Control of Marine Robots in the Era of Data-Driven Intelligence

Fumin Zhang; Lin Hong; Lu Liu; Zhouhua Peng

REVIEW 3 major objections 5 minor 149 references

Control of Marine Robots in the Era of Data-Driven Intelligence

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A new review maps how data-driven control is reshaping marine robotics, from single robots to cooperative fleets.

desk verdict A useful but not systematic review; the open-source resource list and taxonomy are the real value, while the 'comprehensive' claim needs backing. read the letter →

arxiv 2506.21063 v1 pith:ZPL2E5LZ submitted 2025-06-26 cs.RO

Lin Hong , Lu Liu , Zhouhua Peng , Fumin Zhang This is my paper · ORCID

classification cs.RO

keywords marinerobotsdata-drivencontrolmodel-basedcooperativereinforcementlearningopensourcesimulationplatformsdeep

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 marine robot control is undergoing a paradigm shift from purely model-based methods toward data-driven intelligence, and it offers a structured map of the field to guide future research. It claims that a comprehensive review of recent advances, covering both individual robots and cooperative multi-robot systems, reveals a clear trajectory: data-driven methods are not replacing classical control but complementing it, with hybrid controllers as the most promising design. The review also catalogs open-source simulators and robot platforms that support development and validation, and it identifies data efficiency, interpretability, and onboard deployment as the key bottlenecks to real-world autonomy.

What carries the argument

The organizing device is the taxonomy of data-driven control methods, anchored by the standard rigid-body marine robot model $M_i \dot{v}_i + C_i(v_i)v_i + D_i(v_i)v_i + g_i(\eta_i) = \tau_i + \tau_{\text{env}}$, which applies to both single and multi-robot systems. This equation lets the paper classify methods by how they handle unknown dynamics: learning the residual term $f_{\text{residual}}(x,u)$ with Gaussian processes, lifting states into a linear Koopman space, embedding physical laws via PINNs, or skipping the model entirely with reinforcement learning. The taxonomy carries the argument by showing that each approach trades off interpretability, sample efficiency, and real-time deployability, and that hybrid designs are the emerging resolution of those trade-offs.

What would settle it

A systematic database search following the paper's own categories—model-based data-driven, model-free data-driven, and hybrid methods for both single and cooperative marine robots—over the same time window would either confirm the taxonomy's completeness or reveal substantial omitted work and misclassified algorithms.

Watch

Extended reading notes

Core claim

The central claim is that the next-generation control framework for marine robots will be built on integrating data-driven intelligence with classical and modern control theory, not on wholesale replacement. The paper organizes the literature into a three-part taxonomy: model-based data-driven methods (neural networks, Gaussian processes, Koopman operators, physics-informed neural networks), model-free data-driven methods (deep reinforcement learning, imitation learning, deep imitation reinforcement learning), and hybrid integrations such as MPC with RL and SMC with RL. For cooperative systems, it extends the framework to multi-agent reinforcement learning and deep MARL across formation, flocking, target enclosing, coverage, game-based competition, and cross-domain coordination. The review grounds these categories in a unified six-degree-of-freedom dynamic model and positions open-source simulators and robot platforms as enabling infrastructure for the field.

Load-bearing premise

The roadmap's usefulness depends on the roughly 150 cited works and the curated open-source list being representative of the field, yet the paper does not document how those works were searched, screened, or quality-filtered.

Editorial extensions

If this is right

Engineers designing marine robot controllers should expect hybrid MPC+RL and SMC+RL architectures to be the practical sweet spot in the near term, combining safety guarantees with learned adaptability.
The taxonomy provides a decision protocol: choose model-based data-driven methods when a nominal model exists, model-free methods when it does not, and hybrid approaches when both formal guarantees and adaptability are required.
Open-source platforms such as BlueROV2 and HoloOcean are turning into standard testbeds, enabling cross-laboratory comparison and faster validation of control strategies.
Progress toward high-level autonomy will be paced by three bottlenecks: data efficiency, model interpretability, and onboard computational deployment.
Data-driven intelligence will complement rather than replace classical and modern control methods in marine robotics.

