Pith. sign in

REVIEW 2 major objections 5 minor 30 references

Efficient variable-length hanging tether parameterization for marsupial robot planning in 3D environments

T0 review · 2 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read A parabola stand-in makes hanging-tether robot planning 25x faster.

desk verdict Real speedup in RRT* from the parabola decision problem; the trajectory-optimizer speedup over [19] rests on baseline times from another machine. read the letter →

arxiv 2502.04467 v1 pith:3IWEYYLS submitted 2025-02-06 cs.RO

S. Martínez-Rozas , D. Alejo , F. Caballero , L. Merino , M.A. Pérez-Cutiño , F. Rodriguez , V. Sánchez-Canales , I. Ventura
show 1 more author
J.M. Díaz-Bañez
This is my paper · ORCID
classification cs.RO MSC 68T40
keywords hangingtethermarsupialrobotsUAV-UGVcatenaryparabolaapproximationmotionplanningtrajectoryoptimizationRRT*
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 make path and trajectory planning for a marsupial robot pair — a ground vehicle and an aerial vehicle joined by a hanging tether — fast enough for practical and eventually real-time use. Its central move is to replace the catenary curve, which is expensive to compute and must be checked thousands of times during planning, with a parabola, then fit the final catenary only when needed. The paper reports that planning with the parabola decision problem is 1.9 to 25 times faster than the catenary baseline in RRT*, and that trajectory optimization time drops by about two orders of magnitude while average feasibility of the optimized trajectory rises to 97.3%. A sympathetic reader would take away that the parabola is not just a heuristic shortcut but a parameterization that can carry the whole planning pipeline from sampling to optimization.

What carries the argument

The load-bearing object is the Parabola Decision Problem and its iterative solver: starting from the straight segment $AB$, the algorithm finds the set of polygons the current parabola intersects, takes their convex hull, and jumps to the longest parabola through a hull vertex, repeating until the parabola is collision-free or exceeds length $L$. Two geometric facts carry the argument: two distinct parabolas through the same endpoints meet at most twice, so the region under the longest parabola contains the regions under all shorter ones (Lemma 1), and a catenary length-monotonicity result (Corollary 6) that turns fitting the catenary into a bisection on length, Algorithm 2. The optimization stage then uses the parabola or catenary parameters $(p,q,r)$ or $(a,x_0,y_0)$ as decision variables, which lets the solver compute tether-obstacle residuals and gradients directly.

What would settle it

Run the same planning pipeline in a scenario with a large sag-to-span ratio or steeply unequal suspension heights, measure the maximum vertical distance between the BySampling catenary and its source parabola, and inflate obstacles by the paper's $\tau$; if that distance exceeds $\tau$ at any obstacle vertex, a collision-free parabola can produce a catenary that intersects the inflated obstacle, breaking the guarantee.

Watch

Extended reading notes

Core claim

The central claim is that the hanging-tether planning problem can be reorganized around a parabola decision problem without losing collision-free guarantees: instead of asking whether some catenary of length at most $L$ connects the two suspension points while avoiding obstacles, the planner asks whether such a parabola exists, solves that with an iterative geometric algorithm, and then converts the winning parabola into a catenary by bisection. The authors prove that the longest valid parabola is found by repeatedly moving to the parabola of maximum length through the convex hull of the obstacles the current parabola hits, which guarantees termination and keeps the invariant that any shorter parabola is invalid. The collision-free status of the final catenary is maintained by inflating obstacles by a tolerance $\tau$ chosen from the measured parabola-catenary approximation error, or by rechecking and adjusting after fitting. The trajectory optimizer directly includes the three curve parameters of the tether in the state vector, so gradients with respect to the tether's shape are available analytically; this is what cuts optimization time by roughly two orders of magnitude.

Load-bearing premise

The plan assumes that the maximum vertical gap between the fitted catenary and the parabola never exceeds the inflated-obstacle tolerance $\tau = 0.035L$ (about 0.6 m in the tests); that bound was observed in 100 random trials and not proved, so a new geometry with a larger gap could put the supposedly collision-free catenary inside an obstacle.

