Pith. sign in

REVIEW 4 major objections 3 minor 29 references

Energy-Aware Wind-Resilient Routing for Truck-Assisted Multi-UAV Delivery under Wind Uncertainty

T0 review · 4 major / 3 minor · reviewed 2026-08-16 · deepseek-v4-flash

Pith's one-line read This paper proposes an online routing framework that keeps delivery drones energy-safe under uncertain wind by building conservative uncertainty margins into every edge cost and return-feasibility check.

desk verdict A solid engineering integration of wind-aware replanning with return-feasibility checks, but the central safety bound is unvalidated and the reported margin is partly by construction. read the letter →

arxiv 2608.11641 v1 pith:GRXHITBY submitted 2026-08-12 eess.SY cs.ROcs.SY

classification eess.SYcs.ROcs.SY
keywords truck-UAVdeliverywinduncertaintyenergy-awareroutingreturnfeasibilityrisk-sensitiveplanningtime-dependentenergygraphonlinereplanningUAVsafety
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 argues that routing decisions for truck-assisted multi-UAV delivery should be made against a time-varying energy graph whose edge costs include not only the current wind estimate but a conservative margin for how wrong that estimate might be. Its central proposal, EWR, replans each UAV's forward path online while checking at every step whether the UAV can still return to the truck or depot with the margin intact. The authors test this in simulated delivery graphs with wind data replayed from a public truck-UAV delivery dataset and from a real weather station, and they report higher mission success rates and lower wind-induced return-failure rates than shortest-path, energy-aware, and online-replanning baselines. The result matters because low-battery, high-wind conditions are precisely where static energy models underestimate headwind, crosswind, and return-feasibility risk.

What carries the argument

The central object is a time-dependent directed energy graph $G_t=(V,E,c_t)$ whose edge cost is $c_t(e)=E_t(e)+\lambda U_t(e)$: predicted traversal energy plus a wind-uncertainty penalty. Three components carry the argument: the bounded-error wind model $W(t,e)=\hat{W}_t(e)+\Delta W_t(e)$ with $\|\Delta W_t(e)\|\le \epsilon_W(t,e)$; the conservative reachability filter that rejects near-critical edges when the crosswind estimate plus its error bound exceeds a lateral control margin; and the adaptive safety margin $E_{\mathrm{safe}}(t,e)$ built from an emergency reserve, wind and battery state-of-charge uncertainty, and mission phase, with wind-induced energy uncertainty upper bounded by the Lipschitz product $L_E(e)\epsilon_W(t,e)$. The recursive use of these bounds makes every online replan an energy-feasibility check rather than merely a shortest-path update.

What would settle it

Replay the same weather-station wind log through EWR while recording, at each decision step, the actual deviation between the true wind and the delayed estimate; if that deviation exceeds the assumed $\epsilon_W(t,e)$ on any step that is followed by a return failure or a negative minimum margin, the conservative bound is not conservative. A cleaner version of the experiment inserts a short gust event exceeding $\epsilon_W$ into the log and checks whether EWR's return-failure rate rises to the level of Online-Replan or D* Lite.

Watch

Extended reading notes

Core claim

Under delayed, noisy wind observations, a UAV's safe-return guarantee cannot come from the current wind estimate alone. EWR's claim is that routing with an explicit uncertainty budget—a bounded confidence radius on wind error, a conservative power-model Lipschitz bound, and an adaptive safety margin—keeps the minimum residual energy margin positive throughout the flight, whereas replanning methods that trust the latest estimate let the margin go negative and end in return failures. In the paper's experiments this shows up as roughly 92% mission success at full battery and above 75% at half battery, an 86.5% success rate under real weather-station wind replay with a 2.8% return-failure rate, and per-step planning times under 160 ms at the largest tested graph size.

Load-bearing premise

The safety margin only works if the true wind never differs from the delayed estimate by more than the confidence radius the planner assumes, and the paper does not show how that radius is estimated or validated, while it explicitly acknowledges that sudden gusts are not modeled—so an unexpected gust bigger than the assumed bound would break the return-feasibility guarantee precisely where it matters.

