Long-Term Mapping of the Douro River Plume with Multi-Agent Reinforcement Learning

George J. Pappas; Jo\~ao Borges de Sousa; Milad Mesbahi; Nicol\`o Dal Fabbro; Renato Mendes

REVIEW 4 major objections 6 minor 32 references

Long-Term Mapping of the Douro River Plume with Multi-Agent Reinforcement Learning

T0 review · 4 major / 6 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read A learned multi-agent policy maps the Douro river plume for days, cutting error roughly in half while more than doubling fleet endurance when scaled from 3 to 6 vehicles.

desk verdict The multi-day, energy-aware plume mapping formulation is genuinely useful, but the paper's own Table I contradicts the central 'consistently outperforms' claim. read the letter →

arxiv 2510.03534 v5 pith:AAQZPR6A submitted 2025-10-03 cs.MA cs.LGcs.SYeess.SYstat.ML

Nicolò Dal Fabbro , Milad Mesbahi , Renato Mendes , João Borges de Sousa , George J. Pappas This is my paper · ORCID

classification cs.MAcs.LGcs.SYeess.SYstat.ML

keywords riverplumemappingmulti-agentreinforcementlearningautonomousunderwatervehiclesGaussianprocessregressionlong-termenvironmentalmonitoringDelft3Dsimulationenergy-awarenavigationDouro

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

The paper tries to establish that a small fleet of energy-limited AUVs can keep a useful, low-error map of a fast-moving river plume for multiple days if their headings and speeds are chosen by a central learned controller rather than by fixed coverage patterns or myopic uncertainty sampling. It claims that a multi-agent deep reinforcement learning policy, fed with Gaussian-process estimates of the salinity field, consistently beats both physically unrealizable uniform-random sampling and the best published front-tracking and Voronoi coverage baselines. It further claims that increasing the fleet from 3 to 6 vehicles can more than double mission endurance while improving accuracy, because the policy learns when the slower cruise speed suffices. If true, this is a path to practical long-term autonomous monitoring of dynamic coastal features.

What carries the argument

The load-bearing component is the multi-head Q-network: a shared state-value head feeds two advantage heads, one over eight directions and one over two speed levels, so the server can command heading and speed separately for each AUV. Around it sits a spatiotemporal Gaussian process whose kernel is separable in space (exponential) and time (linear decay plus a 12.5-hour tidal oscillation fitted to historical data), and a reward that adds individual credit to global MSE to fix the credit-assignment problem. Together they convert raw salinity measurements into commands over 30-minute slots with only intermittent, roughly 160-byte uplinks.

What would settle it

Compare the learned policy's real-world salinity maps against independent CTD casts during a high-discharge Douro event. If the map MSE is no better than a simple lawnmower survey, or the fleet runs out of energy far sooner than the simulated 3-6 day missions, the central claim does not transfer to the ocean.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that long-horizon plume mapping reduces to a sequential decision problem that a centralized multi-head Q-network can learn to solve, provided the state encodes a spatiotemporal Gaussian-process regression map plus recent trajectory traces and wind, and the reward mixes global MSE, a salinity-contrast term, a fleet-level speed penalty, and per-agent credit for sampling where the previous estimate was wrong. The authors show in Delft3D simulations that this learned policy tracks the Douro plume over days, generalizes to months and years not used in training, and outperforms single-agent EIBV front tracking and multi-agent Voronoi coverage, with fl

Load-bearing premise

The entire result rests on the Delft3D simulation being a faithful stand-in for the real Douro plume; every policy and kernel is trained and tested inside that model, with no at-sea validation.

Editorial extensions

If this is right

Even a single AUV with this policy keeps the plume in view over days, whereas the EIBV front-tracking benchmark loses track over long horizons.
With 3 or 6 vehicles, the learned policy produces roughly half the mapping MSE of adaptive Voronoi coverage on the March 2018 test month.
Raising the fleet from 3 to 6 vehicles can more than double mission endurance, in addition to improving accuracy.
Varying one reward weight traces an endurance-accuracy frontier, so operators can dial in the desired trade-off.
Generalization across unseen months and years (2016-2018) suggests the policy captures persistent dynamics rather than overfitting one season.