Editorial extensions

If this is right

  • RRT* path planning for tethered UAV-UGV systems can run fast enough for local replanning instead of only offline computation.
  • Pushing more of the tether geometry into the optimizer makes the optimized trajectories more often collision-free, specifically 97.3% average feasibility for the parabola parameterization.
  • The same decision-problem structure should transfer to other pairs of suspension points and fixed maximum tether length, since the geometric lemmas do not depend on the robot models.
  • Computing catenary fits only after a parabola passes the decision test concentrates the expensive transcendental solves on configurations that are already likely feasible.

Reading between the lines

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

  • If the empirical $\tau$ bound holds generally, obstacle inflation makes the final catenary collision-free, but the authors verified it only on 100 random trials; a proof or a distribution-free bound would be needed before safety-critical deployment.
  • The curve-parameter-in-state-vector idea could carry over to tethers with buoyancy, drag, or other effects that make the catenary model invalid, as long as a low-dimensional parametric curve still describes the sag.
  • A testable extension is to relax the fixed-suspension assumption: allowing the tether to lie partly on the ground or wrap around obstacles would change the decision problem, but the parabola-based speedup would still apply to the curved segments.
  • The two-orders-of-magnitude optimization gain suggests that the bottleneck in prior work was gradient computation through transcendental equations, not the optimizer itself; removing that bottleneck is the deeper architectural lesson.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. The paper addresses planning for a marsupial UAV-UGV system connected by a variable-length hanging tether. To avoid the computational cost of repeatedly fitting catenaries, the authors propose a Parabola Decision Problem (PDP) that finds collision-free parabolic arcs quickly, followed by a numerical fit to a catenary. They also reparameterize the trajectory optimization state to include the three parameters of the tether curve (parabola or catenary) instead of only the tether length. Experiments in five scenarios, each repeated 100 times, compare the proposed parabola-based RRT* and trajectory optimizer against the catenary-based method of [19]. Reported results show 1.9x-25x RRT* speedups, roughly two-orders-of-magnitude optimization speedups, and average trajectory-optimizer feasibility of 97.3% for the parabola parameterization.

Significance. If the speedup results are confirmed, this is a practically significant contribution: it enables near-real-time planning for tethered marsupial systems, which previously required offline computation. The idea of using the parabola-catenary similarity for a decision problem, and of including curve parameters directly in the optimization state, is novel and clearly explained. The paper is also strengthened by an anonymous public source-code repository, repeated randomized experiments, and an internally controlled RRT* comparison between the parabola and catenary models on the same machine. However, the headline optimization-time speedup is currently quantified against the reported timings of [19] rather than a same-hardware rerun, and the stated collision-free guarantee is based on an empirical error bound rather than a proof. Both issues are addressable and do not invalidate the core algorithmic idea.

major comments (2)
  1. [Section VI-B (Table III) and Section VI] The claim of "two orders of magnitude" speedup in trajectory optimization is not based on a controlled comparison. The text states "All the experiments have been run on the same computer, an eight-core AMD Ryzen 7 6800H" (Section VI) but later says "We use the results of [19] as reported in the paper" (Section VI-B2). The baseline timings for [19] in Table III therefore were not produced on the same hardware, so the apparent 100x speedup may be partly due to hardware differences. Since this speedup is a central contribution, please re-run [19] on the same machine and report the timings, or clearly qualify the comparison as cross-hardware and provide the original hardware specifications.
  2. [Section III-C] The statement that "using BySampling and tau = 0.6 m, we can ensure that Algorithm 2 solves the CDP problem and computes a collision-free catenary" is not justified. The value tau = 0.035*L is chosen because the mean vertical deviation "never exceeded this value in the experiments" over 100 random trials; this is an empirical observation, not a formal guarantee. If the final method relies on the reevaluation step described later in the same section rather than on obstacle inflation, the text should say so explicitly and the tau-based guarantee should be presented as a heuristic. If a rigorous guarantee is claimed, a proven error bound for the BySampling fitting method is needed.
minor comments (5)
  1. [Section III-B2 (proof of Theorem 5)] In the proof, the sentence "From Theorem 5, r1 = A1/L1 = C1/L3..." should refer to Theorem 4 (the area-length ratio property from [23]), not Theorem 5, which is the statement being proved.
  2. [Section III-B4] The choice of approximation parameters for the benchmarking experiments (epsilon = 10^-2 for ByFitting, n = 5 for BySampling) is not justified; a sentence explaining how these values were selected would improve reproducibility.
  3. [Section V-B, Eq. (13)] The text says the system is "a relatively simple linear system of equations for p, q, r and y," but y is not an unknown in Eq. (13); the unknowns are p, q, r (with aC given). Please correct this description.
  4. [Section III-C] There is a typo: "The election of tau strongly depends on the approximation method used" should read "The selection of tau...".
  5. [Table III] The table reports negative mean accelerations for the [19] rows (e.g., S1.1 and S5.1). If these are artifacts of the sign convention, please clarify; otherwise explain why the baseline has negative mean accelerations contrary to the desired value rho_aa = 0.0.

