obstacle rich lattice locomotion via compliant lateral undulation to open terrain sidewinding, [PITH_FULL_IMAGE:figures/full_fig_p006_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 1 linked inside Pith

[34]

Mechanical intelligence simplifies control in terrestrial limbless locomotion,

T. Wang, C. Pierce, V . Kojouharov, B. Chong, K. Diaz, H. Lu, and D. I. Goldman, “Mechanical intelligence simplifies control in terrestrial limbless locomotion,”Science Robotics, vol. 8, no. 85, p. eadi2243, 2023

2023
[35]

Anisotropic body compliance facilitates robotic sidewinding in complex environments,

V . Kojouharov, T. Wang, M. Fernandez, J. Maeng, and D. I. Gold- man, “Anisotropic body compliance facilitates robotic sidewinding in complex environments,” in2024 International Conference on Robotics and Automation (ICRA). IEEE, 2024

2024
[1]

Biologically inspired robots,

S. Hirose, “Biologically inspired robots,”Snake-Like Locomotors and Manipulators, 1993

1993
[2]

Biologically inspired snake-like robots,

S. Hirose and M. Mori, “Biologically inspired snake-like robots,” in 2004 IEEE International Conference on Robotics and Biomimetics. IEEE, 2004, pp. 1–7

2004
[3]

A review on modelling, implementation, and control of snake robots,

P. Liljeb ¨ack, K. Y . Pettersen, Ø. Stavdahl, and J. T. Gravdahl, “A review on modelling, implementation, and control of snake robots,” Robotics and Autonomous Systems, vol. 60, no. 1, pp. 29–40, 2012

2012
[4]

Design of a modular snake robot,

C. Wright, A. Johnson, A. Peck, Z. McCord, A. Naaktgeboren, P. Gi- anfortoni, M. Gonzalez-Rivero, R. Hatton, and H. Choset, “Design of a modular snake robot,” in2007 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 2007, pp. 2609–2614

2007
[5]

Design and architecture of the unified modular snake robot,

C. G. Wright, A. D. Buchan, B. Brown, J. C. Geist, M. Schwerin, D. Rollinson, M. Tesch, and H. Choset, “Design and architecture of the unified modular snake robot,” in2012 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2012, pp. 4347–4354

2012
[6]

Design and architecture of a series elastic snake robot,

D. Rollinson, Y . Bilgen, B. Brown, F. Enner, S. Ford, C. Layton, J. Rembisz, M. Schwerin, A. Willig, P. Velagapudi, and H. Choset, “Design and architecture of a series elastic snake robot,” in2014 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2014, pp. 4630–4636

2014

Show all 43 references

[7]

Online optimization of swimming and crawling in an amphibious snake robot,

A. Crespi and A. J. Ijspeert, “Online optimization of swimming and crawling in an amphibious snake robot,”IEEE Transactions on Robotics, vol. 24, no. 1, pp. 75–87, 2008

2008
[8]

AmphiBot I: an amphibious snake-like robot,

A. Crespi, A. Badertscher, A. Guignard, and A. J. Ijspeert, “AmphiBot I: an amphibious snake-like robot,”Robotics and Autonomous Systems, vol. 50, no. 4, pp. 163–175, 2005

2005
[9]

Snake robot obstacle-aided locomotion: Modeling, simula- tions, and experiments,

A. A. Transeth, R. I. Leine, C. Glocker, K. Y . Pettersen, and P. Lil- jeb¨ack, “Snake robot obstacle-aided locomotion: Modeling, simula- tions, and experiments,”IEEE Transactions on Robotics, vol. 24, no. 1, pp. 88–104, 2008

2008
[10]

Parameterized and scripted gaits for modular snake robots,

M. Tesch, K. Lipkin, I. Brown, R. Hatton, A. Peck, J. Rembisz, and H. Choset, “Parameterized and scripted gaits for modular snake robots,”Advanced Robotics, vol. 23, no. 9, pp. 1131–1158, 2009

2009
[11]

Shape-based compliance in locomotion,

M. Travers, J. Whitman, P. Schiebel, D. I. Goldman, and H. Choset, “Shape-based compliance in locomotion,” inRobotics: Science and Systems (RSS), 2016