Reading between the lines

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

The taxonomy implies a maturity ladder: Gaussian process and Koopman methods currently need substantial offline design, while deep reinforcement learning is the most flexible but least trusted; a practical engineering path may start hybrid and gradually increase the learned component as trust grows.
The paper's observation that multi-robot open-source simulators are scarce suggests that cooperative marine robotics research may be slowed more by a lack of shared, communication-aware simulation infrastructure than by algorithm deficits.
The future direction on LLM-empowered control could be tested concretely: giving a residual reinforcement learning policy a natural-language mission description should improve sample efficiency more in semantically rich tasks like 'survey the reef' than in purely geometric tasks like 'go to coordinates x, y'.
Because the same unified dynamic model underlies USVs, AUVs, and gliders, transfer of data-driven control methods across robot types may be under-exploited relative to its potential.

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.

Desk Editor's Note

read the letter

This is a review paper, so judge it as one. The main value is the taxonomy and the curated open-source resource list. The split into model-based data-driven (NN, GP, Koopman, PINN) and model-free (DRL, IL, DIRL) for single robots, plus cooperative control and open-source simulators/robots, gives a newcomer a decent map. The standard equations are correct, the writing is clear, and the future directions are sensible. It does what a roadmap should do: point you to the right papers and tools.

The soft spots are about the claim of being 'comprehensive.' There is no search protocol, inclusion criteria, or quality filter. That means the representativeness of the 150 cited works is unverifiable. The key tables (1 and 2) are in supplemental material, which is fine in a journal but absent here, so the actual resource list is not available to the reader. There's also an uncited claim that MarineGym is 'the first simulator specifically designed for RL'—that needs a citation or softening. The self-citation cluster is noticeable but not disqualifying; these authors are central to cooperative USV control, so citing their own work is expected. Still, without a protocol, you can't rule out network bias.

None of this undermines the paper's internal correctness. It's a solid entry point, not a systematic evidence map. The right fix is to either soften 'comprehensive' or add a methods paragraph describing how papers were selected and make the tables accessible. A referee could reasonably ask for that.

Yes, it deserves a serious referee. It's a review by established researchers, likely intended for an invited venue, and it will be a useful reference if the missing methodology is added. I'd send it out, with the caveat that the referee should check the MarineGym claim and the completeness of the resource lists.

Referee Report

3 major / 5 minor

Summary. The paper is a survey of data-driven control for marine robots, covering both single-robot and cooperative multi-robot systems. It organizes recent methods into model-based data-driven, model-free data-driven, and hybrid categories, reviews cooperative scenarios and MARL/DMARL approaches, and lists open-source simulators and robot platforms. It also presents the standard 6-DOF rigid-body marine robot model and outlines future research directions, with the stated aim of serving as a roadmap for next-generation marine robot control in the era of data-driven intelligence.

Significance. If the survey's coverage is genuinely representative, the paper provides a useful synthesis of a rapidly growing field and a practical resource list of simulators and open-source platforms. The taxonomy of model-based data-driven methods (NN, GP, Koopman, PINN) and model-free methods (DRL, IL, DIRL) is reasonable, and the mathematical background equations (1)-(8) are presented correctly. The open-source resource discussion, despite its shortcomings, addresses a real community need. However, the central claim of being a 'comprehensive review' and 'roadmap' depends on the representativeness of the 150 cited works and the completeness of the resource tables, and that foundation is not currently demonstrated.

major comments (3)