Circularity Check

1 steps flagged · score 2.0 of 10

Core parabola parameterization is independent; the only mildly circular step is the tau safety margin, which is calibrated from its own approximation-error experiments and then presented as a guarantee.

  1. fitted input called prediction [Section III-C, 'Ensuring a collision-free catenary' (sentence after Table I)]
    "setting it to the average approximation error plus its standard deviation is a good trade-off, particularly τ = 0.035∗L ... Note that the mean vertical distance between the parabola and the catenary never exceeded this value in the experiments. ... Therefore, using BySampling and τ = 0.6 m, we can ensure that Algorithm 2 solves the CDP problem and computes a collision-free catenary for the marsupial robotic system."

    The collision-free guarantee for the fitted catenary rests on the bound tau = 0.035*L, which is not proven but is set to the average approximation error plus one standard deviation measured in the same 100 random approximation experiments summarized in Table I. The paper then converts this empirically calibrated statistic into an unconditional 'ensure' statement for the CDP. In any new scenario, the claimed collision-freeness is an assumption inherited from the fitted error distribution, not a derived result. This is a fitted input presented as a guarantee, though it is not the central parameterization contribution and does not make the overall speedup claim tautological.

full rationale

The central contribution is self-contained: the parabola model is introduced as an approximation of the catenary, the Parabola Decision Problem is solved through geometric lemmas (Lemma 1, Remark 2) and Algorithm 1, and the catenary is recovered by a bisection fitting procedure (Algorithm 2). None of these steps defines the target result in terms of itself. The only mild self-reference is the tau calibration in Section III-C, where the safety margin is taken from the mean and standard deviation of the same approximation-error experiments that are used to argue accuracy, and then treated as a collision-freeness guarantee. This is an overclaim but not a circular derivation of the planner's main result. The trajectory-optimization speedup relative to [19] uses reported numbers from a prior paper by overlapping authors ('We use the results of [19] as reported in the paper'); however, that comparison is a reproducibility and hardware-comparability concern rather than a reduction of the new method's claim to its own inputs, and the RRT* speedups in Table II are internally controlled. Overall, the paper's core derivation is independent, so the circularity score is low.

Assumptions & free parameters 5 free parameters · 7 assumptions · 0 invented entities

The method introduces no new physical entities. It relies on standard catenary and parabola mathematics, a set of domain assumptions about tether statics and obstacle representation, and a small number of hand-tuned or empirically calibrated parameters. The most fragile entries are the ad hoc error bound tau and the optimizer weights, both of which are tuned rather than derived.