Reading between the lines

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

A direct test the authors do not report: deploy the same policy on real LAUVs during a Douro high-flow event and compare CTD-verified maps against the simulated MSE; that would separate simulator generalization from real-ocean generalization.
Because the method is agnostic to the measured scalar, the same server-AUV loop could map temperature, turbidity, or dye plumes, provided the GP kernel is refit to the target process's correlation scale and period.
The reward decomposition (global MSE + individual credit + fleet speed penalty) is a reusable recipe for other multi-vehicle monitoring tasks where energy and communication are the binding constraints.
The fitted tidal-period kernel hints that encoding known environmental cycles into the surrogate is a cheap way to improve long-horizon estimates; that likely transfers to any tidally forced coastal process.

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.

Desk Editor's Note

read the letter

Two things you should know. First, this is a genuinely useful system: spatiotemporal GPR combined with a multi-head DQN to control a fleet of AUVs over multiple days, with an explicit energy-reward term. The multi-day, endurance-aware setting is new for plume mapping, and the Delft3D simulations are a credible testbed. Second, the headline claim outruns the evidence. The paper says it 'consistently outperforms' baselines, but Table I shows the Voronoi benchmark [32] beating it on October 2017 with N=6 (2.04 vs 2.79 MSE). That's a direct contradiction of a central claim, not a minor wrinkle.

What it does well: the GP kernel is fit to historical data, including a tidal-period term, and the reward function splits global MSE and per-agent credit, which addresses a real credit-assignment issue. The comparisons against an 'ideal' EIBV baseline [18] and a Voronoi benchmark [32] are appropriate, and in most tested months the method is better. The energy-speed tradeoff, with two cruise speeds and a fleet-level penalty, is a sensible way to handle endurance.

Where it's soft: no error bars or seed counts on any RL result, so we can't tell if the reported gaps are meaningful. The baselines are 'empirically tuned' and the best results shown, which is test-set tuning. The evaluation protocol is underspecified: no fixed mission horizon for MSE, and endurance is reported only for the proposed method, not the baselines. The endurance-doubling claim isn't visible in Table I—no row has N=6 endurance as double the N=3 value, and March 2018 actually decreases. The simulator-fidelity issue is real but secondary; every result is internal to Delft3D, so generalization is within the model, not to the physical ocean.

The paper is not weak in conception. The formulation is valuable, and the method plausibly works. But the claims as written are stronger than the data. A revision with seed statistics, a fixed evaluation protocol, and a softened or properly supported 'consistently outperforms' claim could make this a solid contribution.

For you: this is worth bringing to a reading group if you work in marine robotics or adaptive sampling, mainly to discuss what evidence is needed for 'consistent' improvement. I wouldn't cite it until the claims are backed up. But I'd send it to serious peer review—the core idea deserves referee time, and the flaws are fixable.

Referee Report

4 major / 6 minor

Summary. The paper addresses long-term mapping of the Douro River plume with a fleet of AUVs. It formulates the problem as minimizing time-averaged MSE over a grid subject to trajectory constraints and ocean-current dynamics (Sec. II). The proposed solution combines a GPR estimator with a separable spatiotemporal kernel fit to historical Delft3D data (Sec. III-A) and a multi-head DQN that selects direction and speed for each AUV (Sec. III-B). A central server communicates intermittently with the AUVs. Experiments in the Delft3D simulator compare against uniform sampling, ideal rotations, an ideal EIBV benchmark [18], and an adaptive Voronoi benchmark [32] (Sec. IV). The authors claim consistent outperformance, generalization across unseen months/years, and endurance gains from scaling the fleet.

Significance. If substantiated, the method would be a meaningful contribution to adaptive sampling and multi-agent reinforcement learning for environmental monitoring. The architecture is sensible, the effort to fit the kernel to empirical correlations is valuable, and the comparison against EIBV and Voronoi baselines is commendable. However, the present evidence is not sufficient to support the central claims: one cell of Table I directly contradicts 'consistently outperforms', no error bars or seed statistics are provided, and the endurance-doubling example is not reported with numeric data. These are correctable with additional experiments and more careful phrasing.

major comments (4)

