Pith. sign in

REVIEW 3 major objections 5 minor 35 references

Towards Reliable Aerial Ground Vehicle Collaboration: An Integrated Planning and Autonomy Framework for Field Deployment

T0 review · 3 major / 5 minor · reviewed 2026-07-11 · grok-4.5

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

Pith's one-line read An integrated DRL planner, YAML mission API, vehicle autonomy stacks, and online rendezvous replanner make energy-limited UAV–UGV team missions reliable enough for outdoor field use.

desk verdict Solid systems integration paper: DRL routing + YAML API + RARP + outdoor demos work as a package, but the headline reliability number is sim-only and field landing is still manual. read the letter →

arxiv 2607.07350 v1 pith:VPDSU26T submitted 2026-07-07 cs.RO

Md Safwan Mondal , Luca Russo , James D. Humann , James M. Dotterweich , Pranav Bhounsule This is my paper
classification cs.RO
keywords UAV–UGVcollaborationcooperativeroutingdeepreinforcementlearningrendezvous-awarereplanningenergy-constrainedmissionplanningfieldautonomyYAMLAPIsearchandrescue
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

Long UAV missions that must visit many scattered areas of interest fail when batteries run out. This paper shows that pairing the UAV with a road-bound ground vehicle for mobile recharging can work in the field if planning and execution are tightly coupled. A deep-reinforcement-learning planner chooses both the order of aerial visits and the ground rendezvous points so total mission time is shorter than classic hierarchical heuristics. Those plans are turned into synchronized actions through a simple two-layer YAML interface, executed by standard PX4 and ROS 2 stacks, and protected online by a lightweight rendezvous-aware replanner that trims actions when wind or delays threaten the battery margin. Outdoor trials, including a search-and-rescue run that uses a vision-language model for hazard spotting, demonstrate that the full pipeline keeps sorties feasible under real disturbances.

What carries the argument

The Rendezvous-Aware Replanner (RARP): a lightweight online module that, on timing deviation, checks remaining endurance, rolls back the UAV action prefix, and re-selects a feasible road-network rendezvous so the team never violates the battery safety margin.

What would settle it

Repeat the same multi-sortie missions with fully autonomous precision landing on a moving pad under stronger wind or longer endurance budgets; if energy-margin violations rise above the reported 20 percent or rendezvous fail repeatedly, the reliability claim does not hold.

Watch

Extended reading notes

Core claim

A complete planning-to-execution stack—DRL cooperative routing that jointly selects UAV visit order and UGV rendezvous locations, a standardized two-layer YAML mission API, PX4/MAVSDK and ROS 2/Nav2 autonomy layers, and an online Rendezvous-Aware Replanner—enables reliable outdoor UAV–UGV missions under strict energy limits. The DRL planner reduces mission time relative to multi-echelon heuristics; the replanner cuts energy-margin violations from 83 percent to 20 percent; and field experiments on a 50 m square site confirm coordinated takeoff, survey, landing, mid-mission task insertion, and VLM-based hazard detection.

Load-bearing premise

The outdoor trials—small flat grass field, moderate wind, simulated recharging, and manual final landings—are enough to claim the full autonomy stack is reliable for field deployment.

Editorial extensions

If this is right

  • Mission planners can treat UGV road points as dynamic recharging nodes rather than fixed depots, shrinking total flight time for multi-AOI surveillance.
  • A YAML state/action contract lets heterogeneous PX4 and ROS 2 robots exchange plans without custom middleware.
  • Online rollback of UAV actions when endurance is threatened becomes a standard safety layer for any fuel-limited aerial–ground team.
  • Mid-mission AOI insertion and VLM hazard labeling can be folded into the same receding-horizon loop without redesigning the autonomy stacks.