Editorial extensions

If this is right

  • Adding an uncertainty margin to edge costs and an explicit return-feasibility check at each replan step is what separates EWR from reactive replanners in the paper's tests; without the margin, Online-Replan and D* Lite let the minimum energy margin go negative under the lower battery budget.
  • The reported per-step planning times below 160 ms at 352 nodes and 3486 edges suggest that the margin-carrying graph search remains fast enough for online use in medium-scale truck-UAV delivery.
  • A positive minimum energy margin, not just a higher success count, is the operational claim: EWR reports positive $M_{\min}$ under both battery budgets, while most baselines report near-zero or negative margins.
  • On real weather-station wind logs, EWR reports an 86.5% success rate and a 2.8% return-failure rate at full battery, versus 58-78% success and 8-20% failure for the baselines, indicating that the benefit survives realistic wind variability.

Reading between the lines

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

  • If the margin framework is sound, it can be layered on top of any existing replanner as a safety wrapper: the return-feasibility filter checks a candidate path before execution, so it could guard non-margin baselines without redesigning them.
  • A natural next step the paper does not take is to make the confidence radius $\epsilon_W(t,e)$ an online-estimated quantity, for example from the residual between forecast and observed wind, which would turn a fixed margin into an adaptive one.
  • Combining the margin approach with gust-prediction models would directly address the paper's stated limitation that sudden wind gusts are not modeled; the current worst-case bound only covers errors inside $\epsilon_W$.
  • The framework's metric $M_{\min}$ could double as a real-time alarm: a controller watching it cross zero has an immediate trigger for safe abort, which fits the paper's distinction between aborts and return failures.
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

4 major / 3 minor

Summary. The manuscript proposes Energy-Aware Wind-Resilient Routing (EWR) for truck-assisted multi-UAV delivery. The environment is modeled as a time-dependent directed energy graph whose edge costs are computed from delayed noisy wind estimates, a bounded-error model, an adaptive safety margin, and a crosswind reachability filter. EWR replans online while checking that each UAV can safely return to the truck or depot. Experiments on synthetic graphs with replayed ATS and ASOS wind logs compare EWR against six baselines, reporting higher mission success rates, lower return failure rates, positive minimum energy margins, and sub-160 ms planning times.

Significance. If the safety margins and uncertainty bounds are fully specified and validated, the framework would be a useful contribution: it directly targets energy-feasible routing under wind uncertainty, a real operational concern, and it evaluates against multiple baselines with two wind-data sources. The paper is honest about its limitations, including unmodeled gusts and battery aging. The main empirical direction—that uncertainty-aware replanning with return-feasibility checks improves success rates—is plausible and worth publishing once the missing model definitions and validation are supplied.

major comments (4)
  1. [III-B, Eq. (1), Eq. (9), Eq. (5)] The bounded-error model in Eq. (1) and the Lipschitz energy propagation in Eq. (9) are load-bearing but unspecified. The paper does not define how the confidence radius epsilon_W(t,e) is estimated from sensor data, how the Lipschitz bound L_E(e) is computed, or what functional forms P_parallel(t,e) and P_perp(t,e) take in Eq. (5). As a result, the conservative uncertainty margin E_safe + Delta E_W and the return-feasibility guarantee cannot be reproduced or tested. The conclusion's admission that "unforeseen sudden strong wind gusts" are not modeled further means that the guarantee holds only if the realized wind error stays inside epsilon_W, which is never validated against the replayed ASOS logs. Please specify these quantities and validate the bound empirically, or weaken the safety claims accordingly.
  2. [V-B, Eq. (12), Fig. 4(b)] The reported minimum energy margin M_min is defined by subtracting exactly the quantities that EWR enforces as constraints: E_safe(t,e_t) and Delta E_W(t,e_t). Consequently, a positive M_min is in part a bookkeeping consequence of the routing rule rather than an independent safety certificate. For baselines it is unclear whether E_safe and Delta E_W are set to zero or to the same values; without this, the comparison in Fig. 4(b) may be biased. Please report a raw minimum margin (E_rem - E_t - E_ret without EWR's own safety terms) and state the baseline convention.
  3. [V-C, Fig. 3(a)] Figure 3(a) reports mission success rates without error bars even though each setting uses 100 trials. Table III provides standard deviations for the real-wind-log setting, but the synthetic-setting figure does not; without variance information, the claimed improvements over Online-Replan and D* Lite cannot be assessed for statistical significance. Add error bars or confidence intervals to the figure.
  4. [V-A, VI] The experimental setup does not describe how the wind logs are mapped to graph edges, what values of epsilon_W are used in simulation, or how the "wind discretizations" mentioned in the conclusion are varied. Section V-A only says the logs are streamed to update edge costs, and no experiment in the paper varies a discretization. Provide these implementation details or remove the claim about wind discretizations.
minor comments (3)
  1. [Table II] The columns for edges and planning time are concatenated in Table II (e.g., "1863.8" reads as 186 edges and 3.8 ms). Add a column separator or reformat the table.
  2. [III-B] The notation for airspeed is inconsistent: Eq. (4) uses V_A while Eq. (5) uses hat V_G(t,e); the definitions in Section III-A should state the relationship between these quantities.
  3. [V-A] The description of the "public ATS-based truck-UAV delivery dataset" (Ref. [26]) is minimal; specify how the 100 rounds are generated, how customers and UAVs are sampled, and how task allocation is performed before EWR routing.