[Section IV, Table I] In the October 2017 row with N=6, the proposed method's MSE is 2.79 while the Voronoi baseline [32] achieves 2.04. This directly contradicts Contribution (ii), 'consistently outperforms baselines and benchmarks in both single- and multi-agent settings.' Since Table I is the primary evidence for cross-season generalization, either the claim must be weakened or the table must be augmented with repeated-seed statistics and a statistical test showing the gap is within noise.
[Section IV, Multi-agent performance] No error bars, seed counts, or variance information are provided for any RL result. Figures 9 and 10 report box plots, but the underlying distributions, numbers of runs, and training seeds are unspecified. Without this, the reported differences (including the 2.79 vs 2.04 gap) cannot be distinguished from stochastic variation. The evaluation protocol should be described: number of training seeds, number of evaluation episodes, how MSE is aggregated over grid and time, and whether endurance values in Table I come from a single run.
[Section IV, Fig. 11 and Table I] The contribution states that scaling from 3 to 6 vehicles more than doubles mission lifetime. Table I (η2=50) shows no such doubling; for example, Oct ’17 endurance goes from 13.0 to 15.6 days, and Mar ’18 endurance decreases from 3.5 to 3.1 days. The only support is the Fig. 11 caption. The authors should report the exact configuration (η2 value, month) where the doubling occurs and provide the numeric data behind that figure, otherwise the endurance-scaling claim is unsupported.
[Section IV, Baselines] The tuning of baselines is described only qualitatively: 'We empirically tune the exploration-exploitation hyperparameters in [32]' and the threshold for [18] is selected as best from {28,30,32,34}. To ensure fair comparison, the selected hyperparameter values and the criterion for 'best' should be reported. This is load-bearing because the outperformance claims in Contributions (ii) and (iii) rest on these comparisons.

minor comments (6)

[Table I] The text near Table I contains garbled strings such as '/uni00000015/uni00000016/...' which appear to be rendering artifacts. These must be removed or fixed before publication.
[Table I] Endurance columns are reported only for the proposed method, not for the baselines, so the reader cannot evaluate the endurance comparison. Please include baseline endurance or state explicitly that baselines were run with a fixed mission duration.
[Section III-A] The memory window M is introduced but its value (24 frames) appears only in the baselines description. Specify M for all experiments in the training setup.
[Section III-A, Eq. (13)] The temporal kernel parameter T0 is described as the tidal period. Clarify whether T0 is fixed a priori or fitted to the data, and how its value is determined.
[Abstract and Contributions] The word 'consistently' in the abstract and Contributions overstates the evidence. Consider hedging with 'in the majority of test cases' or similar, until statistical evidence is available.
[Section IV] The paper should explicitly remind readers in the abstract and conclusion that all results are internal to the Delft3D simulator and that no in-situ validation is presented; the current phrasing 'generalizes to unseen conditions' is limited to the simulator.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the central RL/GPR pipeline is trained and evaluated on disjoint simulator months against external baselines; the paper's self-citations are domain tooling, not load-bearing.

full rationale

Walking the derivation chain: the objective (7) defines long-term mapping as minimizing MSE over evaluation grid G; the GPR estimator (10)-(13) is fit to historical Delft3D correlations; the multi-head DQN policy is trained on 2018 months and evaluated on unseen months/years in Table I. No equation reduces the reported MSE or endurance to the fitted parameters by construction. The reward (18) uses ground-truth MSE during training, but the evaluation metric (8) is independent and the comparisons against 'ideal [18]' and [32] are external benchmarks, not algebraic consequences of the fitted kernel or reward weights. Self-citations [5], [11], [24] concern LAUV specifications and Delft3D configuration, not the claimed algorithmic gain; no uniqueness theorem or ansatz is imported from the authors' prior work. Per the reviewing rule, I flag that Table I itself undermines 'consistently outperforms' (Oct '17, N=6: [32] MSE 2.04 vs Ours 2.79) and that no visible row supports the 'more than doubling' endurance claim; however, these are evidence-quality/correctness issues, not circularity, and therefore do not raise the circularity score.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on a set of engineering and modeling choices: the salinity field is a GP with a fitted separable kernel; the AUV model is a point mass with ocean flow; the Delft3D simulation is a valid proxy; the reward function is hand-tuned. There are no invented physical entities. The free parameters include GP kernel hyperparameters, reward weights, and the memory window.

free parameters (4)

