pith. sign in

arxiv: 2605.05502 · v1 · submitted 2026-05-06 · 🧮 math.OC · cs.SY· eess.SY

A Computationally Tractable Path-Planning Method for Airborne Wind Energy Systems

Manuel C.R.M. Fernandes , Fernando A.C.C. Fontes This is my paper

Pith reviewed 2026-05-08 15:54 UTC · model grok-4.3

classification 🧮 math.OC cs.SYeess.SY
keywords airborne wind energypath planningLissajous curvenonlinear programmingcrosswind flightpower maximizationreel-out phase
0
0 comments X

The pith

Optimizing Lissajous curve parameters in a nonlinear program yields efficient power-maximizing paths for airborne wind energy systems.

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

This paper shows that the path-planning problem for crosswind flight in airborne wind energy systems can be solved by optimizing the shape parameters of a Lissajous curve inside a nonlinear program. The goal is to maximize average power output during the reel-out phase while keeping the curve's curvature within feasible limits for the flying device. A sympathetic reader would care because full optimal control methods are too slow for repeated use, and this offers a faster way to generate good reference trajectories. If the Lissajous family is rich enough, the resulting paths should deliver near-optimal energy harvest without excessive computation.

Core claim

The authors formulate the reference path selection for the reel-out phase as a nonlinear optimization problem. They parameterize the desired flight path using a Lissajous curve and adjust its parameters to maximize the average power produced, subject to constraints on curvature. This yields a computationally tractable method that serves as an alternative to more demanding optimal control techniques and learning-based approaches for designing geometric flight paths in crosswind airborne wind energy systems.

What carries the argument

The nonlinear program that optimizes the parameters of a Lissajous curve to maximize average power production subject to curvature constraints.

If this is right

  • The method allows faster computation of reference paths compared to full optimal control solvers.
  • It supports real-time or frequent replanning under changing wind conditions.
  • Curvature constraints keep generated trajectories physically feasible for the tethered device.
  • Optimization focuses specifically on the reel-out phase where the system generates energy.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The parameterization could extend to reel-in phases or multi-cycle operations if the same curve family proves flexible there.
  • Coupling the planned paths with low-level flight controllers might reveal how tracking errors affect net power output.
  • Similar parametric-curve optimizations could apply to other tethered aerial vehicles that must follow smooth periodic trajectories.

Load-bearing premise

That Lissajous curves can approximate the power-optimal trajectories closely enough for practical purposes.

What would settle it

Running a high-accuracy optimal control solver on the same power model and finding that its achieved average power exceeds the Lissajous-based solution by more than a small margin.

Figures

Figures reproduced from arXiv: 2605.05502 by Fernando A.C.C. Fontes, Manuel C.R.M. Fernandes.

Figure 1
Figure 1. Figure 1: Global and Local coordinate systems. where s is an angular parameter spanning between 0 and 2π, nβ and nφ are the number of cycles the path describes over each axis, and its ratio nβ nφ is the relative cycle rate along both coordinates. This ratio defines the number of lobes in the geometric path. A closed geometric path is only obtained for rational ratios. As an example, an ellipse would have nβ nφ = 1 w… view at source ↗
Figure 2
Figure 2. Figure 2: Path parametrisation in the (φ, β) plane 2.2 Acting Forces The system is considered to be dominated by the aerodynamic (F⃗ aer) and tether force (F⃗ tether). The assumption of a perfectly taut tether causes the tether force to be radial. The aerodynamic force is divided into the lift and drag forces as F⃗ aer = F⃗ lif t + F⃗ drag (3) where Flif t = 1 2 ρAcLv 2 a (4) Fdrag = 1 2 ρAcDv 2 a (5) 5 view at source ↗
Figure 3
Figure 3. Figure 3: Turning lift on a spherical surface. We can also compute the required lateral acceleration for the kite to follow the path as al = v 2 k R0(s) = v 2 kκ(s) (11) where R0 is the local curvature radius of the path and κ is the curvature of the path  κ(s) = 1 R0(s)  . 6 view at source ↗
Figure 4
Figure 4. Figure 4: Maximum and minimum elevation angle constraints depiction. view at source ↗
Figure 5
Figure 5. Figure 5: Elliptical paths results. 11 view at source ↗
Figure 6
Figure 6. Figure 6: Figure-of-eight paths results. 12 view at source ↗
Figure 7
Figure 7. Figure 7: Elliptical and figure-of-eight paths in 3D. view at source ↗
Figure 8
Figure 8. Figure 8: Geometric similarity of the aerodynamic forces and the apparent velocity components. view at source ↗
Figure 9
Figure 9. Figure 9: Geometric similarity verification by triangle rotation. view at source ↗
read the original abstract