Reading between the lines

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

  • The same YAML API and RARP logic could be reused for multi-UAV fleets sharing one or more ground rechargers if collision constraints are added to the decoder mask.
  • Predictive wind or latency models inside RARP would turn reactive trimming into proactive early returns, further lowering the residual 20 percent violation rate.
  • Vision-based relative localization for docking would remove the last manual step and open GPS-denied SAR sites.
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

3 major / 5 minor

Summary. The paper presents an integrated planning-and-autonomy framework for energy-constrained UAV–UGV cooperative routing. A DRL encoder–decoder policy jointly selects the UAV AOI visitation order and UGV rendezvous points on a road network, outperforming the Maini et al. hierarchical heuristic on mission time (Tables I–II). A two-layer YAML mission API maps planner outputs to synchronized action primitives; distributed stacks (PX4/MAVSDK for the UAV, ROS 2/Nav2 for the UGV) execute them; and a lightweight Rendezvous-Aware Replanner (RARP, Algorithm 1) trims sorties and reassigns rendezvous under timing drift. RARP is evaluated in Gazebo SITL (Table IV: energy-margin violations 83.33%→20.00%). Outdoor trials on a 50 m × 50 m field (six AOIs, five RNPs) demonstrate takeoff, survey, rendezvous, mid-mission AOI insertion, and a VLM-based SAR coverage scenario; final pad landing is performed manually.

Significance. If the claims hold, the work supplies a practical end-to-end bridge from cooperative routing to outdoor hardware that is still rare in the UAV–UGV literature: a learned joint planner, a minimal standardized mission API, platform-specific autonomy stacks, and an online safety layer, all exercised together in the field. The DRL vs. heuristic comparisons and the RARP simulation metrics are concrete and useful. Explicit credit is due for the full-stack integration, the receding-horizon replan-at-rendezvous design, and the outdoor demos (including dynamic task insertion and VLM hazard detection). The main limitation is that the strongest quantitative reliability number is simulation-only and that precision landing remains manual, so the transfer of “reliable field deployment” is only partially evidenced.

major comments (3)
  1. [Abstract, §V.C, Table IV, §VI] Abstract, §V.C and Table IV vs. §VI: The headline reliability claim (energy-margin violations reduced from 83.33% to 20.00%) is obtained exclusively in Gazebo SITL under injected wind, GPS noise and a 15% speed reduction. Section VI field results (Figs. 7–10) report qualitative trajectory/energy alignment and mid-mission AOI insertion on flat grass under 12–24 km/h wind, but never report RARP trigger counts, residual endurance violations, or the same energy-margin metric outdoors. The abstract and contribution list juxtapose the RARP number with “validated through outdoor field experiments,” which overstates what the field evidence currently supports. Either add outdoor RARP statistics (or an ablation with RARP disabled) or clearly separate the simulation reliability claim from the qualitative field demonstration and temper the abstract language.
  2. [§VI.B, Fig. 7c] §VI.B explicitly states that final precision landing on the 30 in × 30 in pad is performed manually; vision-based landing is validated only in Gazebo. Autonomous docking under outdoor wind and relative-state uncertainty is a load-bearing part of the claimed “complete autonomy stack” and of closed-loop energy management. The manuscript should either (i) report autonomous landing success rates in the field or (ii) rephrase the autonomy and reliability claims to acknowledge that recovery remains human-in-the-loop, and discuss the implications for the energy-margin guarantees that RARP is intended to protect.
  3. [§IV, Eq. (4), Tables I–II] §IV and Tables I–II: The DRL planner is the source of the mission-time gains, yet the architecture, reward, and training details are deferred to prior work [30], [31]. For a journal paper whose first contribution is this planner, the present manuscript needs enough self-contained description (state features, masking rules, cost definition used for REINFORCE, and how UGV road travel time is folded into the joint objective) that a reader can interpret the reported gaps without the earlier papers. At minimum, state the exact cost c used in Eq. (4) and confirm that the same endurance and recharging models are used for both DRL and the Maini heuristic.
minor comments (5)
  1. [Fig. 5] Fig. 5 caption and body: “RNP : Road netwok points” — typo (“network”).
  2. [§III, Algorithm 1] §III objective (1) and later text use both F_a and T_full / T_rem for endurance; a single consistent symbol would help.
  3. [Table III, Fig. 3] Table III and Fig. 3: action names differ slightly in spelling (allow_take_off_from_UGV vs. allow_takeoff_from_UGV). Align the YAML schema and the table.
  4. [§II.B] Related work §II.B is thorough; a short explicit comparison table (planning method, energy model, outdoor vs. indoor, landing autonomy) would make the claimed gap clearer.
  5. [Header] arXiv header and journal running head show 2026 dates; correct for the camera-ready version.

Circularity Check

1 steps flagged · score 1.0 of 10

No significant circularity: engineering integration paper whose performance claims rest on external baselines, simulation trials, and field demos rather than definitional or self-forced results.