2016
[12]

Modeling, analysis, and synthesis of serpentine locomotion with a multilink robotic snake,

M. Saito, M. Fukaya, and T. Iwasaki, “Modeling, analysis, and synthesis of serpentine locomotion with a multilink robotic snake,” IEEE Control Systems Magazine, vol. 22, no. 1, pp. 64–81, 2002

2002
[13]

Kinematic gait synthesis for snake robots,

C. Gong, M. J. Travers, H. C. Astley, L. Li, J. R. Mendelson, D. I. Goldman, and H. Choset, “Kinematic gait synthesis for snake robots,” The International Journal of Robotics Research, vol. 35, no. 1-3, pp. 100–113, 2016

2016
[14]

Obstacles are beneficial to me! scaffold- based locomotion of a snake-like robot using decentralized control,

T. Kano and A. Ishiguro, “Obstacles are beneficial to me! scaffold- based locomotion of a snake-like robot using decentralized control,” in2013 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2013, pp. 3273–3278

2013
[15]

Perception-driven obstacle-aided locomotion for snake robots: the state of the art, challenges and possibilities,

F. Sanfilippo, J. Azpiazu, G. Marafioti, A. A. Transeth, Ø. Stavdahl, and P. Liljeb ¨ack, “Perception-driven obstacle-aided locomotion for snake robots: the state of the art, challenges and possibilities,”Applied Sciences, vol. 7, no. 4, p. 336, 2017

2017
[16]

A’sidewinding’locomotion gait for hyper-redundant robots,

J. W. Burdick, J. Radford, and G. S. Chirikjian, “A’sidewinding’locomotion gait for hyper-redundant robots,” in
[17]

Sidewinding with minimal slip: Snake and robot ascent of sandy slopes,

H. Marvi, C. Gong, N. Gravish, H. Astley, M. Travers, R. L. Hatton, J. R. Mendelson III, H. Choset, D. L. Hu, and D. I. Goldman, “Sidewinding with minimal slip: Snake and robot ascent of sandy slopes,”Science, vol. 346, no. 6206, pp. 224–229, 2014

2014
[18]

Modulation of orthogonal body waves enables high maneuverability in sidewinding locomotion,

H. C. Astley, C. Gong, J. Dai, M. Travers, M. M. Serrano, P. A. Vela, H. Choset, J. R. Mendelson III, D. L. Hu, and D. I. Goldman, “Modulation of orthogonal body waves enables high maneuverability in sidewinding locomotion,”Proceedings of the National Academy of Sciences, vol....

2015
[19]

Modeling rolling gaits of a snake robot,

W. Zhen, C. Gong, and H. Choset, “Modeling rolling gaits of a snake robot,” in2015 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2015, pp. 3741–3746

2015
[20]

Ladder climbing with a snake robot,

T. Takemori, M. Tanaka, and F. Matsuno, “Ladder climbing with a snake robot,” in2018 IEEE/RSJ International Conference on Intelli- gent Robots and Systems (IROS). IEEE, 2018, pp. 8140–8145

2018
[21]

Hoop-passing motion for a snake robot to realize motion transition across different environments,

T. Takemori, M. Tanaka, and F. Matsuno, “Hoop-passing motion for a snake robot to realize motion transition across different environments,” IEEE Transactions on Robotics, vol. 37, no. 5, pp. 1696–1711, 2021

2021
[22]

Adaptive helical rolling of a snake robot to a straight pipe with irregular cross-sectional shape,

T. Takemori, M. Tanaka, and F. Matsuno, “Adaptive helical rolling of a snake robot to a straight pipe with irregular cross-sectional shape,” IEEE Transactions on Robotics, vol. 39, no. 1, pp. 437–451, 2022

2022
[23]

Reconstruction of backbone curves for snake robots,

T. Wang, B. Lin, B. Chong, J. Whitman, M. Travers, D. I. Goldman, G. Blekherman, and H. Choset, “Reconstruction of backbone curves for snake robots,”IEEE Robotics and Automation Letters, vol. 6, no. 2, pp. 3264–3270, 2021

2021
[24]