Airborne Wind Energy Systems (AWES) have emerged as a promising renewable energy technology that exploits stronger, more consistent high-altitude winds via tethered airborne devices. Among the various concepts, crosswind systems, where efficient flight control is essential to maximise energy output, offer significant potential. This paper addresses the problem of reference selection for crosswind flight control, focusing on the design of power-maximising geometric flight paths for the reel-out phase of Groundgen systems. To overcome the computational challenges associated with optimal control approaches, a computationally tractable framework is proposed in which a path-planning problem is formulated as a nonlinear program. The method optimises the parameters of a Lissajous curve to maximise the average power production over the reel-out phase, while incorporating curvature constraints. The proposed approach provides an efficient alternative to existing optimal control and learning-based methods.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 2 minor

Summary. The paper proposes a path-planning method for crosswind Airborne Wind Energy Systems (AWES) during the reel-out phase of Groundgen configurations. It formulates the problem as a nonlinear program (NLP) that optimizes the parameters of a Lissajous curve to maximize average power production while enforcing curvature constraints, positioning this as a computationally tractable alternative to full optimal-control and learning-based approaches.

Significance. If the Lissajous family proves sufficiently expressive for near-optimal trajectories and the embedded power model accurately captures aerodynamics and tether dynamics, the method could enable faster reference generation for AWES flight controllers. The explicit NLP structure with a low-dimensional parametrization is a clear strength, but the absence of any reported numerical results, approximation-error bounds, or baseline comparisons in the manuscript prevents assessment of practical gains over direct collocation or pseudospectral optimal control.

major comments (2)
  1. [Abstract / Formulation] The central claim that the Lissajous-based NLP yields power-maximizing paths comparable to optimal-control solutions is unsupported: the manuscript provides no numerical experiments, no comparison against a direct collocation or pseudospectral solver on the same power-production model, and no sensitivity analysis quantifying power loss due to the parametrization restriction (see abstract and the formulation section).
  2. [Path parametrization] No approximation-error bound or expressivity analysis is given for the Lissajous family relative to typical figure-eight or lemniscate crosswind trajectories; without this, it is impossible to determine whether the curvature-constrained optimum lies near the true power-optimal path (see the path parametrization and constraint sections).
minor comments (2)
  1. [Problem formulation] Clarify the exact definition of average power (integral over reel-out time or distance?) and how the curvature constraint is implemented inside the NLP (hard constraint or penalty?).
  2. [Introduction] The abstract states benefits over 'existing optimal control and learning-based methods' but cites no specific references or complexity comparisons; add a brief related-work paragraph with runtime or iteration counts.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their thorough review and valuable feedback on our manuscript. We agree with the identified shortcomings regarding the lack of numerical validation and will revise the paper accordingly to include the necessary experiments and analyses.

read point-by-point responses

  1. Referee: [Abstract / Formulation] The central claim that the Lissajous-based NLP yields power-maximizing paths comparable to optimal-control solutions is unsupported: the manuscript provides no numerical experiments, no comparison against a direct collocation or pseudospectral solver on the same power-production model, and no sensitivity analysis quantifying power loss due to the parametrization restriction (see abstract and the formulation section).

    Authors: We acknowledge that the current manuscript does not include numerical experiments or comparisons with optimal control solvers, which leaves the central claim unsupported. This is a valid criticism. In the revised version, we will add a dedicated numerical results section that includes simulations using the proposed Lissajous parametrization, direct comparisons with a pseudospectral optimal control method on the same power model, and a sensitivity analysis to quantify any power loss due to the restricted parametrization. This will provide evidence for the tractability and near-optimality claims. revision: yes

  2. Referee: [Path parametrization] No approximation-error bound or expressivity analysis is given for the Lissajous family relative to typical figure-eight or lemniscate crosswind trajectories; without this, it is impossible to determine whether the curvature-constrained optimum lies near the true power-optimal path (see the path parametrization and constraint sections).

