Pith. sign in

REVIEW 4 major objections 6 minor 29 references

Moving Matter: Using a Single, Simple Robot to Reconfigure a Connected Set of Building Blocks

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read In simulation, the histogram-based CH2C planner beats the GLC and MWPMexpand heuristics on cost when start and target shapes are linearly separable and all operations cost one unit; it also runs on a Bill-E robot with some manual help.

desk verdict A useful, honest empirical comparison of CH2C against prior planners, with real reproducibility gaps around map generation and statistics. read the letter →

arxiv 2506.23333 v1 pith:BFFAH5EH submitted 2025-06-29 cs.RO cs.CGcs.DS

Javier Garcia , Jonas Friemel , Ramin Kosfeld , Michael Yannuzzi , Peter Kramer , Christian Rieck , Christian Scheffer , Arne Schmidt
show 4 more authors
Harm Kube Dan Biediger Sándor P. Fekete Aaron T. Becker
This is my paper · ORCID
classification cs.ROcs.CGcs.DS
keywords polyominoreconfigurationsingle-robotmanipulationhistogramintermediateconfigurationsBill-Erobotconnectedtilestructuresschedulecostheuristiccomparisonprogrammablematter
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

At bottom, this paper is a practical stress test of a theoretical promise. It implements the recently proposed CH2C algorithm, which reconfigures a connected set of tiles using a single robot that walks on the tiles and carries one tile at a time while keeping the structure connected, and measures it against two existing planners, GLC and MWPMexpand. In 1000-map simulated test sets, CH2C completes every map and achieves the lowest total cost whenever the start and target shapes can be separated by a horizontal or vertical line and pickup, dropoff, and movement each cost one unit. The advantage disappears when the shapes overlap, and it reverses if tile pickups or dropoffs are much more expensive than moving, because CH2C performs many more of those operations. A hardware demonstration on a Bill-E inchworm robot shows that the method is physically executable, with manual assistance needed for some tile placements.

What carries the argument

The mechanism is a three-phase schedule built around histograms, intermediate shapes made of a unit-height base strip with unit-width columns attached. Phase I transforms the start configuration into a histogram by iteratively moving 'free components'—parts of the configuration that can be shifted without disconnecting it—downward; Phase II reconfigures between two opposing histograms by repeatedly moving the topmost, leftmost surplus tile to the topmost, leftmost deficit position; Phase III reverses Phase I to reach the target. This decomposition replaces the full polyomino reconfiguration problem with two easier conversions and is what carries the constant-factor approximation guarantee inherited from the original algorithm.

What would settle it

Run an exhaustive search for optimal schedules on all small polyomino pairs (up to, say, eight tiles) that are separable by a horizontal or vertical line under the unit-cost model, and compare CH2C's schedule cost to optimal and to GLC; if CH2C is not consistently below GLC and within a small constant of optimal on a large majority of instances, the paper's central performance claim fails. A cheaper probe is to re-run the Boxy and Snakey benchmarks with pickup and dropoff cost set to twice the move cost and locate the crossover where CH2C becomes worse than GLC.

Watch

Extended reading notes

Core claim

Central claim: the histogram-based approximation algorithm, whose worst-case guarantee was already known, is also the best practical choice in the regime its theory covers. Specifically, the paper reports that on the Boxy and Snakey benchmark sets, where the bounding boxes of start and target are separated by a line, CH2C yields lower total schedule cost than GLC and MWPMexpand under unit operation costs, with a 100% completion rate while MWPMexpand finishes only 71.4% and 50.5% of the maps and GLC tends to give the highest costs. The same advantage does not hold for overlapping configurations, where CH2C drops to roughly the level of GLC, and it depends on pickup and dropoff costing no more than movement; the paper reports that CH2C uses significantly more pickup/dropoff pairs than the other planners. The authors therefore position CH2C as the method of choice precisely when the separation condition holds and the cost model is uniform, and as less attractive when lifting and placing tiles dominates robot travel.

Load-bearing premise

The benchmark conclusions rest on the generated Boxy and Snakey maps being representative of real reconfiguration tasks, but the generation procedure is only loosely described (randomly adding tiles to the right and top, or in all directions, with no seeds or tile counts given), and the cost model assumes pickup, dropoff, and movement each cost exactly one unit; if real maps or real robot costs differ, the reported advantage may not hold.