  1. self citation load bearing [Section IV.A.3 / end of IV.A]
    "This section summarizes the key components of our DRL routing policy. Since the primary contribution of this paper lies in the integration and real-world deployment of UAV–UGV cooperation, we refer readers to our prior work [30], [31] for a more exhaustive treatment of the DRL policy design."

    The detailed encoder–decoder architecture and REINFORCE training are deferred to overlapping-author prior papers. This is a minor self-citation for implementation detail only; the present paper still re-evaluates the policy against an external heuristic baseline and reports independent metrics, so the citation is not load-bearing for the integration/RARP/field claims.

full rationale

This is a systems/integration paper, not a first-principles derivation. The DRL planner is trained via REINFORCE against a greedy baseline and evaluated against the independent hierarchical heuristic of Maini et al. [19] (Tables I–II); the reported mission-time gaps are empirical, not forced by construction. Self-citations [30],[31] supply architectural details of the attention policy but are not load-bearing for the paper’s new claims (YAML API, autonomy stacks, RARP Algorithm 1, outdoor energy-timeline alignment, mid-mission AOI insertion, VLM SAR demo). RARP’s 83.33%→20.00% reduction (Table IV) is measured under explicit simulated disturbances, not tautological. No equation equates a claimed prediction to a fitted input, no uniqueness theorem is imported to forbid alternatives, and no known empirical pattern is merely renamed. Minor self-citation of prior DRL design is normal and non-circular under the guidelines; the central results remain independently measured.

Assumptions & free parameters 5 free parameters · 5 assumptions · 3 invented entities

The central claim is systems-level: that this integrated stack works outdoors under energy limits. It rests on standard multi-robot modeling choices (fixed endurance, road graph, known speeds), hand-chosen safety and timing parameters (α, endurance budget, dwell/tolerance thresholds), and two engineered artifacts (YAML API, RARP) whose value is shown experimentally rather than derived. No new physical entities are postulated; the free parameters mainly set feasibility margins and training.

free parameters (5)
  • endurance safety factor α = 0.90 (default); 0.95 (relax)
    Default α=0.90 (effective 90 s of 100 s endurance); relaxed once to 0.95 in RARP Phase 3. Directly controls when sorties are declared infeasible and drives the 83%→20% metric.
  • nominal UAV flight endurance F_a / T_full = 100 s
    Set to 100 s in sim and field; mission feasibility and RARP remaining-time checks are defined relative to this budget.
  • fixed recharging service duration T_R = 30 s (field)
    Fixed dock time (30 s in field narrative) between sorties; affects total mission time and schedule synchronization.
  • DRL architecture and training hyperparameters = as listed in §IV.B
    Embedding dim 128, L=3 layers, M=8 heads, logit clip C_p=10, Adam lr 1e-4, decay 0.995, batch 256, 100 epochs, 5.12M instances—chosen by authors and determine reported planner quality.
  • execution tolerances (goal radius, dwell, control rates) = 0.3 m; 80 s risky threshold; etc.
    UGV 0.3 m goal tolerance; UAV geodesic error dwell; 20–50 Hz offboard rates; risky-sortie threshold 80 s—hand-set thresholds that affect success flags and replanner triggers.
assumptions (5)
  • domain assumption UAV flight endurance is a known fixed budget per sortie; energy is tracked as remaining flight time without a higher-fidelity battery model.
    Used throughout problem formulation (Eq. 1), planner fuel state, and RARP T_rem checks (§III, Algorithm 1).
  • domain assumption UGV motion is confined to a known road-network graph G with deterministic travel times along edges.
    Core constraint π_g ⊆ G in Eq. 1 and RARP shortest-path updates.
  • domain assumption Inter-agent UDP/TCP links and a central mission manager provide sufficiently timely state and completion flags for synchronized takeoff/landing and event-driven replanning.
    Communication stack §V.B.3; field setup assumes 30–50 Hz telemetry on 5 GHz Wi-Fi.
  • standard math REINFORCE with greedy baseline and Transformer encoder–decoder is a valid training method for the cooperative routing policy.
    §IV.A.3; standard policy-gradient practice; details deferred to authors’ prior work.
  • domain assumption Hierarchical multi-echelon heuristic of Maini et al. [19] is an appropriate baseline for mission-time comparison.
    Tables I–II comparisons; common in truck-drone literature but not an exact optimum.
invented entities (3)
  • Rendezvous-Aware Replanner (RARP)
    purpose: Online feasibility repair: check remaining endurance, rollback UAV action prefixes, reselect RNP rendezvous, or emergency land.
    Core safety contribution; Algorithm 1 and Table IV. Independent evidence is experimental (sim metrics), not external theory.
  • Two-layer YAML mission API (state YAML + action-plan YAML with ins_id synchronization)
    purpose: Standardize planner I/O and synchronized UAV/UGV action primitives across heterogeneous stacks.
    §V.A, Table III, Fig. 3. Engineering interface; value shown by deployment, not external prior formalization.
  • Integrated UAV–UGV autonomy architecture (central manager + PX4/MAVSDK + ROS 2/Nav2 action executors)
    purpose: Translate symbolic plans into closed-loop flight and ground navigation with takeoff/landing handshakes.
    §V.B; composition of known tools into a mission-specific pipeline claimed as the deployment bridge.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Reliable Aerial Ground Vehicle Collaboration: An Integrated Planning and Autonomy Framework for Field Deployment." pith.science (2026). https://pith.science/paper/VPDSU26T