Circularity Check

1 steps flagged · score 4.0 of 10

EWR's positive M_min safety margin is partly definitional (Eq. 12 restates EWR's own filter terms), but the headline success-rate and return-failure comparison is an independent simulation outcome.

  1. self definitional [Section III-C and Section V-B, Eq. (12)]
    "EWR continuously evaluates segment energy consumption, return-to-home feasibility, and safety margins during mission execution. It not only considers the energy optimality of the current forward path but also explicitly constrains whether a UAV can safely return to the truck or depot. ... To quantify energy safety during execution, we further report the minimum energy margin: Mmin = min_t [ E_i^rem(t) − E_t(e_t) − E_t^ret(u_t) − E_safe(t,e_t) − ΔE_W(t,e_t) ]. A positive Mmin indicates that the UAV maintains return feasibility throughout the mission."

    The terms in M_min are exactly the quantities EWR uses for its safe-return filter: predicted traversal energy E_t, minimum return energy E_ret, the adaptive safety margin E_safe from Eq. (8), and the wind-induced energy uncertainty ΔE_W from Eq. (9). Thus, for any trajectory EWR accepts, M_min >= 0 is a restatement of the routing constraint rather than an independent measurement of safety. Reporting EWR's positive M_min in Fig. 4(b) as evidence of a 'robust energy safety buffer' therefore has a definitional component. The mission success rate and return-failure rate, by contrast, are objective counts of simulated outcomes and do not reduce to the algorithm's own definitions.

full rationale

The paper's central empirical comparison—mission success rate and return-failure rate against SP-NoWind, Energy-SP, Initial-Wind-SP, Online-Replan, D* Lite, and Greedy-Energy—is an independent simulation result on synthetic graphs with replayed wind logs; no fitted parameter is renamed as a prediction, and no load-bearing self-citation chain is present. The only circular element is the minimum energy margin metric: Eq. (12) is composed of the same E_safe and ΔE_W terms that EWR enforces during routing, so a positive M_min under EWR is partly a tautological consequence of the algorithm's own feasibility check. The admitted lack of gust modeling and the unvalidated ε_W bound are correctness and validity concerns, not circularity. Because the headline success/failure claims are not forced by construction, the paper is only partially circular in its safety-margin reporting.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The framework relies on several domain assumptions about wind-error bounds and power models that are stated but not derived or validated; the evaluated algorithm also contains at least seven hyper-parameters whose values are not disclosed.