GP kernel hyperparameters (λ, ℓ, β0, β1, β2, T0) = not reported (fit to Delft3D historical correlations)
Eqs. (12)-(13), Sec. III-A; set by matching empirical correlations from historical Delft3D data. Affects the GPR estimate used in state and reward.
Reward weights η0, η1, η2, η3 = η2 varied (values 0 and 50 shown); others not reported
Eq. (18); balance global MSE, salinity contrast, and fleet speed; η2 is the endurance-MSE tradeoff knob.
Memory window M = not reported
Sec. III-A; only the most recent M time slots are retained in the GPR; directly affects estimator accuracy and the state image.
Initial AUV deployments (positions, headings, speeds) = predetermined
Algorithm 1 line 2; initial conditions, not fitted, but they influence the trajectories and thus the reported MSE.

assumptions (5)

domain assumption The salinity field f(x,t) is a Gaussian process with a separable spatiotemporal kernel of the form (11)-(13).
Sec. III-A (Eq. 9-13); this defines the estimator. If the kernel is misspecified, the GP estimates and thus state and reward are biased.
domain assumption AUVs move according to point-mass dynamics x_dot = u + c(x,t), with ocean flow c from the Delft3D model.
Sec. II (Eq. 3); the entire trajectory prediction depends on this model and on the fidelity of the flow field.
domain assumption The Delft3D numerical model reproduces the real Douro plume with sufficient fidelity for the trained policy to transfer to the ocean.
Sec. I and IV; all training and evaluation is in Delft3D; no real-world validation is provided.
ad hoc to paper The reward function (18) with the selected weights produces a policy that solves the long-term mapping problem (7).
Sec. III-B; hand-designed with multiple terms; no theoretical guarantee, only empirical tuning.
standard math Centralized MARL with full communication is a valid approximation for the intermittent-communication system described.
Sec. III-B; they adopt a decentralized-with-full-communication training scheme [29]; this ignores communication dropouts between slots.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Long-Term Mapping of the Douro River Plume with Multi-Agent Reinforcement Learning." pith.science (2026). https://pith.science/paper/AAQZPR6A

@misc{pith2026251003534,
  author       = {Pith},
  title        = {Pith review of: Long-Term Mapping of the Douro River Plume with Multi-Agent Reinforcement Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AAQZPR6A}},
  note         = {Machine review of arXiv:2510.03534}
}

read the original abstract

We study the problem of long-term (multiple days) mapping of a river plume using multiple autonomous underwater vehicles (AUVs), focusing on the Douro river representative use-case. We propose an energy - and communication - efficient multi-agent reinforcement learning approach in which a central coordinator intermittently communicates with the AUVs, collecting measurements and issuing commands. Our approach integrates spatiotemporal Gaussian process regression (GPR) with a multi-head Q-network controller that regulates direction and speed for each AUV. Simulations using the Delft3D ocean model demonstrate that our method consistently outperforms both single- and multi-agent benchmarks, with scaling the number of agents both improving mean squared error (MSE) and operational endurance. In some instances, our algorithm demonstrates that doubling the number of AUVs can more than double endurance while maintaining or improving accuracy, underscoring the benefits of multi-agent coordination. Our learned policies generalize across unseen seasonal regimes over different months and years, demonstrating promise for future developments of data-driven long-term monitoring of dynamic plume environments.

Figures

Figures reproduced from arXiv: 2510.03534 by the authors.

**Figure 2.** Illustrative example of 10 hours of spatiotemporal evolution of the Douro river plume and AUV mobility (red trajectory), during March 2, 2018. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Salinity map of the Douro river plume (on the left) and a [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (6 more)

**Figure 5.** Figure 5: System architecture. results of our simulation study. LAUVs. The sensing agents considered in this work are light autonomous underwater vehicles (LAUVs) [5]. Each LAUV is equipped with a CTD probe (conductivity, temperature, depth) to measure salinity, navigation sens…

**Figure 6.** Figure 6: Temporal and spatial kernels h(τ) and Ks(x, x′ ) fitting the corresponding empirical correlations. through a simple, interpretable kernel function; (ii) it is highly flexible and reliable when in need of interpolating the sparse and highly irregular samples MN k produc…

**Figure 8.** Figure 8: Comparison of MSE over time for the two baselines (uniform and [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]

**Figure 9.** Figure 9: Example of mapping performance when using a [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]

**Figure 10.** Figure 10: Multi-agent performance for the test month of March 2018. “Unif” [PITH_FULL_IMAGE:figures/full_fig_p007_10.png]