Editorial extensions

If this is right

  • If start and target shapes are linearly separable, CH2C offers a complete planner with lower schedule cost than GLC and MWPMexpand under equal unit operation costs, so the separation condition can serve as a cheap pre-check before choosing a planner.
  • CH2C's 100% completion on the Boxy, Snakey, and Overlapping test sets makes it a more dependable fallback than MWPMexpand, which fails on up to half of the harder maps, even when its cost advantage is small.
  • When the robot platform makes pickups or dropoffs costlier than movement, the planner of choice shifts to GLC or MWPMexpand, since CH2C's lower travel cost is paid for with many more tile lifts.
  • The Bill-E hardware demonstration shows the schedule class is physically achievable for small translations, but real-robot constraints (magnet strength and front clearance) must be handled before the method becomes fully autonomous.

Reading between the lines

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

  • The separation condition in the benchmarks is stricter than the theory requires; testing CH2C on configurations separated by a line but with highly elongated or concave shapes would reveal how the practical advantage degrades as histogram creation becomes more expensive.
  • Because the paper shows a sharp dependence on the pickup/dropoff cost ratio, a natural extension is a hybrid that uses CH2C's shortest-path group moves but selects free components to minimize tile lifts; the paper does not explore this.
  • The linearity of CH2C cost against the TSP baseline in the Boxy and Snakey plots suggests the algorithm's practical makespan may scale predictably with instance size; a scaling study over the number of tiles would test whether the constant-factor guarantee translates into a useful asymptotic advantage over GLC.
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

4 major / 6 minor

Summary. The paper evaluates CH2C, a recently proposed constant-factor approximation algorithm for reconfiguring connected tile configurations with a single robot, against two existing heuristic planners (GLC and MWPMexpand) from prior work. The evaluation is carried out in simulation on three sets of 1000 maps (Boxy, Snakey, Overlapping) under a unit-cost model for move, pickup, and dropoff operations, and is complemented by a hardware demonstration using a Bill-E inchworm robot. The authors report that CH2C achieves the lowest total cost on the separated map sets with 100% completion, at the price of more pickup/dropoff operations, and that its advantage vanishes on overlapping configurations. The paper concludes that CH2C is superior when start and target can be separated by a horizontal or vertical line and when pickup/dropoff costs equal movement costs.

Significance. The paper provides the first practical evaluation of a theoretically grounded reconfiguration algorithm (CH2C) with a constant-factor guarantee, comparing it to established heuristics in a relevant robotic model. The inclusion of a physical demonstration on an inchworm robot, even with disclosed manual assistance, is a useful step toward bridging theory and hardware. The qualitative trends in the scatter plots are suggestive and the comparison against independent baselines (GLC, MWPMexpand) is sensible. However, the strength of the empirical conclusions is limited by the loosely specified map generation, the lack of any statistical measures, and an unresolved gap between the 2-scaled theoretical setting and the unit-tile simulations. If the reported trends are confirmed under better-specified and more diverse benchmarks, the paper would be a valuable reference for practitioners selecting reconfiguration planners.