@misc{pith2026260707350,
  author       = {Pith},
  title        = {Pith review of: Towards Reliable Aerial Ground Vehicle Collaboration: An Integrated Planning and Autonomy Framework for Field Deployment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VPDSU26T}},
  note         = {Machine review of arXiv:2607.07350}
}
read the original abstract

Limited flight endurance significantly restricts the operational range of unmanned aerial vehicles (UAVs) in long duration missions such as surveillance and inspection, where multiple spatially distributed Areas of Interest (AOIs) must be visited. These tasks require efficient routing determining the sequence of visits which directly impacts mission time, energy consumption, and overall feasibility. Pairing UAVs with unmanned ground vehicles (UGVs) for mobile recharging offers a promising solution, but introduces a tightly coupled cooperative routing problem involving UAV route planning, UGV road constrained movement, energy management, and rendezvous scheduling under uncertainty. In this work, we present an integrated planning and autonomy framework for reliable field deployment. We formulate the problem as an energy constrained cooperative routing task and solve it using a Deep Reinforcement Learning (DRL) based planner that jointly optimizes the UAV visitation sequence and rendezvous locations with the UGV, outperforming baseline heuristics in minimizing total mission time. To bridge the gap between planning and execution, we introduce a standardized two layer YAML based mission API that captures environment states and structures lightweight, synchronized action sequences. This framework is supported by a complete autonomy stack using PX4/MAVSDK for UAV control and ROS 2/Nav2 for UGV navigation. Furthermore, we propose a lightweight Rendezvous Aware Replanner (RARP) that operates online to handle environmental uncertainties, reducing energy margin violations from 83.33% to 20.00%. The full system is validated through outdoor field experiments, demonstrating robust cooperative navigation and adaptability in dynamic tasks, including a search and rescue scenario with vision language model (VLM) based hazard detection

Figures

Figures reproduced from arXiv: 2607.07350 by the authors.

Figure 1
Figure 1. Conceptual illustration of the UAV–UGV cooperative mission framework. The mission planner computes coordinated routes for an endurance-limited [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Mission planner architecture. The encoder embeds mission points using multi-head self-attention to capture spatial relationships among AOIs and [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Representative snippet of the mission-state YAML provided to the [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Integrated UAV–UGV autonomy and coordination architecture. (a) UGV autonomy stack based on ROS 2 and Nav2, showing multi-sensor state [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Illustration of the rendezvous-aware replanning (RARP) mechanism. (a) Timeline showing the initial sortie plan, replanned segments, and the final [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Hardware platforms used in the field deployment. The UAV executes [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Sequence of the field experiment demonstrating the three core tasks of the cooperative mission. The composite images illustrate: (a) The UAV [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Qualitative analysis of the experimental field trial. (a) Trajectories of the UGV and UAV, where the UGV follows the road network (solid blue) while [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 9
Figure 9. Figure 9: Dynamic mission execution with mid-mission task insertion. (a) [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Autonomous UAV–UGV search-and-rescue mission with vision–language-based hazard detection. (a) Coverage planning, where the survey region is [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

35 extracted references · 35 canonical work pages

  1. [30]