**Figure 12.** Figure 12: Example of the multi-agent control performance of our solution with [PITH_FULL_IMAGE:figures/full_fig_p008_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 3 linked inside Pith

[32]

Dis- tributed coverage control for time-varying spatial processes,

F. Pratissoli, M. Mantovani, A. Prorok, and L. Sabattini, “Dis- tributed coverage control for time-varying spatial processes,” IEEE Transactions on Robotics, 2025

2025
[18]

Efficient 3d real-time adaptive auv sampling of a river plume front,

M. O. Berild, Y . Ge, J. Eidsvik, G.-A. Fuglstad, and I. Ellingsen, “Efficient 3d real-time adaptive auv sampling of a river plume front,” Frontiers in Marine Science, vol. V olume 10 - 2023, 2024

2023
[1]

Ai-driven marine robotics: Emerging trends in underwater perception and ecosystem monitoring,

S. Raine and T. Fischer, “Ai-driven marine robotics: Emerging trends in underwater perception and ecosystem monitoring,” arXiv preprint arXiv:2509.01878, 2025

arXiv 2025
[2]

The use of emerging autonomous technologies for ocean monitoring: insights and legal challenges,

A. Pastra, T. Johansson, J. Soares, and F. E. Muller-Karger, “The use of emerging autonomous technologies for ocean monitoring: insights and legal challenges,” Frontiers in Marine Science, vol. 12, p. 1561737, 2025

2025
[3]

Observation of a turbid plume using modis imagery: The case of douro estuary (portugal),

R. Mendes, N. Vaz, D. Fern ´andez-N´ovoa, J. Da Silva, M. Decastro, M. G´omez-Gesteira, and J. Dias, “Observation of a turbid plume using modis imagery: The case of douro estuary (portugal),” Remote sensing of environment, vol. 154, pp. 127–138, 2014

2014
[4]

Mixing and transport in coastal river plumes,

A. R. Horner-Devine, R. D. Hetland, and D. G. MacDonald, “Mixing and transport in coastal river plumes,” Annual Review of Fluid Mechanics, vol. 47, no. 1, pp. 569–594, 2015

2015
[5]

3D Tracking of a River Plume Front with an AUV,

D. Teixeira, J. B. d. Sousa, R. Mendes, and J. Fonseca, “3D Tracking of a River Plume Front with an AUV,” in OCEANS 2021: San Diego – Porto, 2021

2021
[6]

Formation control of multiple autonomous underwater vehicles: a review,

T. Yan, Z. Xu, S. X. Yang, and S. A. Gadsden, “Formation control of multiple autonomous underwater vehicles: a review,” Intelligence & Robotics, vol. 3, no. 1, 2023

2023

Show all 32 references

[7]

The douro estuary (portugal): a mesotidal salt wedge,

M. E. Vieira and A. A. Bordalo, “The douro estuary (portugal): a mesotidal salt wedge,” Oceanologica Acta, vol. 23, no. 5, pp. 585– 594, 2000

2000
[8]

Drivers of spatio-temporal patterns of salinity in spanish rivers: a nationwide assessment,

E. Est ´evez, T. Rodriguez-Castillo, A. M. Gonz ´alez-Ferreras, M. Ca ˜nedo-Arg¨uelles, and J. Barquin, “Drivers of spatio-temporal patterns of salinity in spanish rivers: a nationwide assessment,” Philosophical Transactions of the Royal Society B, vol. 374, no. 1764, p. 20180...

2019
[9]

Characterization of river plume dynamics for a better water quality management,

R. Friedland, T. Neumann, S. Piehl, H. Radtke, and G. Schernewski, “Characterization of river plume dynamics for a better water quality management,” Frontiers in Marine Science, vol. 12, p. 1617660, 2025

2025
[10]

Validation Document Delft3D-FLOW; a software system for 3D flow simulations,

H. Gerritsen, E. De Goede, F. Platzek, M. Genseberger, J. Van Kester, and R. Uittenbogaard, “Validation Document Delft3D-FLOW; a software system for 3D flow simulations,” The Netherlands: Delft Hydraulics, Report X, vol. 356, p. M3470, 2007

2007
[11]

Integrated high-resolution numerical model for the nw iberian peninsula coast and main estuarine systems,

M. C. Sousa, A. S. Ribeiro, M. Des, R. Mendes, I. Alvarez, M. Gomez- Gesteira, and J. M. Dias, “Integrated high-resolution numerical model for the nw iberian peninsula coast and main estuarine systems,”Journal of Coastal Research, no. 85, pp. 66–70, 2018