major comments (4)
  1. [Section III-B] The description of the Boxy and Snakey map generators is insufficient for reproducibility and potentially unrepresentative: no random seeds, tile counts, or separation distances are given, and the growth rules ('tiles were randomly added to the right and top of existing tiles' for Boxy; 'prioritizing recently placed tiles' for Snakey) are anisotropic and may systematically favor shapes for which the histogram subroutine is efficient. Because the central conclusion that CH2C yields lower costs for horizontally or vertically separated configurations rests entirely on these two 1000-map benchmarks, the representativeness of the generated maps needs to be justified with concrete parameters, or the claim must be restricted to the tested shape families.
  2. [Section II and Section III-B] The theoretical algorithm in Becker et al. [5] is presented for 2-scaled configurations (2×2 tiles rooted at even grid coordinates), but the simulation maps in Section III-B appear to be ordinary unit-size tile polyominoes. The paper does not state whether the implementation scales the inputs or modifies the algorithm to handle unit tiles. Without this clarification, the connection between the implemented CH2C and the constant-factor guarantee is unclear, and the evaluation may not be testing the algorithm whose performance is claimed in the abstract.
  3. [Section III-A, Figures 5-7] The empirical comparison reports only scatter plots and completion rates, with no error bars, confidence intervals, or statistical tests. Statements such as 'CH2C displays a strongly linear relationship with the TSP cost' and the conclusion that CH2C 'results in lower costs' are not quantified, so the reader cannot assess the variability across the 1000 maps or determine whether the observed differences are robust to map-generation randomness. The authors should provide at least the variance of the costs or a statistical comparison (e.g., paired tests or confidence intervals) to support the qualitative claims.
  4. [Section III-B and Section III-F] The sentence 'All maps belong to the cases shown in Figures 4a and 4c' is ambiguous: it is unclear whether each of the Boxy and Snakey sets uses only a horizontal bisector, only the two-bisector case, or a mixture. This ambiguity matters because Section III-F later introduces yet another 1000-map set specifically for the two-bisector case, and the central claim concerns separation by a horizontal or vertical line. Please clarify the separation geometry of each benchmark set and state explicitly which of Figures 4a, 4b, or 4c each map family belongs to.
minor comments (6)
  1. [Section IV] The hardware demonstration required manual assistance for picking up and placing tiles; the Conclusion acknowledges this, but the abstract's phrase 'practical setting' and the claim of demonstrating feasibility should be qualified to avoid overstatement.
  2. [Section I-C] The definition of schedule says pickup, dropoff, and movement are 'weighted equally'; consider using 'counted equally' for clarity, since the term 'weighted' might suggest tunable weights.
  3. [Figures 5 and 6] The phrase 'strongly linear relationship' in the captions would be more informative with a reported slope, intercept, or R² value for the CH2C versus TSP cost scatter.
  4. [Section III-B] Please specify the number of tiles (or the range) and the spatial separation in the generated maps; these parameters are essential for interpreting the cost scales and for reproducing the experiments.
  5. [Section III-A] The Concorde TSP solver is mentioned without a citation; please add the appropriate reference so that readers can identify the exact solver version used.
  6. [General] Consider adding a data and code availability statement. Releasing the map generators, seeds, and simulation code would greatly enhance the reproducibility of the empirical claims.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: CH2C is evaluated against independent baselines; benchmark-generation concerns affect generality, not derivation.

full rationale

This is an evaluation and benchmarking paper, not a derivation. The central empirical claim — that CH2C results in lower total cost than GLC and MWPMexpand when start and target are separated by a horizontal or vertical line under unit operation costs — is supported by direct simulation comparisons on the Boxy, Snakey, and Overlapping map sets. No parameter is fitted to a subset of data and then presented as a prediction, and no quantity in the paper is defined in terms of the algorithm's own output. The CH2C algorithm and its constant-factor guarantee are cited from the companion paper [5] with substantial author overlap, but that citation is contextual rather than load-bearing: the present paper does not rely on the guarantee to establish its experimental conclusions, and the simulations are reported as independent measurements. The map generators in Section III-B are only loosely specified and are author-designed, which is a legitimate threat to external validity — the maps may oversample shapes that are easy for histogram-based reconfiguration — but that is a sampling and representativeness concern, not circularity under the seven enumerated patterns. The paper also honestly reports limiting evidence: degraded performance on overlapping configurations, significantly more pickup/dropoff pairs, sensitivity to free-component strategy and bisector orientation, and the need for manual assistance in the hardware demonstration. These are falsifiable observations rather than re-statements of assumptions. No equation, construction, or benchmark reduces to its own input by construction. Therefore no circular step is present and the score is 0.

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

The paper introduces no new entities or fitted parameters. It relies on three domain assumptions: fidelity of the CH2C implementation, the unit-cost model as the objective, and the representativeness of the generated test maps. All three are plausible but unverified due to missing code and data.

assumptions (3)
  • domain assumption The CH2C algorithm as implemented faithfully realizes the algorithm described in Becker et al. [5], including its constant-factor approximation guarantee for well-separated configurations.
    The paper does not prove or re-derive the algorithm; it cites [5] whose authors overlap heavily with the present paper. No code is provided to verify the implementation.
  • domain assumption The unit-cost model, where each move, pickup, and dropoff costs 1, is an adequate proxy for real robot operation and is the objective being optimized.
    All simulation costs are computed under this model; the paper acknowledges that different relative pickup/dropoff costs change the ranking in Section III-D.
  • domain assumption The random map generators for Boxy and Snakey produce representative instances of the intended application domain.
    The generation procedure is only loosely described in Section III-B, with no seeds, distributions, or tile counts, so the representativeness of the instance sets is unverifiable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Moving Matter: Using a Single, Simple Robot to Reconfigure a Connected Set of Building Blocks." pith.science (2026). https://pith.science/paper/BFFAH5EH