    Authors: We agree that without an expressivity analysis, it is difficult to assess how well the Lissajous family approximates the optimal paths. We will include in the revision an analysis of the Lissajous curve's ability to represent typical crosswind trajectories such as figure-eights and lemniscates. This will involve deriving or numerically estimating approximation errors and discussing the impact on the curvature-constrained power maximization problem. revision: yes

Circularity Check

0 steps flagged

No circularity; direct NLP formulation on chosen parametrization

full rationale

The paper formulates the path-planning task as a nonlinear program that directly optimizes the parameters of a pre-selected Lissajous curve family to maximize average power subject to curvature constraints. This is a standard modeling and optimization choice with no evidence that the objective function, power model, or constraints reduce by construction to quantities already fitted or defined inside the paper. No self-citation chains, uniqueness theorems, or ansatzes imported from prior author work are invoked in the abstract or reader's summary to justify the central result. The derivation therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on an unstated but load-bearing model that maps geometric path parameters to instantaneous power; the paper treats this mapping as given and does not derive it.

free parameters (1)
  • Lissajous curve parameters
    These are decision variables optimized inside the NLP rather than fixed a priori; their optimal values are outputs of the method.
axioms (1)
  • domain assumption Power production is a known, differentiable function of the instantaneous flight-path geometry and tether length rate.
    Required to turn the path-planning task into a well-posed nonlinear program.

pith-pipeline@v0.9.0 · 5452 in / 1193 out tokens · 34554 ms · 2026-05-08T15:54:37.084147+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

34 extracted references · 25 canonical work pages

  1. [1]

    Airborne wind energy systems: A review of the technologies.Renewable & Sustainable Energy Reviews, 2015

    Antonello Cherubini, Andrea Papini, Rocco Vertechy, and Marco Fontana. Airborne wind energy systems: A review of the technologies.Renewable & Sustainable Energy Reviews, 2015. doi: 10. 1016/j.rser.2015.07.053

    2015

  2. [2]

    Marine Pollution Bulletin 123, 73–82

    Chris Vermillion, Mitchell Cobb, Lorenzo Fagiano, Rachel Leuthold, Moritz Diehl, Roy S. Smith, Tony A. Wood, Sebastian Rapp, Roland Schmehl, David J. Olinger, and Michael A. Demetriou. Electricity in the air: Insights from two decades of advanced control research and experimental flight testing of airborne wind energy systems.Annual Reviews in Control, 20...

    work page doi:10.1016/j 2021

  3. [3]

    Publications Office, 2018

    European Commission.Study on challenges in the commercialisation of airborne wind energy sys- tems. Publications Office, 2018. doi: doi/10.2777/87591

    work page doi:10.2777/87591 2018

  4. [4]

    Optimal control for power generating kites.2007 European Control Conference (ECC), 2007

    Boris Houska and Moritz Diehl. Optimal control for power generating kites.2007 European Control Conference (ECC), 2007. doi: 10.23919/ECC.2007.7068861

    work page doi:10.23919/ecc.2007.7068861 2007

  5. [5]

    Optimal crosswind towing and power generation with tethered kites.Journal of Guidance, Control, and Dynamics, 31(1):81–93, 2008

    Paul Williams, Bas Lansdorp, and Wubbo Ockels. Optimal crosswind towing and power generation with tethered kites.Journal of Guidance, Control, and Dynamics, 31(1):81–93, 2008. doi: 10.2514/ 1.30089

    2008

  6. [7]

    Performance assessment of a rigid wing airborne wind energy pumping system.Energy,

    Giovanni Licitra, Jonas Koenemann, Adrian B¨ urger, Paul Williams, Richard Ruiterkamp, and Moritz Diehl. Performance assessment of a rigid wing airborne wind energy pumping system.Energy,

  7. [8]

    doi: 10.1016/J.ENERGY.2019.02.064

    work page doi:10.1016/j.energy.2019.02.064 2019

  8. [9]

    Lu´ ıs Tiago Paiva and Fernando A. C. C. Fontes. Optimal control algorithms with adaptive time-mesh refinement for kite power systems.Energies, 11(3), 2018. ISSN 1996-1073. doi: 10.3390/en11030475

    work page doi:10.3390/en11030475 2018

  9. [10]

    Numerical trajectory optimization of airborne wind energy systems with stroboscopic averaging methods.IEEE Control Systems Letters, 2025

    Jakob Harzer, Jochem De Schutter, and Moritz Diehl. Numerical trajectory optimization of airborne wind energy systems with stroboscopic averaging methods.IEEE Control Systems Letters, 2025. doi: 10.1109/LCSYS.2025.3577225

    work page doi:10.1109/lcsys.2025.3577225 2025

  10. [11]

    Control of airborne wind energy systems based on nonlinear model predictive control & moving horizon estimation

    S´ ebastien Gros, Mario Zanon, and Moritz Diehl. Control of airborne wind energy systems based on nonlinear model predictive control & moving horizon estimation. In2013 European Control Conference (ECC), pages 1017–1022, 2013. doi: 10.23919/ECC.2013.6669713

    work page doi:10.23919/ecc.2013.6669713 2013

  11. [12]