    An attention-aware deep reinforcement learning framework for uav-ugv collaborative route planning,

    M. S. Mondal, S. Ramasamy, J. D. Humann, J. M. Dotterweich, J.- P. F. Reddinger, M. A. Childers, and P. Bhounsule, “An attention-aware deep reinforcement learning framework for uav-ugv collaborative route planning,” in2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2024, pp. 13 687–13 694

    work page 2024

  2. [31]

    How to coordinate uavs and ugvs for efficient mission planning? optimizing energy-constrained cooperative routing with a drl framework,

    M. S. Mondal, S. Ramasamy, L. Russo, J. D. Humann, J. M. Dot- terweich, and P. Bhounsule, “How to coordinate uavs and ugvs for efficient mission planning? optimizing energy-constrained cooperative routing with a drl framework,” inProceedings of the Robotics: Science and Systems Conference (RSS), 2025

    work page 2025

  3. [1]

    Multi criteria decision analysis (mcda) of unmanned aerial vehicles (uavs) as a part of standard response to emergencies,

    S. Mehmood, S. Ahmed, A. S. Kristensen, and D. Ahsan, “Multi criteria decision analysis (mcda) of unmanned aerial vehicles (uavs) as a part of standard response to emergencies,” in4th international conference on green computing and engineering technologies. Gyancity International Publishers, 2018, p. 31

    work page 2018

  4. [2]

    Semi-autonomous hybrid ugv-uav system for search and rescue missions,

    C. John De Britto, R. Jothi Sri, and V . Sandhiya, “Semi-autonomous hybrid ugv-uav system for search and rescue missions,” inInternational Conference on Universal Threats in Expert Applications and Solutions. Springer, 2025, pp. 359–374

    work page 2025

  5. [3]

    Air to ground collaboration for energy-efficient path planning for ground robots,

    M. Wei and V . Isler, “Air to ground collaboration for energy-efficient path planning for ground robots,” in2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2019, pp. 1949–1954

    work page 2019

  6. [4]

    Devices, systems, and methods for automated monitoring enabling precision agriculture,

    J. Das, G. Cross, C. Qu, A. Makineni, P. Tokekar, Y . Mulgaonkar, and V . Kumar, “Devices, systems, and methods for automated monitoring enabling precision agriculture,” in2015 IEEE international conference on automation science and engineering (CASE). IEEE, 2015, pp. 462– 469

    work page 2015

  7. [5]

    The vehicle routing problem with drones: Several worst-case results,

    X. Wang, S. Poikonen, and B. Golden, “The vehicle routing problem with drones: Several worst-case results,”Optimization Letters, vol. 11, no. 4, pp. 679–697, 2017

    work page 2017

  8. [6]

    The flying sidekick traveling salesman problem: Optimization of drone-assisted parcel delivery,

    C. C. Murray and A. G. Chu, “The flying sidekick traveling salesman problem: Optimization of drone-assisted parcel delivery,”Transportation Research Part C: Emerging Technologies, vol. 54, pp. 86–109, 2015

    work page 2015

Show all 35 references
  1. [7]

    Hu and R

    X. Hu and R. H. Assaad, “The use of unmanned ground vehicles (mobile robots) and unmanned aerial vehicles (drones) in the civil infrastructure asset management sector: Applications, robotic platforms, sensors, and algorithms,”Expert Systems with Applications, vol. 232, p. 120897, 2023

    2023

  2. [8]

    Unmanned aerial and ground vehicle (uav- ugv) system prototype for civil infrastructure missions,

    B. Hament and P. Oh, “Unmanned aerial and ground vehicle (uav- ugv) system prototype for civil infrastructure missions,” in2018 IEEE international conference on consumer electronics (ICCE). IEEE, 2018, pp. 1–4

    2018

  3. [9]

    A cooperative uav/ugv platform for wildfire detection and fighting,

    C. Phan and H. H. Liu, “A cooperative uav/ugv platform for wildfire detection and fighting,” in2008 Asia Simulation Conference-7th Inter- national Conference on System Simulation and Scientific Computing. IEEE, 2008, pp. 494–498

    2008

  4. [10]

    Sensor planning for a symbiotic uav and ugv system for precision agriculture,

    P. Tokekar, J. Vander Hook, D. Mulla, and V . Isler, “Sensor planning for a symbiotic uav and ugv system for precision agriculture,”IEEE transactions on robotics, vol. 32, no. 6, pp. 1498–1511, 2016