@misc{pith2026250623333,
  author       = {Pith},
  title        = {Pith review of: Moving Matter: Using a Single, Simple Robot to Reconfigure a Connected Set of Building Blocks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BFFAH5EH}},
  note         = {Machine review of arXiv:2506.23333}
}
read the original abstract

We implement and evaluate different methods for the reconfiguration of a connected arrangement of tiles into a desired target shape, using a single active robot that can move along the tile structure. This robot can pick up, carry, or drop off one tile at a time, but it must maintain a single connected configuration at all times. Becker et al. (CCCG 2025) recently proposed an algorithm that uses histograms as canonical intermediate configurations, guaranteeing performance within a constant factor of the optimal solution if the start and target configuration are well-separated. We implement and evaluate this algorithm, both in a simulated and practical setting, using an inchworm type robot to compare it with two existing heuristic algorithms.

Figures

Figures reproduced from arXiv: 2506.23333 by the authors.

Figure 1
Figure 1. Demonstration of a reconfiguration using the proposed [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The approximation algorithm by Becker et al. [ [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. The four possible cases of separation of the start and target configuration. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figures from the paper (7 more)
Figure 5
Figure 5. Figure 5: Simulation results for test set Boxy. CH2C displays a strongly linear relationship with the TSP cost. MWPMexpand fails often, and GLC completes all maps but has the highest costs. Example maps are shown. IV. Hardware Demonstration We utilized the Bill-E bot platform [2…
Figure 9
Figure 9. Figure 9: (Top-left) Map with U-shaped starting configuration. [PITH_FULL_IMAGE:figures/full_fig_p006_9.png]
Figure 7
Figure 7. Figure 7: Simulation results for test set Overlapping. This time, all the maps are created to have overlap. As a result, CH2C no longer performs much better than GLC, but at least completes all maps unlike MWPMexpand. 0 2000 4000 6000 8000 10000 12000 14000 16000 TSP Cost 0 200 …
Figure 10
Figure 10. Figure 10: Cost difference when using CH2C with a horizontal and a vertical bisector. The x-axis shows the difference between the horizontal and vertical gaps between the starting and target configurations’ bounding boxes. when the starting and target configurations can be separ…
Figure 8
Figure 8. Figure 8: Number of pickup/dropoff pairs for the different [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]
Figure 11
Figure 11. Figure 11: Starting from a configuration (left), our [PITH_FULL_IMAGE:figures/full_fig_p007_11.png]
Figure 12
Figure 12. Figure 12: The Bill-E bot translating a 2 × 2 configuration sideways. The bottom-left of each frame shows a top view of the map and robot. tiles may be unreachable from the outside of the shape and auxiliary tiles may be necessary to main￾tain connectivity [25]. CH2C could be au…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 29 canonical work pages

  1. [5]

    Efficient reconfiguration of tile arrange- ments by a single active robot,

    A. T. Becker, S. P. Fekete, J. Friemel, R. Kos- feld, P. Kramer, H. Kube, C. Rieck, C. Scheffer, and A. Schmidt, “Efficient reconfiguration of tile arrange- ments by a single active robot,” in Canadian Conference on Computational Geometry (CCCG) , 2025

    work page 2025

  2. [1]

    Compacting squares: Input-sensitive in- place reconfiguration of sliding squares,

    H. A. Akitaya, E. D. Demaine, M. Korman, I. Kostit- syna, I. Parada, W. Sonke, B. Speckmann, R. Uehara, and J. Wulms, “Compacting squares: Input-sensitive in- place reconfiguration of sliding squares,” in Scandinavian Symposium on Algorithm Theory (SW AT), 2022

    work page 2022

  3. [2]

    Sliding squares in parallel,

    H. A. Akitaya, S. P. Fekete, P. Kramer, S. Molaei, C. Rieck, F. Stock, and T. Wallner, “Sliding squares in parallel,” in European Symposium on Algorithms (ESA) , 2025

    work page 2025

  4. [3]

    Pushing lines helps: Efficient universal centralised transformations for programmable matter,