    Flight trajectory optimization of fly-gen airborne wind energy systems through a harmonic balance method.Wind Energy Science, 2022

    Filippo Trevisi, Iv´ an Castro-Fern´ andez, Gregorio Pasquinelli, Carlo Emanuele Dionigi Riboldi, and Alessandro Croce. Flight trajectory optimization of fly-gen airborne wind energy systems through a harmonic balance method.Wind Energy Science, 2022. doi: 10.5194/WES-7-2039-2022

    work page doi:10.5194/wes-7-2039-2022 2022

  12. [13]

    Michael Erhard, Greg Horn, and Moritz Diehl. A quaternion-based model for optimal control of an airborne wind energy system.ZAMM - Journal of Applied Mathematics and Mechanics / Zeitschrift f¨ ur Angewandte Mathematik und Mechanik, 97(1):7–24, 2017. doi: 10.1002/zamm.201500180. 20

    work page doi:10.1002/zamm.201500180 2017

  13. [14]

    Manuel C. R. M. Fernandes, Lu´ ıs Tiago Paiva, and Fernando A. C. C. Fontes. Optimal path and path-following control in airborne wind energy systems.Computational methods in applied sciences,

  14. [15]

    doi: 10.1007/978-3-030-57422-2 26

    work page doi:10.1007/978-3-030-57422-2

  15. [16]

    Andreas H¨ ohl, Michael Maiworm, and Rolf Findeisen. Path following or tracking model predictive control for towing kites – a question of formulation or learning? In2023 IEEE Conference on Control Technology and Applications (CCTA), pages 459–465, 2023. doi: 10.1109/CCTA54093. 2023.10252231

    work page doi:10.1109/ccta54093 2023

  16. [17]

    Zgraggen, Lorenzo Fagiano, and Manfred Morari

    Aldo U. Zgraggen, Lorenzo Fagiano, and Manfred Morari. Real-time optimization and adaptation of the crosswind flight of tethered wings for airborne wind energy.IEEE Transactions on Control Systems and Technology, 2015. doi: 10.1109/tcst.2014.2332537

    work page doi:10.1109/tcst.2014.2332537 2015

  17. [18]

    Fathy, and Chris Vermillion

    Mitchell Cobb, Kira Barton, Hosam K. Fathy, and Chris Vermillion. Iterative learning-based path optimization for repetitive path planning, with application to 3-d crosswind flight of airborne wind energy systems.IEEE Transactions on Control Systems Technology, 2020. doi: 10.1109/TCST. 2019.2912345

    work page doi:10.1109/tcst 2020

  18. [19]

    and Bergveld, H

    Ali Baheri and Christopher Vermillion. Waypoint optimization using bayesian optimization: A case study in airborne wind energy systems.American Control Conference, 2019. doi: 10.23919/ ACC45564.2020.9147518

    work page arXiv 2019

  19. [20]

    Traction power generation with tethered wings.Green energy and technology, 2013

    Roland Schmehl, Michael Noom, and Rolf van der Vlugt. Traction power generation with tethered wings.Green energy and technology, 2013. doi: 10.1007/978-3-642-39965-7 2

    work page doi:10.1007/978-3-642-39965-7 2013

  20. [21]

    Miles L. Loyd. Crosswind kite power (for large-scale wind power production).Journal of Energy, 4 (3):106–111, 1980. doi: 10.2514/3.48021. URLhttps://doi.org/10.2514/3.48021

    work page doi:10.2514/3.48021 1980

  21. [22]

    Rautakorpi, and Risto Silvennoinen

    Ivan Argatov, P. Rautakorpi, and Risto Silvennoinen. Estimation of the mechanical energy output of the kite wind generator.Renewable Energy, 2009. doi: 10.1016/J.RENENE.2008.11.001

    work page doi:10.1016/j.renene.2008.11.001 2009

  22. [23]

    Luchsinger

    Rolf H. Luchsinger. Pumping cycle kite power.Green energy and technology, 2013. doi: 10.1007/ 978-3-642-39965-7 3

    2013

  23. [24]

    Quasi-steady model of a pumping kite power system.Renewable Energy, 2019

    Rolf van der Vlugt, Anna Bley, Michael Noom, and Roland Schmehl. Quasi-steady model of a pumping kite power system.Renewable Energy, 2019. doi: 10.1016/J.RENENE.2018.07.023

    work page doi:10.1016/j.renene.2018.07.023 2019

  24. [25]

    Unified engineering models for the per- formance and cost of ground-gen and fly-gen crosswind airborne wind energy systems.Renewable Energy, 162:893–907, 2020