free parameters (6)
  • λ = not reported
    Risk-sensitivity coefficient in edge cost Eq. (6); no value or tuning procedure given.
  • α_w, α_b, α_p = not reported
    Weights in adaptive safety margin Eq. (8); no values given.
  • β_W, β_τ, β_g = not reported
    Weights in uncertainty penalty Eq. (7); no values given.
  • ε_W(t,e) = not reported
    Confidence radius of wind estimation error; central to Eqs. (1), (4), (9) but no model or value supplied.
  • ρ_⊥ = 0.75
    Lateral control margin in reachability filter Eq. (4); fixed to 0.75 in Table I.
  • E_res = 10 Wh
    Emergency energy reserve in Eq. (8); fixed to 10 Wh in Table I.
assumptions (5)
  • domain assumption Wind estimation error is bounded by a known confidence radius: ||ΔW_t(e)|| ≤ ε_W(t,e).
    Used in Eqs. (1), (4), and (9) to justify the reachability filter and the energy uncertainty bound; no procedure for deriving ε_W from sensor noise is given.
  • domain assumption Edge energy consumption decomposes into additive power terms P0, P∥, P⊥, P_load (Eq. 5).
    The energy model is stated without deriving P∥ and P⊥; the safety margin depends on this model, so a wrong model would break the margin guarantee.
  • ad hoc to paper The Lipschitz bound L_E(e) in Eq. (9) can be computed for each edge.
    No formula or algorithm for L_E is supplied; it is invoked to upper-bound ΔE_W.
  • domain assumption Graph, wind logs, and simulation parameters reflect real truck-UAV operating conditions.
    The experiments use synthetic graphs with replayed wind logs; no validation against physical flights is provided.
  • standard math Dijkstra and D* Lite algorithms are correct.
    Used as baseline and inside EWR; standard, unremarkable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Energy-Aware Wind-Resilient Routing for Truck-Assisted Multi-UAV Delivery under Wind Uncertainty." pith.science (2026). https://pith.science/paper/GRXHITBY

@misc{pith2026260811641,
  author       = {Pith},
  title        = {Pith review of: Energy-Aware Wind-Resilient Routing for Truck-Assisted Multi-UAV Delivery under Wind Uncertainty},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GRXHITBY}},
  note         = {Machine review of arXiv:2608.11641}
}
read the original abstract

Energy feasibility under wind uncertainty is a critical safety issue for low-altitude air-ground delivery. In truck-UAV systems, UAVs complete assigned deliveries and safely return to a mobile truck or depot, while wind-induced propulsion costs vary online and are only partially observable. Existing routing methods often rely on static or deterministic energy models, which may underestimate headwind, crosswind, battery-voltage, and return-feasibility risks. This paper proposes Energy-Aware Wind-Resilient Routing (EWR), an online risk-sensitive planning framework for wind-aware and energy-safe UAV routing. The delivery environment is represented as a time-dependent directed energy graph whose edge costs are updated using delayed noisy wind estimates, payload states, and conservative uncertainty margins. Experiments using synthetic delivery graphs with replayed wind logs from a public truck-UAV delivery dataset show that EWR improves mission success rates and reduces wind-induced return failures.

Figures

Figures reproduced from arXiv: 2608.11641 by the authors.

Figure 1
Figure 1. Overview of the proposed framework. II. RELATED WORK A. Energy-Aware UAV Routing Recent studies introduced aerodynamic or empirically cal￾ibrated energy models for fixed-wing and rotary-wing UAVs, showing that minimum-distance routes are not necessarily minimum-energy routes [13], [14]. In drone-delivery sce￾narios, energy constraints have been incorporated into VRP, TSP-D, and truck-drone routing formulations, wher… view at source ↗
Figure 2
Figure 2. Time-varying routing visualization for the proposed wind-aware [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. (a) Return safety comparison under wind uncertainty. (b) Minimum [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 24 canonical work pages

  1. [1]

    Cooperative air-ground instant delivery by uavs and crowdsourced taxis,

    J. Gao, Q. Wang, X. Zhang, J. Shi, X. Zhao, Q. Han, and Y . Pan, “Cooperative air-ground instant delivery by uavs and crowdsourced taxis,” in2024 IEEE 40th International Conference on Data Engineering (ICDE), 2024, pp. 4153–4166

  2. [2]

    Hierarchical Reinforcement Learning for Cooperative Air-Ground Delivery in Urban System

    S. Lei, C. mei Ma, H. Wen, Y . Li, L. Chen, Q. Yang, F. Tsung, L. Chen, S. Ruan, and Y . Liang, “Hierarchical reinforcement learning for cooperative air-ground delivery in urban system,”ArXiv, vol. abs/2602.12913, 2026. [Online]. Available: https://api.semanticscholar.org/CorpusID:285607462

  3. [3]

    Trideliver: Cooperative air-ground instant delivery with uavs, couriers, and crowdsourced ground vehicles,

    J. Gao, Y . Pan, Q. Wang, W. Hou, Y . Deng, L. Jiang, and Y . Fang, “Trideliver: Cooperative air-ground instant delivery with uavs, couriers, and crowdsourced ground vehicles,” 2026. [Online]. Available: https://arxiv.org/abs/2604.09049

  4. [4]