2018
[12]

Two-dimensional mapping and tracking of a coastal upwelling front by an autonomous underwater vehicle,

Y . Zhang, J. G. Bellingham, J. P. Ryan, B. Kieft, and M. J. Stanway, “Two-dimensional mapping and tracking of a coastal upwelling front by an autonomous underwater vehicle,” in2013 OCEANS - San Diego, 2013

2013
[13]

AUV Adaptive Sampling Methods: A Review,

J. Hwang, N. Bose, and S. Fan, “AUV Adaptive Sampling Methods: A Review,” Applied Sciences, vol. 9, no. 15, 2019

2019
[14]

Hierarchical probabilistic regression for auv-based adaptive sampling of marine phenomena,

J. Das, J. Harvey, F. Py, H. Vathsangam, R. Graham, K. Rajan, and G. S. Sukhatme, “Hierarchical probabilistic regression for auv-based adaptive sampling of marine phenomena,” in 2013 IEEE International Conference on Robotics and Automation, 2013, pp. 5571–5578

2013
[15]

Sampling-based robotic information gathering algorithms,

G. Hollinger and G. Sukhatme, “Sampling-based robotic information gathering algorithms,” The International Journal of Robotics Research, vol. 33, pp. 1271–1287, 08 2014

2014
[16]

Informative path planning in random fields via mixed integer programming,

S. Dutta, N. Wilde, and S. L. Smith, “Informative path planning in random fields via mixed integer programming,” in 2022 IEEE 61st Conference on Decision and Control (CDC), 2022

2022
[17]

3-D adaptive AUV sam- pling for classification of water masses,

Y . Ge, J. Eidsvik, and T. Mo-Bjørkelund, “3-D adaptive AUV sam- pling for classification of water masses,” IEEE Journal of Oceanic Engineering, vol. 48, no. 3, pp. 626–639, 2023

2023
[19]

A multiagent deep reinforcement learning approach for path planning in autonomous surface vehicles: The Ypacara ´ı lake patrolling case,

S. Y . Luis, D. G. Reina, and S. L. T. Mar ´ın, “A multiagent deep reinforcement learning approach for path planning in autonomous surface vehicles: The Ypacara ´ı lake patrolling case,” IEEE access, vol. 9, pp. 17 084–17 099, 2021

2021
[20]

Adaptive path planning using Gaussian process regression: a reinforcement learning approach,

L. Zhao, “Adaptive path planning using Gaussian process regression: a reinforcement learning approach,” in Fourth International Conference on Signal Processing and Computer Science (SPCS 2023), 2023

2023
[21]

Cooperative deep reinforcement learning for dynamic pollution plume monitoring using a drone fleet,

M. S. Assenine, W. Bechkit, I. Mokhtari, H. Rivano, and K. Benatchba, “Cooperative deep reinforcement learning for dynamic pollution plume monitoring using a drone fleet,” IEEE Internet of Things Journal, vol. 11, no. 5, pp. 7325–7338, 2023

2023
[22]

Deep learning based active spatial channel gain prediction using a swarm of unmanned aerial vehicles,

E. Krijestorac and D. Cabric, “Deep learning based active spatial channel gain prediction using a swarm of unmanned aerial vehicles,” arXiv preprint arXiv:2310.04547, 2023

2023 arXiv
[23]

Wildfire front monitoring with multiple uavs using deep q-learning,

A. Viseras, M. Meißner, and J. Marchal, “Wildfire front monitoring with multiple uavs using deep q-learning,” IEEE Access, 2021

2021
[24]

Trajectory optimization for underwater vehi- cles in time-varying ocean flows,

M. Aguiar, J. B. de Sousa, J. M. Dias, J. E. da Silva, R. Mendes, and A. S. Ribeiro, “Trajectory optimization for underwater vehi- cles in time-varying ocean flows,” in 2018 IEEE/OES Autonomous Underwater Vehicle Workshop (AUV). IEEE, 2018

2018
[25]

Errors in dynamical fields inferred from oceanographic cruise data: Part II. The impact of the lack of synopticity,