    A. Almethen, O. Michail, and I. Potapov, “Pushing lines helps: Efficient universal centralised transformations for programmable matter,” Theoretical Computer Science , vol. 830, 2020

    work page 2020

  5. [4]

    On efficient connectivity-preserving transforma- tions in a grid,

    ——, “On efficient connectivity-preserving transforma- tions in a grid,” Theoretical Computer Science, vol. 898, 2022

    work page 2022

  6. [6]

    Transformation of modular robots by rotation: 3 + 1 musketeers for all orthogonally convex shapes,

    M. Connor and O. Michail, “Transformation of modular robots by rotation: 3 + 1 musketeers for all orthogonally convex shapes,” Journal of Computer and System Sci- ences, vol. 150, 2025

    work page 2025

  7. [7]

    Algorithmic approaches to reconfigurable assembly systems,

    A. Costa, A. Abdel-Rahman, B. Jenett, N. Gershenfeld, I. Kostitsyna, and K. Cheung, “Algorithmic approaches to reconfigurable assembly systems,” in Aerospace Con- ference, 2019

    work page 2019

  8. [8]

    Con- nected reconfiguration of lattice-based cellular structures by finite-memory robots,

    S. P. Fekete, E. Niehs, C. Scheffer, and A. Schmidt, “Con- nected reconfiguration of lattice-based cellular structures by finite-memory robots,” Algorithmica, vol. 84, no. 10, 2022

    work page 2022

Show all 29 references
  1. [9]

    Reconfiguration planning for heterogeneous self-reconfiguring robots,

    R. Fitch, Z. J. Butler, and D. Rus, “Reconfiguration planning for heterogeneous self-reconfiguring robots,” in International Conference on Intelligent Robots and Sys- tems (IROS), 2003

    2003

  2. [10]

    Reconfiguration planning among obstacles for het- erogeneous self-reconfiguring robots,

    ——, “Reconfiguration planning among obstacles for het- erogeneous self-reconfiguring robots,” in International Conference on Robotics and Automation (ICRA) , 2005

    2005

  3. [11]

    Mmic-i: A robotic platform for assembly integration and internal locomotion through mechanical meta-material structures,

    O. Formoso, G. Trinh, D. Catanoso, I.-W. Park, C. Gregg, and K. Cheung, “Mmic-i: A robotic platform for assembly integration and internal locomotion through mechanical meta-material structures,” in International Conference on Robotics and Automation (ICRA) , 2023

    2023

  4. [12]

    Efficient shape reconfiguration by hybrid programmable matter,

    J. Friemel, D. Liedtke, and C. Scheffer, “Efficient shape reconfiguration by hybrid programmable matter,” in Eu- ropean Workshop on Computational Geometry (EuroCG), 2025

    2025

  5. [13]

    Connected reconfiguration of polyominoes amid obstacles using RRT ∗,

    J. Garcia, M. Yannuzzi, P. Kramer, C. Rieck, and A. T. Becker, “Connected reconfiguration of polyominoes amid obstacles using RRT ∗,” in International Conference on Intelligent Robots and Systems (IROS) , 2022

    2022

  6. [14]

    Reconfiguration of a 2D structure using spatio-temporal planning and load transferring,

    J. Garcia, M. Yannuzzi, P. Kramer, C. Rieck, S. P. Fekete, and A. T. Becker, “Reconfiguration of a 2D structure using spatio-temporal planning and load transferring,” in International Conference on Robotics and Automation (ICRA), 2024

    2024

  7. [15]

    Form- ing tile shapes with simple robots,

    R. Gmyr, K. Hinnenthal, I. Kostitsyna, F. Kuhn, D. Rudolph, C. Scheideler, and T. Strothmann, “Form- ing tile shapes with simple robots,” Natural Computing, vol. 19, no. 2, 2020

    2020

  8. [16]

    Ultralight, strong, and self-reprogrammable mechanical metamaterials,

    C. E. Gregg, D. Catanoso, O. I. B. Formoso, I. Kostitsyna, M. E. Ochalek, T. J. Olatunde, I. W. Park, F. M. Sebas- tianelli, E. M. Taylor, G. T. Trinh, and K. C. Cheung, “Ultralight, strong, and self-reprogrammable mechanical metamaterials,” Science Robotics, vol. 9, no. 86, 2024

    2024

  9. [17]