Path-planning for perception-driven obstacle-aided snake robot locomotion,

K. G. Hanssen, A. A. Transeth, F. Sanfilippo, P. Liljeb ¨ack, and Ø. Stavdahl, “Path-planning for perception-driven obstacle-aided snake robot locomotion,” in2020 16th International Workshop on Advanced Motion Control (AMC). IEEE, 2020, pp. 98–104

2020
[25]

Autonomous decentralized shape-based navigation for snake robots in dense envi- ronments,

G. Sartoretti, T. Wang, G. Chuang, Q. Li, and H. Choset, “Autonomous decentralized shape-based navigation for snake robots in dense envi- ronments,” in2021 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2021, pp. 9276–9282

2021
[26]

Proprioceptive-inertial autonomous locomotion for articu- lated robots,

F. Ruscelli, G. Sartoretti, J. Nan, Z. Feng, M. Travers, and H. Choset, “Proprioceptive-inertial autonomous locomotion for articu- lated robots,” in2018 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2018, pp. 3436–3441

2018
[27]

Range-sensor-based semiau- tonomous whole-body collision avoidance of a snake robot,

M. Tanaka, K. Kon, and K. Tanaka, “Range-sensor-based semiau- tonomous whole-body collision avoidance of a snake robot,”IEEE Transactions on Control Systems Technology, vol. 23, no. 5, pp. 1927– 1934, 2015

1927
[28]

Perception-action coupling target tracking control for a snake robot via reinforcement learning,

Z. Bing, C. Lemke, F. O. Morin, Z. Jiang, L. Cheng, K. Huang, and A. Knoll, “Perception-action coupling target tracking control for a snake robot via reinforcement learning,”Frontiers in Neurorobotics, vol. 14, p. 591128, 2020

2020
[29]

Sensnake: A snake robot with contact force sensing for studying locomotion in complex 3-d terrain,

D. Ramesh, Q. Fu, and C. Li, “Sensnake: A snake robot with contact force sensing for studying locomotion in complex 3-d terrain,” in2022 International Conference on Robotics and Automation (ICRA). IEEE, 2022, pp. 2068–2075

2022
[30]

Physical intelligence as a new paradigm,

M. Sitti, “Physical intelligence as a new paradigm,”Extreme Mechan- ics Letters, vol. 46, p. 101340, 2021

2021
[31]

Robotic modelling of snake traversing large, smooth obstacles reveals stability benefits of body compliance,

Q. Fu and C. Li, “Robotic modelling of snake traversing large, smooth obstacles reveals stability benefits of body compliance,”Royal Society open science, vol. 7, no. 2, p. 191192, 2020

2020
[32]

Shape-based compliance in locomotion

M. J. Travers, J. Whitman, P. E. Schiebel, D. I. Goldman, and H. Choset, “Shape-based compliance in locomotion.” inRobotics: Science and Systems, 2016

2016
[33]

Incorporating frictional anisotropy in the design of a robotic snake through the exploitation of scales,

M. M. Serrano, A. H. Chang, G. Zhang, and P. A. Vela, “Incorporating frictional anisotropy in the design of a robotic snake through the exploitation of scales,” in2015 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2015, pp. 3729–3734

2015
[36]

Aquamilr: Mechanical intelligence simplifies control of undulatory robots in cluttered fluid environments,

T. Wang, N. Mankame, M. Fernandez, V . Kojouharov, and D. I. Goldman, “Aquamilr: Mechanical intelligence simplifies control of undulatory robots in cluttered fluid environments,” in2025 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2025, pp. 14 671–14 677

2025
[37]

Aquamilr+: Design of an unteth- ered limbless robot for complex aquatic terrain navigation,

M. Fernandez, T. Wang, G. Tunnicliffe, D. Dortilus, P. Gunnarson, J. O. Dabiri, and D. I. Goldman, “Aquamilr+: Design of an unteth- ered limbless robot for complex aquatic terrain navigation,” in2025 IEEE International Conference on Robotics and Automation (ICRA), Atlanta, USA, 2025

2025
[38]

Three- dimensional hydro-cluttered locomotion by an undulatory robot,

T. Wang, M. Fernandez, G. Tunnicliffe, N. Cornell, J. Duong, D. Dor- tilus, Z. J. Xu, P. Meza, S. Lublinsky, D. Parikh,et al., “Three- dimensional hydro-cluttered locomotion by an undulatory robot,” arXiv preprint arXiv:2606.06829, 2026