D. Gomis, A. Pascual, and M. A. Pedder, “Errors in dynamical fields inferred from oceanographic cruise data: Part II. The impact of the lack of synopticity,” Journal of Marine Systems, vol. 56, no. 3-4, pp. 334–351, 2005

2005
[26]

Carlton, Marine propellers and propulsion

J. Carlton, Marine propellers and propulsion. Butterworth- Heinemann, 2018

2018
[27]

R. S. Sutton, A. G. Barto et al., Reinforcement learning: An introduction. MIT press Cambridge, 1998

1998
[28]

Deep implicit coordination graphs for multi-agent reinforcement learning,

S. Li, J. K. Gupta, P. Morales, R. Allen, and M. J. Kochenderfer, “Deep implicit coordination graphs for multi-agent reinforcement learning,” arXiv preprint arXiv:2006.11438, 2020

2006 arXiv
[29]

Grandmaster level in StarCraft II using multi-agent reinforcement learning,

O. Vinyals, I. Babuschkin, W. M. Czarnecki, M. Mathieu, A. Dudzik, J. Chung, D. H. Choi, R. Powell, T. Ewalds, P. Georgiev et al., “Grandmaster level in StarCraft II using multi-agent reinforcement learning,” nature, vol. 575, no. 7782, pp. 350–354, 2019

2019
[30]

Distributed multi-agent target search and tracking with gaussian process and reinforcement learning,

J. Kim, D. Jang, and H. J. Kim, “Distributed multi-agent target search and tracking with gaussian process and reinforcement learning,” International Journal of Control, Automation and Systems, vol. 21, no. 9, p. 3057–3067, 2023

2023
[31]

Credit assignment for collective multiagent RL with global rewards,

D. T. Nguyen, A. Kumar, and H. C. Lau, “Credit assignment for collective multiagent RL with global rewards,” Advances in neural information processing systems, 2018

2018

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.

[1] [32]

Dis- tributed coverage control for time-varying spatial processes,

F. Pratissoli, M. Mantovani, A. Prorok, and L. Sabattini, “Dis- tributed coverage control for time-varying spatial processes,” IEEE Transactions on Robotics, 2025

2025

[2] [18]

Efficient 3d real-time adaptive auv sampling of a river plume front,

M. O. Berild, Y . Ge, J. Eidsvik, G.-A. Fuglstad, and I. Ellingsen, “Efficient 3d real-time adaptive auv sampling of a river plume front,” Frontiers in Marine Science, vol. V olume 10 - 2023, 2024

2023

[3] [1]

Ai-driven marine robotics: Emerging trends in underwater perception and ecosystem monitoring,

S. Raine and T. Fischer, “Ai-driven marine robotics: Emerging trends in underwater perception and ecosystem monitoring,” arXiv preprint arXiv:2509.01878, 2025

arXiv 2025

[4] [2]

The use of emerging autonomous technologies for ocean monitoring: insights and legal challenges,

A. Pastra, T. Johansson, J. Soares, and F. E. Muller-Karger, “The use of emerging autonomous technologies for ocean monitoring: insights and legal challenges,” Frontiers in Marine Science, vol. 12, p. 1561737, 2025

2025

[5] [3]

Observation of a turbid plume using modis imagery: The case of douro estuary (portugal),

R. Mendes, N. Vaz, D. Fern ´andez-N´ovoa, J. Da Silva, M. Decastro, M. G´omez-Gesteira, and J. Dias, “Observation of a turbid plume using modis imagery: The case of douro estuary (portugal),” Remote sensing of environment, vol. 154, pp. 127–138, 2014

2014

[6] [4]

Mixing and transport in coastal river plumes,

A. R. Horner-Devine, R. D. Hetland, and D. G. MacDonald, “Mixing and transport in coastal river plumes,” Annual Review of Fluid Mechanics, vol. 47, no. 1, pp. 569–594, 2015

2015

[7] [5]

3D Tracking of a River Plume Front with an AUV,

D. Teixeira, J. B. d. Sousa, R. Mendes, and J. Fonseca, “3D Tracking of a River Plume Front with an AUV,” in OCEANS 2021: San Diego – Porto, 2021

2021

[8] [6]

Formation control of multiple autonomous underwater vehicles: a review,

T. Yan, Z. Xu, S. X. Yang, and S. A. Gadsden, “Formation control of multiple autonomous underwater vehicles: a review,” Intelligence & Robotics, vol. 3, no. 1, 2023

2023