    2016

  5. [11]

    A hybrid path planning method in unmanned air/ground vehicle (uav/ugv) cooperative systems,

    J. Li, G. Deng, C. Luo, Q. Lin, Q. Yan, and Z. Ming, “A hybrid path planning method in unmanned air/ground vehicle (uav/ugv) cooperative systems,”IEEE Transactions on Vehicular Technology, vol. 65, no. 12, pp. 9585–9596, 2016

    2016

  6. [12]

    Cooperative path planning for target tracking in urban environments using unmanned air and ground vehicles,

    H. Yu, K. Meier, M. Argyle, and R. W. Beard, “Cooperative path planning for target tracking in urban environments using unmanned air and ground vehicles,”IEEE/ASME transactions on mechatronics, vol. 20, no. 2, pp. 541–552, 2014

    2014

  7. [13]

    Online path planning of messenger uav in air–ground coordination over urban areas,

    Y . Wang, B. Xin, Y . Ding, B. He, and J. Chen, “Online path planning of messenger uav in air–ground coordination over urban areas,”IEEE Transactions on Vehicular Technology, vol. 74, no. 5, pp. 7135–7151, 2024

    2024

  8. [14]

    Aerial-ground robots collaborative 3d mapping in gnss-denied environments,

    Y . Yue, C. Zhao, Y . Wang, Y . Yang, and D. Wang, “Aerial-ground robots collaborative 3d mapping in gnss-denied environments,” in2022 International Conference on Robotics and Automation (ICRA). IEEE, 2022, pp. 10 041–10 047

    2022

  9. [15]

    Radiation search operations using scene understand- ing with autonomous uav and ugv,

    G. Christie, A. Shoemaker, K. Kochersberger, P. Tokekar, L. McLean, and A. Leonessa, “Radiation search operations using scene understand- ing with autonomous uav and ugv,”Journal of Field Robotics, vol. 34, no. 8, pp. 1450–1468, 2017

    2017

  10. [16]

    Cooperative routing for an air–ground vehicle team—exact algorithm, transformation method, and heuristics,

    S. G. Manyam, K. Sundar, and D. W. Casbeer, “Cooperative routing for an air–ground vehicle team—exact algorithm, transformation method, and heuristics,”IEEE Transactions on Automation Science and Engi- neering, vol. 17, no. 1, pp. 537–547, 2019

    2019

  11. [17]

    A two-echelon cooperated routing problem for a ground vehicle and its carried unmanned aerial vehicle,

    Z. Luo, Z. Liu, and J. Shi, “A two-echelon cooperated routing problem for a ground vehicle and its carried unmanned aerial vehicle,”Sensors, vol. 17, no. 5, p. 1144, 2017

    2017

  12. [18]

    On cooperation between a fuel constrained uav and a refueling ugv for large scale mapping applications,

    P. Maini and P. Sujit, “On cooperation between a fuel constrained uav and a refueling ugv for large scale mapping applications,” in2015 international conference on unmanned aircraft systems (ICUAS). IEEE, 2015, pp. 1370–1377

    2015

  13. [19]

    Cooperative aerial–ground vehicle route planning with fuel constraints for coverage applications,

    P. Maini, K. Sundar, M. Singh, S. Rathinam, and P. Sujit, “Cooperative aerial–ground vehicle route planning with fuel constraints for coverage applications,”IEEE Transactions on Aerospace and Electronic Systems, vol. 55, no. 6, pp. 3016–3028, 2019

    2019

  14. [20]

    Algorithms and experiments on routing of unmanned aerial vehicles with mobile recharging stations,

    K. Yu, A. K. Budhiraja, S. Buebel, and P. Tokekar, “Algorithms and experiments on routing of unmanned aerial vehicles with mobile recharging stations,”Journal of Field Robotics, vol. 36, no. 3, pp. 602– 616, 2019

    2019

  15. [21]

    Attention, learn to solve routing problems!