free parameters (5)
  • Obstacle inflation tolerance tau = 0.6 m (also stated as 0.035*L)
    Chosen from 100 random experiments as average approximation error plus one standard deviation; used to claim the fitted catenary is collision-free.
  • Tether length safety margin = 5% shorter and 5% longer than parabola length
    Ad hoc margin to keep the catenary within length constraints after approximation.
  • ByFitting bisection error epsilon = 1e-2
    Stop criterion for the catenary-fitting bisection in Algorithm 2.
  • BySampling number of points n = 5
    Number of sampled points in the nonlinear catenary fitting method.
  • Optimizer weights and thresholds = gamma_eg=0.2, gamma_ea=0.25, gamma_ot=0.25, rho_oa=1.2, rho_ot=0.1, beta=10, etc.
    Hand-tuned objective weights and collision thresholds used in the trajectory optimization; listed in Section VI-B without a tuning procedure.
assumptions (7)
  • domain assumption The tether follows a static catenary curve between two suspension points.
    Used throughout, e.g., Equation (1), assuming uniform weight, no wind, no bending stiffness, and negligible dynamics.
  • domain assumption The tether lies in the vertical plane that passes through the two suspension points.
    Reduces the 3D problem to a 2D plane pi in Section III, which is standard for a hanging cable but ignores out-of-plane forces.
  • domain assumption Obstacles in the 2D plane can be represented as convex polygons or convex hulls of clusters.
    Algorithm 1 requires convex polygons; the implementation computes convex hulls of point-cloud clusters, which may merge or split obstacles in ways not analyzed.
  • standard math Two different parabolas in the family P_AB(v) intersect at most twice, including the endpoints A and B.
    Used in Lemma 1 to conclude that the regions R(v) are nested by length. This is a standard property of quadratic differences.
  • standard math For a catenary, the ratio of area under the curve to arc length over any horizontal interval is constant (Theorem 4 from [23]).
    Invoked in the proof of Theorem 5 to derive a contradiction about intersecting catenaries.
  • ad hoc to paper The empirical approximation-error statistics observed in 100 random experiments bound the true error for all planning scenarios.
    Section III-C sets tau = mean plus one standard deviation and then claims this 'ensures' collision-free catenaries, which is a statistical heuristic, not a proven worst-case bound.
  • domain assumption The parabola and the tether never touch the ground during planning.
    Algorithm 1 returns 'None' if the parabola touches the ground; the method does not handle ground-contact or dragging tether states.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient variable-length hanging tether parameterization for marsupial robot planning in 3D environments." pith.science (2026). https://pith.science/paper/3IWEYYLS

@misc{pith2026250204467,
  author       = {Pith},
  title        = {Pith review of: Efficient variable-length hanging tether parameterization for marsupial robot planning in 3D environments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3IWEYYLS}},
  note         = {Machine review of arXiv:2502.04467}
}
read the original abstract

This paper presents a novel approach to efficiently parameterize and estimate the state of a hanging tether for path and trajectory planning of a UGV tied to a UAV in a marsupial configuration. Most implementations in the state of the art assume a taut tether or make use of the catenary curve to model the shape of the hanging tether. The catenary model is complex to compute and must be instantiated thousands of times during the planning process, becoming a time-consuming task, while the taut tether assumption simplifies the problem, but might overly restrict the movement of the platforms. In order to accelerate the planning process, this paper proposes defining an analytical model to efficiently compute the hanging tether state, and a method to get a tether state parameterization free of collisions. We exploit the existing similarity between the catenary and parabola curves to derive analytical expressions of the tether state.

Figures

Figures reproduced from arXiv: 2502.04467 by the authors.

Figure 1
Figure 1. Simplified 2D sketch showing an example for motion planning of [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Polygons inside the T region, which is delimited by the black trapezoid ABFAFBA. A and B are the suspension points; FA and FB are their projections on the floor. The blue polygon is weakly inscribed in the parabolic region R(v ∗) and AB, delimited by the blue parabola. III. EFFICIENT COMPUTATION OF A COLLISION-FREE CATENARY CURVE BETWEEN TWO POINTS In this section, we present a new method for efficient computation o… view at source ↗
Figure 3
Figure 3. Example of a 2D projection, in which each cluster is represented with [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Scenarios considered for validation. S1: Open/constrained space with arc as obstacle. S2: Narrow/constrained space with denied access to UGV. S3: [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Violin plots of the distribution of the execution times of the Decision Problem test bench with Parabola and Catenary methods. In blue, the approximate [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 28 canonical work pages

  1. [19]

    Path and trajectory planning of a tethered uav-ugv marsupial robotic system,

    S. Mart ´ınez-Rozas, D. Alejo, F. Caballero, and L. Merino, “Path and trajectory planning of a tethered uav-ugv marsupial robotic system,” IEEE Robotics and Automation Letters , vol. 8, no. 10, pp. 6475–6482, 2023

    work page 2023

  2. [1]

    Tethered unmanned aerial vehicles—a systematic review,

    M. N. Marques, S. A. Magalh ˜aes, F. N. Dos Santos, and H. S. Mendonc ¸a, “Tethered unmanned aerial vehicles—a systematic review,” Robotics, vol. 12, no. 4, 2023. [Online]. Available: https://www.mdpi.com/2218-6581/12/4/117

    work page 2023

  3. [2]