    Wind field modeling for formation planning in multi-drone systems,

    M. Park and T.-C. Au, “Wind field modeling for formation planning in multi-drone systems,” in2024 IEEE International Conference on Robotics and Automation (ICRA), 2024, pp. 12 375–12 381

  5. [5]

    Energy-aware multi- uav coverage mission planning with optimal speed of flight,

    D. Datsko, F. Nekovar, R. Penicka, and M. Saska, “Energy-aware multi- uav coverage mission planning with optimal speed of flight,”IEEE Robotics and Automation Letters, vol. 9, no. 3, pp. 2893–2900, 2024

  6. [6]

    Physics-informed probabilistic learning of low-altitude urban wind from onboard motion data,

    M. Chen, “Physics-informed probabilistic learning of low-altitude urban wind from onboard motion data,” inUrbanAI: Harnessing Artificial Intelligence for Smart Cities, 2025. [Online]. Available: https://openreview.net/forum?id=wfYXDYtoYV

  7. [7]

    Learning to predict the wind for safe aerial vehicle planning,

    F. Achermann, N. R. J. Lawrance, R. Ranftl, A. Dosovitskiy, J. J. Chung, and R. Siegwart, “Learning to predict the wind for safe aerial vehicle planning,” in2019 International Conference on Robotics and Automation (ICRA), 2019, pp. 2311–2317

  8. [8]

    End-to-end thermal updraft detec- tion and estimation for autonomous soaring using temporal convolutional networks,

    C. Gall, W. Fichter, and A. Ahmad, “End-to-end thermal updraft detec- tion and estimation for autonomous soaring using temporal convolutional networks,” in2024 IEEE International Conference on Robotics and Automation (ICRA), 2024, pp. 17 875–17 881

Show all 29 references
  1. [9]

    Real-time path planning for uavs in windy environments without computational fluid dynamics,

    A. Shrivastava, S. Gupta, and Z. Obradovic, “Real-time path planning for uavs in windy environments without computational fluid dynamics,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 40, no. 22, 2026, pp. 18 540–18 548

  2. [10]

    Energy-optimized planning in non-uniform wind fields with fixed-wing aerial vehicles,

    Y . Duan, F. Achermann, J. Lim, and R. Siegwart, “Energy-optimized planning in non-uniform wind fields with fixed-wing aerial vehicles,” in2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2024, pp. 3116–3122

  3. [11]

    Perception-aware cooperative path planning for multi-uav systems in urban wind fields via deep reinforcement learning,

    J. Ding, L. Wang, S. Jin, and D. Wang, “Perception-aware cooperative path planning for multi-uav systems in urban wind fields via deep reinforcement learning,”Sensors, vol. 26, no. 10, p. 2960, 2026

  4. [12]

    Time-optimal path planning in a constant wind for uncrewed aerial vehicles using dubins set classification,

    B. Moon, S. Sachdev, J. Yuan, and S. Scherer, “Time-optimal path planning in a constant wind for uncrewed aerial vehicles using dubins set classification,”IEEE Robotics and Automation Letters, vol. 9, no. 3, pp. 2176–2183, 2024

  5. [13]

    A power consumption model for multi-rotor small unmanned aircraft systems,

    Z. Liu, R. Sengupta, and A. Kurzhanskiy, “A power consumption model for multi-rotor small unmanned aircraft systems,” in2017 international conference on unmanned aircraft systems (ICUAS). IEEE, 2017, pp. 310–315

  6. [14]

    Comprehensive energy consumption model for unmanned aerial vehi- cles, based on empirical studies of battery performance,

    H. V . Abeywickrama, B. A. Jayawickrama, Y . He, and E. Dutkiewicz, “Comprehensive energy consumption model for unmanned aerial vehi- cles, based on empirical studies of battery performance,”IEEE access, vol. 6, pp. 58 383–58 394, 2018

  7. [15]

    Vehicle routing problems for drone delivery,

    K. Dorling, J. Heinrichs, G. G. Messier, and S. Magierowski, “Vehicle routing problems for drone delivery,”IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 47, no. 1, pp. 70–85, 2016

  8. [16]

    Energy-aware uav path planning with adaptive speed,

    J. Diller and Q. Han, “Energy-aware uav path planning with adaptive speed,” inProceedings of the 2023 International Conference on Au- tonomous Agents and Multiagent Systems, 2023, pp. 923–931