    W. Kool, H. Van Hoof, and M. Welling, “Attention, learn to solve routing problems!”arXiv preprint arXiv:1803.08475, 2018

    2018 arXiv

  16. [22]

    Reinforcement learning based truck-and-drone coordinated delivery,

    G. Wu, M. Fan, J. Shi, and Y . Feng, “Reinforcement learning based truck-and-drone coordinated delivery,”IEEE Transactions on Artificial Intelligence, vol. 4, no. 4, pp. 754–763, 2021

    2021

  17. [23]

    Control of multiple uavs for persistent surveillance: Algorithm and flight test results,

    N. Nigam, S. Bieniawski, I. Kroo, and J. Vian, “Control of multiple uavs for persistent surveillance: Algorithm and flight test results,”IEEE transactions on control systems technology, vol. 20, no. 5, pp. 1236– 1251, 2011

    2011

  18. [24]

    Persistent surveillance for unmanned aerial vehicles subject to charging and temporal logic constraints,

    K. Leahy, D. Zhou, C.-I. Vasile, K. Oikonomopoulos, M. Schwager, and C. Belta, “Persistent surveillance for unmanned aerial vehicles subject to charging and temporal logic constraints,”Autonomous Robots, vol. 40, no. 8, pp. 1363–1378, 2016

    2016

  19. [25]

    Collaborative mapping of an earthquake damaged building via ground and aerial robots,

    N. Michael, S. Shen, K. Mohta, V . Kumar, K. Nagatani, Y . Okada, S. Kiribayashi, K. Otake, K. Yoshida, K. Ohnoet al., “Collaborative mapping of an earthquake damaged building via ground and aerial robots,” inField and service robotics: results of the 8th international confere...

    2013

  20. [26]

    Collaborative navigation for flying and walking robots,

    P. Fankhauser, M. Bloesch, P. Kr ¨usi, R. Diethelm, M. Wermelinger, T. Schneider, M. Dymczyk, M. Hutter, and R. Siegwart, “Collaborative navigation for flying and walking robots,” in2016 IEEE/RSJ Interna- tional conference on intelligent robots and systems (IROS). IEEE, 2016, ...

    2016

  21. [27]

    Coordinated landing of a quadrotor on a skid-steered ground vehicle in the presence of time delays,

    J. M. Daly, Y . Ma, and S. L. Waslander, “Coordinated landing of a quadrotor on a skid-steered ground vehicle in the presence of time delays,”Autonomous Robots, vol. 38, no. 2, pp. 179–191, 2015

    2015

  22. [28]

    Aerial online mapping on- board system by real-time object detection for ugv path generation in unstructured outdoor environments,

    J. Lee, J. Lim, S. Pyo, and J. Lee, “Aerial online mapping on- board system by real-time object detection for ugv path generation in unstructured outdoor environments,”Journal of Field Robotics, vol. 40, no. 7, pp. 1754–1765, 2023

    2023

  23. [29]

    Air-ground col- laboration for language-specified missions in unknown environments,

    F. Cladera, Z. Ravichandran, J. Hughes, V . Murali, C. Nieto-Granda, M. A. Hsieh, G. J. Pappas, C. J. Taylor, and V . Kumar, “Air-ground col- laboration for language-specified missions in unknown environments,” arXiv preprint arXiv:2505.09108, 2025

    2025 arXiv

  24. [32]

    Terra: A path planning al- gorithm for cooperative ugv–uav exploration,

    F. Ropero, P. Munoz, and M. D. R-Moreno, “Terra: A path planning al- gorithm for cooperative ugv–uav exploration,”Engineering Applications of Artificial Intelligence, vol. 78, pp. 260–272, 2019

    2019

  25. [33]

    Data model and simulation for persistent mission planning with energy- sharing autonomous ground and air vehicles,

    J. Humann, S. C. Ortega, J. Glenn, J. L. Folsom, S. Ramasamy, M. S. Mondal, P. Bhounsule, J.-P. Reddinger, and J. Dotterweich, “Data model and simulation for persistent mission planning with energy- sharing autonomous ground and air vehicles,” in2024 Winter Simulation Conferen...

    2024

  26. [34]

    Smac hybrid-a* planner configuration guide,

    Nav2 Documentation Team, “Smac hybrid-a* planner configuration guide,” https://docs.nav2.org/configuration/packages/smac/ configuring-smac-hybrid.html, 2026, accessed: 2026-05-01

    2026

  27. [35]

    Openai api documentation,

    OpenAI, “Openai api documentation,” https://platform.openai.com/docs, 2024, gPT-4o model, accessed 2025

    2024

Pith tools

Reviewed July 11, 2026 · model on record in the stance chip above.