    A motion planning approach for marsu- pial robotic systems,

    P. G. Stankiewicz, S. Jenkins, G. E. Mullins, K. C. Wolfe, M. S. Johannes, and J. L. Moore, “A motion planning approach for marsu- pial robotic systems,” in 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Madrid, 2018 . IEEE, 2018, pp. 1–9

    work page 2018

  4. [3]

    The power-tethered uav-ugv team: A collaborative strategy for navigation in partially-mapped environments,

    C. Papachristos and A. Tzes, “The power-tethered uav-ugv team: A collaborative strategy for navigation in partially-mapped environments,” in 22nd Mediterranean Conference on Control and Automation , 2014, pp. 1153–1158

    work page 2014

  5. [4]

    Marsupial and shape-shifting robots for urban search and rescue,

    R. Murphy, “Marsupial and shape-shifting robots for urban search and rescue,” IEEE Intelligent Systems and their Applications , vol. 15, no. 2, pp. 14–19, 2000

    work page 2000

  6. [5]

    Power supply system anal- ysis for tethered drones application,

    W. Walendziuk, D. Oldziej, and M. Slowik, “Power supply system anal- ysis for tethered drones application,” in 2020 International Conference Mechatronic Systems and Materials (MSM) , 2020, pp. 1–6

    work page 2020

  7. [6]

    Indoor uav localization using a tether,

    J. D. X. Xiao, Y . Fan and R. Murphy, “Indoor uav localization using a tether,” in 2018 IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR), Philadelphia, PA, 2018. IEEE, 2018, pp. 1–6

    work page 2018

  8. [7]

    Optimization-based trajectory planning for tethered aerial robots,

    S. Mart ´ınez-Rozas, D. Alejo, F. Caballero, and L. Merino, “Optimization-based trajectory planning for tethered aerial robots,” in 2021 IEEE International Conference on Robotics and Automation (ICRA), 2021, pp. 362–368

    work page 2021

Show all 30 references
  1. [8]

    Lockwood, A Book of Curves , 1st ed

    E. Lockwood, A Book of Curves , 1st ed. Cambridge University Press, ISBN: 103156, 1961

    1961

  2. [9]

    Autonomous visual assistance for robot operations using a tethered uav,

    X. Xiao, J. Dufek, and R. R. Murphy, “Autonomous visual assistance for robot operations using a tethered uav,” in Field and Service Robotics, G. Ishigami and K. Yoshida, Eds. Singapore: Springer Singapore, 2021, pp. 15–29

    2021

  3. [10]

    A multi-model framework for tether-based drone localization,

    R. Lima and G. Pereira, “A multi-model framework for tether-based drone localization,” Intell Robot Syst , vol. 108, no. 20, 2023

    2023

  4. [11]

    Tethered uav with combined multi-rotor and water jet propulsion for forest fire fighting,

    B. Viegas, C.and Chehreh and J. e. a. Andrade, “Tethered uav with combined multi-rotor and water jet propulsion for forest fire fighting,” Intell Robot Syst , vol. 104, no. 21, 2022

    2022

  5. [12]

    On the development of a tether- based drone localization system,

    R. R. Lima and G. A. S. Pereira, “On the development of a tether- based drone localization system,” in 2021 International Conference on Unmanned Aircraft Systems (ICUAS) , 2021, pp. 195–201

    2021

  6. [13]

    The catenary robot: Design and control of a cable propelled by two quadrotors,

    D. S. D’Antonio, G. A. Cardona, and D. Salda ˜na, “The catenary robot: Design and control of a cable propelled by two quadrotors,” IEEE Robotics and Automation Letters , vol. 6, no. 2, pp. 3857–3863, 2021

    2021

  7. [14]

    Catenary-based visual servoing for tether shape control between underwater vehicles,

    M. Laranjeira, C. Dune, and V . Hugel, “Catenary-based visual servoing for tether shape control between underwater vehicles,” Ocean Engineer- ing, vol. 200, p. 107018, 2020

    2020

  8. [15]

    Benchmarking tether-based uav motion primitives,

    X. Xiao, J. Dufek, and R. Murphy, “Benchmarking tether-based uav motion primitives,” in 2019 IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR) , 2019, pp. 51–55

    2019

  9. [16]