2026 arXiv
[39]

A’sidewinding’locomotion gait for hyper-redundant robots,

J. W. Burdick, J. Radford, and G. S. Chirikjian, “A’sidewinding’locomotion gait for hyper-redundant robots,” Advanced Robotics, vol. 9, no. 3, pp. 195–216, 1994

1994
[40]

Frequency modulation of body waves to improve performance of sidewinding robots,

B. Chong, T. Wang, J. M. Rieser, B. Lin, A. Kaba, G. Blekherman, H. Choset, and D. I. Goldman, “Frequency modulation of body waves to improve performance of sidewinding robots,”The International Journal of Robotics Research, vol. 40, no. 12-14, pp. 1547–1562, 2021

2021
[41]

Robophysical modeling of bilaterally activated and soft limbless locomotors,

P. E. Schiebel, M. C. Maisonneuve, K. Diaz, J. M. Rieser, and D. I. Goldman, “Robophysical modeling of bilaterally activated and soft limbless locomotors,” inBiomimetic and Biohybrid Systems: 9th International Conference, Living Machines 2020, Freiburg, Germany, July 28–30, 20...

2020
[42]

Directional compliance in obstacle-aided navigation for snake robots,

T. Wang, J. Whitman, M. Travers, and H. Choset, “Directional compliance in obstacle-aided navigation for snake robots,” in2020 American Control Conference (ACC). IEEE, 2020, pp. 2458–2463

2020
[1993]

IEEE, 1993, pp

Proceedings IEEE International Conference on Robotics and Automation. IEEE, 1993, pp. 101–106

1993

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.

[1] [34]

Mechanical intelligence simplifies control in terrestrial limbless locomotion,

T. Wang, C. Pierce, V . Kojouharov, B. Chong, K. Diaz, H. Lu, and D. I. Goldman, “Mechanical intelligence simplifies control in terrestrial limbless locomotion,”Science Robotics, vol. 8, no. 85, p. eadi2243, 2023

2023

[2] [35]

Anisotropic body compliance facilitates robotic sidewinding in complex environments,

V . Kojouharov, T. Wang, M. Fernandez, J. Maeng, and D. I. Gold- man, “Anisotropic body compliance facilitates robotic sidewinding in complex environments,” in2024 International Conference on Robotics and Automation (ICRA). IEEE, 2024

2024

[3] [1]

Biologically inspired robots,

S. Hirose, “Biologically inspired robots,”Snake-Like Locomotors and Manipulators, 1993

1993

[4] [2]

Biologically inspired snake-like robots,

S. Hirose and M. Mori, “Biologically inspired snake-like robots,” in 2004 IEEE International Conference on Robotics and Biomimetics. IEEE, 2004, pp. 1–7

2004

[5] [3]

A review on modelling, implementation, and control of snake robots,

P. Liljeb ¨ack, K. Y . Pettersen, Ø. Stavdahl, and J. T. Gravdahl, “A review on modelling, implementation, and control of snake robots,” Robotics and Autonomous Systems, vol. 60, no. 1, pp. 29–40, 2012

2012

[6] [4]

Design of a modular snake robot,

C. Wright, A. Johnson, A. Peck, Z. McCord, A. Naaktgeboren, P. Gi- anfortoni, M. Gonzalez-Rivero, R. Hatton, and H. Choset, “Design of a modular snake robot,” in2007 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 2007, pp. 2609–2614

2007

[7] [5]

Design and architecture of the unified modular snake robot,

C. G. Wright, A. D. Buchan, B. Brown, J. C. Geist, M. Schwerin, D. Rollinson, M. Tesch, and H. Choset, “Design and architecture of the unified modular snake robot,” in2012 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2012, pp. 4347–4354

2012

[8] [6]

Design and architecture of a series elastic snake robot,

D. Rollinson, Y . Bilgen, B. Brown, F. Enner, S. Ford, C. Layton, J. Rembisz, M. Schwerin, A. Willig, P. Velagapudi, and H. Choset, “Design and architecture of a series elastic snake robot,” in2014 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2014, pp. 4630–4636

2014