  9. [17]

    Uav coverage path planning under varying power constraints using deep reinforcement learning,

    M. Theile, H. Bayerlein, R. Nai, D. Gesbert, and M. Caccamo, “Uav coverage path planning under varying power constraints using deep reinforcement learning,” in2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2020, pp. 1444–1449

  10. [18]

    Energy-aware planning-scheduling for autonomous aerial robots,

    A. Seewald, H. G. de Marina, H. S. Midtiby, and U. P. Schultz, “Energy-aware planning-scheduling for autonomous aerial robots,” in 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2022, pp. 2946–2953

  11. [19]

    Reliable lifelong planning a*: Technique for re-optimizing reliable shortest paths when travel time distribution updating,

    W. Teng and B. Y . Chen, “Reliable lifelong planning a*: Technique for re-optimizing reliable shortest paths when travel time distribution updating,”Transportation Research Part E: Logistics and Transportation Review, vol. 188, p. 103635, 2024

  12. [20]

    D* lite with reset: Improved version of d* lite for complex environment,

    A. T. Le, M. Q. Bui, T. D. Le, and N. Peter, “D* lite with reset: Improved version of d* lite for complex environment,” in2017 First IEEE International Conference on Robotic Computing (IRC). IEEE, 2017, pp. 160–163

  13. [21]

    Safe path planning based on anytime dynamic a-star algorithm for mobile robots in indoor environment,

    W.-L. Yeh, M. A. Haq, and S.-J. Ruan, “Safe path planning based on anytime dynamic a-star algorithm for mobile robots in indoor environment,” in2025 International Seminar on Intelligent Technology and Its Applications (ISITIA). IEEE, 2025, pp. 59–63

  14. [22]

    Dynamic shortest path and transitive closure algorithms: A survey,

    D. P. Martin, “Dynamic shortest path and transitive closure algorithms: A survey,”arXiv preprint arXiv:1709.00553, 2017

  15. [23]

    Comparative analysis of uav path planning algorithms for efficient navigation in urban 3d environ- ments,

    H. Cheriet, K. K. Badra, and C. Samira, “Comparative analysis of uav path planning algorithms for efficient navigation in urban 3d environ- ments,” in2024 International Conference of the African Federation of Operational Research Societies (AFROS). IEEE, 2024, pp. 1–8

  16. [24]

    Shortest path planning for fleets of delivery drones: Evaluating the benefit of including a temporal dimension in the routing network,

    C. Nolan and P. Cuffe, “Shortest path planning for fleets of delivery drones: Evaluating the benefit of including a temporal dimension in the routing network,” in2025 35th Irish Signals and Systems Conference (ISSC), 2025, pp. 1–6

  17. [25]

    Koenig and M

    S. Koenig and M. Likhachev, “D*lite,” inEighteenth National Con- ference on Artificial Intelligence. USA: American Association for Artificial Intelligence, 2002, p. 476–483

  18. [26]

    Delivery with uavs: a simulated dataset via ats,

    G. Rigoni, C. M. Pinotti, Bhumika, D. Das, and S. K. Das, “Delivery with uavs: a simulated dataset via ats,” in2022 IEEE 95th Vehicular Technology Conference:(VTC2022-Spring). IEEE, 2022, pp. 1–6

  19. [27]

    Empirical power consumption model for uavs,

    H. V . Abeywickrama, B. A. Jayawickrama, Y . He, and E. Dutkiewicz, “Empirical power consumption model for uavs,” in2018 IEEE 88th Vehicular Technology Conference (VTC-Fall), 2018, pp. 1–5

  20. [28]

    1-Minute Page 1 Surface Weather Observations from the Automated Surface Observing System Network,

    NOAA National Centers for Environmental Information, “1-Minute Page 1 Surface Weather Observations from the Automated Surface Observing System Network,” https://www.ncei.noaa.gov/access/metadata/landing- page/bin/iso?id=gov.noaa.ncdc:C00386, accessed: 2026-06-19

  21. [29]

    Iowa Environmen- tal Mesonet: Download ASOS/AWOS/METAR Data,

    Iowa Environmental Mesonet, “Iowa Environmen- tal Mesonet: Download ASOS/AWOS/METAR Data,” https://mesonet.agron.iastate.edu/request/download.phtml, accessed: 2026-06-19

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.