    Oxpecker: A tethered uav for inspection of stone-mine pillars,

    B. Martinez Rocamora, R. R. Lima, K. Samarakoon, J. Rathjen, J. N. Gross, and G. A. S. Pereira, “Oxpecker: A tethered uav for inspection of stone-mine pillars,” Drones, vol. 7, no. 2, 2023. [Online]. Available: https://www.mdpi.com/2504-446X/7/2/73

    2023

  10. [17]

    Entanglement definitions for tethered robots: Exploration and analysis,

    G. Battocletti, D. Boskos, D. Toli ´c, I. Palunko, and B. D. Schutter, “Entanglement definitions for tethered robots: Exploration and analysis,”

  11. [18]

    An efficient strategy for path planning with a tethered marsupial robotics system,

    J. Capit ´an, J. M. D ´ıaz-B´a˜nez, M. A. P ´erez-Cuti˜no, F. Rodr ´ıguez, and I. Ventura, “An efficient strategy for path planning with a tethered marsupial robotics system,” 2024. [Online]. Available: https://arxiv.org/abs/2408.02141

    2024 arXiv

  12. [20]

    A fresh look at the catenary,

    F. Behroozi, “A fresh look at the catenary,” European Journal of Physics, vol. 35, no. 5, p. 055007, 2014

    2014

  13. [21]

    An algorithm for the parabo1ic approxima- tion of the catenary applicable in both inclined and level spans,

    A. Hatibovic and P. K ´ad´ar, “An algorithm for the parabo1ic approxima- tion of the catenary applicable in both inclined and level spans,” in 2018 International IEEE Conference and Workshop in ´Obuda on Electrical and Power Engineering (CANDO-EPE) . IEEE, 2018, pp. 000 217– 000 222

    2018

  14. [22]

    Comparison of the length of the catenary curve and its parabolic approximation in the span of an overhead line,

    A. Hatibovic, P. K ´ad´ar, and G. Morva, “Comparison of the length of the catenary curve and its parabolic approximation in the span of an overhead line,” in 2020 IEEE 3rd International Conference and Workshop in ´Obuda on Electrical and Power Engineering (CANDO- EPE). IEEE, 2...

    2020

  15. [23]

    A property characterizing the catenary,

    E. Parker, “A property characterizing the catenary,” Mathematics Mag- azine, vol. 83, no. 1, pp. 63–64, 2010

    2010

  16. [24]

    Sampling-based algorithms for optimal motion planning,

    S. Karaman and E. Frazzoli, “Sampling-based algorithms for optimal motion planning,” The International Journal of Robotics Research , vol. 30, no. 7, pp. 846–894, 2011

    2011

  17. [25]

    Driving on point clouds: Motion planning, trajectory optimization, and terrain assessment in generic nonplanar environments,

    P. Kr ¨usi, P. Furgale, M. Bosse, and R. Siegwart, “Driving on point clouds: Motion planning, trajectory optimization, and terrain assessment in generic nonplanar environments,” Journal of Field Robotics, vol. 34, no. 5, pp. 940–984, 2017. [Online]. Available: https://onlineli...

    2017 doi

  18. [26]

    3d distance fields: a survey of techniques and applications,

    M. Jones, J. Baerentzen, and M. Sramek, “3d distance fields: a survey of techniques and applications,” IEEE Transactions on Visualization and Computer Graphics, vol. 12, no. 4, pp. 581–599, 2006

    2006

  19. [27]

    Rviz: a toolkit for real domain data visualization,

    H. Kam, S.-H. Lee, T. Park, and C.-H. Kim, “Rviz: a toolkit for real domain data visualization,” Telecommunication Systems, vol. 60, pp. 1– 9, 10 2015

    2015

  20. [28]

    Ceres solver,

    S. Agarwal, K. Mierle, and Others, “Ceres solver,” http://ceres-solver. org, 2021

    2021

  21. [29]

    Comparison of the length of the catenary curve and its parabolic approximation in the span of an overhead line,

    A. Hatibovic, P. K ´ad´ar, and G. Morva, “Comparison of the length of the catenary curve and its parabolic approximation in the span of an overhead line,” in 2020 IEEE 3rd International Conference and Workshop in ´Obuda on Electrical and Power Engineering (CANDO- EPE), 2020, p...

    2020

  22. [2024]

    Available: https://arxiv.org/abs/2402.04909

    [Online]. Available: https://arxiv.org/abs/2402.04909

    arXiv

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.