    Filippo Trevisi, Mac Gaunaa, and Michael McWilliam. Unified engineering models for the per- formance and cost of ground-gen and fly-gen crosswind airborne wind energy systems.Renewable Energy, 162:893–907, 2020. ISSN 0960-1481. doi: 10.1016/j.renene.2020.07.129

    work page doi:10.1016/j.renene.2020.07.129 2020

  25. [26]

    The betz limit applied to airborne wind energy.Renewable Energy, 2018

    Marcelo De Lellis, Romeu Reginatto, Ramiro Saraiva, and Alexandre Trofino. The betz limit applied to airborne wind energy.Renewable Energy, 2018. doi: 10.1016/J.RENENE.2018.04.034

    work page doi:10.1016/j.renene.2018.04.034 2018

  26. [27]

    Filippo Trevisi, C. E. D. Riboldi, and Alessandro Croce. Refining the airborne wind en- ergy system power equations with a vortex wake model.Wind Energy Science, 2023. doi: 10.5194/WES-8-1639-2023

    work page doi:10.5194/wes-8-1639-2023 2023

  27. [28]

    Karakouzian, and Frederic Bourgault

    Mojtaba Kheiri, Samson Victor, Sina Rangriz, Mher M. Karakouzian, and Frederic Bourgault. Aerodynamic performance and wake flow of crosswind kite power systems.Energies, 15(7), 2022. ISSN 1996-1073. doi: 10.3390/en15072449

    work page doi:10.3390/en15072449 2022

  28. [29]

    Karakouzian, Mojtaba Kheiri, and Fr´ ed´ eric Bourgault

    Mher M. Karakouzian, Mojtaba Kheiri, and Fr´ ed´ eric Bourgault. A survey of two analytical wake models for crosswind kite power systems.Physics of Fluids, 34(9):097111, 09 2022. ISSN 1070-6631. doi: 10.1063/5.0102388

    work page doi:10.1063/5.0102388 2022

  29. [30]

    Roque, Lu´ ıs Tiago Paiva, Manuel C.R.M

    Lu´ ıs A.C. Roque, Lu´ ıs Tiago Paiva, Manuel C.R.M. Fernandes, Dalila B.M.M. Fontes, and Fer- nando A.C.C. Fontes. Layout optimization of an airborne wind energy farm for maximum power gen- eration.Energy Reports, 6:165–171, 2020. ISSN 2352-4847. doi: https://doi.org/10.1016/j.egyr.2019. 21 08.037. URLhttps://www.sciencedirect.com/science/article/pii/S23...

    work page doi:10.1016/j.egyr.2019 2020

  30. [31]

    Rui Carvalho da Costa, Lu´ ıs A. C. Roque, Lu´ ıs Tiago Paiva, Manuel C. R. M. Fernandes, Dalila B. M. M. Fontes, and Fernando A. C. C. Fontes. Airborne wind energy farms: Layout optimization combining nsga–ii and brkga. In Hossein Moosaei, Ilias Kotsireas, and Panos M. Pardalos, editors, Dynamics of Information Systems, pages 291–299, Cham, 2025. Springe...

    2025

  31. [32]

    da Costa, Manuel C

    Rui C. da Costa, Manuel C. R. M. Fernandes, Lu´ ıs A. C. Roque, Lu´ ıs Tiago Paiva, Dalila B. M. M. Fontes, and Fernando A. C. C. Fontes. Airborne wind energy farms layout: synchronization strategies for power smoothing.Optimization Letters, January 2026. ISSN 1862-4480. doi: 10.1007/ s11590-025-02273-7. URLhttps://doi.org/10.1007/s11590-025-02273-7

    work page doi:10.1007/s11590-025-02273-7 2026

  32. [33]

    Power curve modelling and scaling of fixed- wing ground-generation airborne wind energy systems.Wind energy science, 2024

    Rishikesh Joshi, Roland Schmehl, and Michiel Kruijff. Power curve modelling and scaling of fixed- wing ground-generation airborne wind energy systems.Wind energy science, 2024. doi: 10.5194/ WES-9-2195-2024

    2024

  33. [34]

    URLhttps://www.upwindresearchproject.com

    UPWIND Project, 2025. URLhttps://www.upwindresearchproject.com. Accessed on 2025-07- 11

    2025

  34. [35]

    Matlab version: 9.13.0 (r2023b), 2023

    The MathWorks Inc. Matlab version: 9.13.0 (r2023b), 2023. URLhttps://www.mathworks.com. Software. 22

    2023