    Efficient shape formation by 3D hybrid programmable matter: An algorithm for low diameter intermediate structures,

    K. Hinnenthal, D. Liedtke, and C. Scheideler, “Efficient shape formation by 3D hybrid programmable matter: An algorithm for low diameter intermediate structures,” in Symposium on Algorithmic Foundations of Dynamic Networks (SAND) , 2024

    2024

  10. [18]

    Hopkins and M

    K. Hopkins and M. Beard, The colosseum . Harvard University Press, 2012

    2012

  11. [19]

    BILL-E: Robotic platform for locomotion and manipulation of lightweight space structures,

    B. Jenett and K. Cheung, “BILL-E: Robotic platform for locomotion and manipulation of lightweight space structures,” in Adaptive Structures Conference, 2017

    2017

  12. [20]

    Material–robot system for assembly of discrete cellular structures,

    B. Jenett, A. Abdel-Rahman, K. Cheung, and N. Ger- shenfeld, “Material–robot system for assembly of discrete cellular structures,” IEEE Robotics and Automation Let- ters, vol. 4, no. 4, 2019

    2019

  13. [21]

    De- sign of multifunctional hierarchical space structures,

    B. Jenett, C. Gregg, D. Cellucci, and K. Cheung, “De- sign of multifunctional hierarchical space structures,” in Aerospace Conference, 2017

    2017

  14. [22]

    Buoyancy en- abled autonomous underwater construction with cement blocks,

    S. Lensgraf, D. Balkcom, and A. Q. Li, “Buoyancy en- abled autonomous underwater construction with cement blocks,” in International Conference on Robotics and Automation (ICRA), 2023

    2023

  15. [23]

    Decoding mod- ular reconfigurable robots: A survey on mechanisms and design,

    G. Liang, D. Wu, Y. Tu, and T. L. Lam, “Decoding mod- ular reconfigurable robots: A survey on mechanisms and design,” The International Journal of Robotics Research , vol. 44, no. 5, 2025

    2025

  16. [24]

    On the transformation capability of feasible mechanisms for pro- grammable matter,

    O. Michail, G. Skretas, and P. G. Spirakis, “On the transformation capability of feasible mechanisms for pro- grammable matter,” Journal of Computer and System Sciences, vol. 102, 2019

    2019

  17. [25]

    Hiding sliding cubes: Why reconfiguring mod- ular robots is not easy (media exposition),

    T. Miltzow, I. Parada, W. Sonke, B. Speckmann, and J. Wulms, “Hiding sliding cubes: Why reconfiguring mod- ular robots is not easy (media exposition),” inSymposium on Computational Geometry (SoCG) , 2020

    2020

  18. [26]

    Reconfiguration algo- rithms for cubic modular robots with realistic movement constraints,

    MIT–NASA Space Robots Team, J. Brunner, K. C. Che- ung, E. D. Demaine, J. Diomidova, C. Gregg, D. H. Hendrickson, and I. Kostitsyna, “Reconfiguration algo- rithms for cubic modular robots with realistic movement constraints,” in Scandinavian Symposium on Algorithm Theory (SW AT), 2024

    2024

  19. [27]

    Recognition and reconfigura- tion of lattice-based cellular structures by simple robots,

    E. Niehs, A. Schmidt, C. Scheffer, D. E. Biediger, M. Yan- nuzzi, B. Jenett, A. Abdel-Rahman, K. C. Cheung, A. T. Becker, and S. P. Fekete, “Recognition and reconfigura- tion of lattice-based cellular structures by simple robots,” in International Conference on Robotics and Au...

    2020

  20. [28]

    SOLL-E: A module trans- port and placement robot for autonomous assembly of discrete lattice structures,

    I. Park, D. Catanoso, O. Formoso, C. Gregg, M. Ochalek, T. Olatunde, F. Sebastianelli, P. Spino, E. Taylor, G. Trinh, and K. C. Cheung, “SOLL-E: A module trans- port and placement robot for autonomous assembly of discrete lattice structures,” in International Conference on Int...

    2023

  21. [29]

    Reconfiguration of DNA molecular arrays driven by in- formation relay,

    J. Song, Z. Li, P. Wang, T. Meyer, C. Mao, and Y. Ke, “Reconfiguration of DNA molecular arrays driven by in- formation relay,” Science, vol. 357, no. 6349, 2017

    2017

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.