[Section 1 and Section 5] The paper's central claim, stated in the Abstract and Section 1, is that it provides a 'comprehensive review' and 'roadmap' for data-driven marine robot control. Yet no search strategy, inclusion criteria, time window for literature coverage, or quality filter is described in Sections 1 or 5. Without such a protocol, the reader cannot verify that the 150 cited works and the curated resource lists are representative rather than selective. This is a load-bearing omission for a survey whose usefulness is precisely its coverage; please add a methodology paragraph describing how papers were searched, screened, and selected, and state any inclusion/exclusion criteria.
[Section 5.1.1] The claim that 'MarineGym addresses this shift as the first simulator specifically designed for RL' is made without a citation, and the tables that would support the resource inventories (Tables 1 and 2) are relegated to Supplemental Material, which was not available for review. For a survey whose value depends on the accuracy and completeness of its open-source resource list, these items need to be verifiable: please cite the MarineGym source and move Tables 1 and 2 into the main text or otherwise make them available to reviewers and readers.
[Reference list] The reference list has a visible concentration of works from the authors' own research networks, including refs. 23, 94, 95, 97, 100, 101, 106, 121-123, 126, 127, 131, 132, and 147. While self-citation is not improper per se, the absence of a documented selection protocol makes it impossible to rule out over-representation of one methodological school, which would skew the roadmap. Please either add a coverage analysis or explicitly discuss the scope limitations and the potential for selection bias in the survey.

minor comments (5)

[Abstract] The phrase 'The rapid evolution of machine learning have opened new avenues' contains a subject-verb agreement error; it should be 'has opened'.
[Figure 1 margin note] The glossary note 'Mrine robots' is misspelled; it should be 'Marine robots'.
[Section 4.1.3] The phrase 'such as USVs, AUVs, and and unmanned aerial vehicles (UAVs)' contains a duplicated 'and'; please correct.
[References] Reference 107 appears to be a duplicate of Reference 65 (Du et al. 2022, 'Safe deep reinforcement learning-based adaptive control for usv interception mission'). Please consolidate or remove the duplicate.
[Supplemental Material references] The text refers to 'Supplemental Material 1' through 'Supplemental Material 5' and to 'Table 1' and 'Table 2' in Section 5, but these materials are not included with the manuscript; please clarify their availability or integrate key content into the main text.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper is a narrative review that reports and organizes cited results; no derived quantity or fitted parameter is shown to be equivalent to its own inputs.

full rationale

This is a narrative review, not a derivation. It introduces a standard 6-DOF marine robot model (Equations 1–5) only as background taxonomy, and then surveys data-driven methods (Sections 3 and 4) and open-source resources (Section 5) by citing external works. No prediction, theorem, or fitted quantity is derived from the review's own claims, and no parameter is fitted and then renamed as a finding. The paper's roadmap value depends on whether the 150 works and the open-source list are representative, and there is no documented search strategy; several factual claims (e.g., 'MarineGym addresses this shift as the first simulator specifically designed for RL', Section 5.1.1) are uncited, and Tables 1 and 2 are relegated to Supplemental Material. These are completeness and verifiability limitations, which would be correctness or reporting risks, but they do not make the review circular: the text's claims are summaries of external results, not equivalences to its own inputs. The visible self-citation cluster (e.g., refs. 23, 94–95, 97, 100–101, 106, 121–123, 126–127, 131–132, 147) is normal scholarly citation in a specialized field and is not used to justify a uniqueness theorem or to forbid alternatives, so it does not raise the circularity score.

Assumptions & free parameters 0 free parameters · 1 assumptions · 0 invented entities

The review introduces no free parameters and no invented entities. It relies on the standard 6-DOF rigid-body model as a unifying frame for the surveyed methods.

assumptions (1)

domain assumption The 6-DOF rigid-body model in Eq. 5 (inertia M, Coriolis C, damping D, restoring g, control tau, environment tau_env) represents the dynamics of all marine robot classes surveyed.
Section 2 frames the entire review on this model; if a platform (for example, soft-bodied robots or aerial-aquatic hybrids) violates it, some reviewed control claims may not transfer.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Control of Marine Robots in the Era of Data-Driven Intelligence." pith.science (2026). https://pith.science/paper/ZPL2E5LZ

@misc{pith2026250621063,
  author       = {Pith},
  title        = {Pith review of: Control of Marine Robots in the Era of Data-Driven Intelligence},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZPL2E5LZ}},
  note         = {Machine review of arXiv:2506.21063}
}

read the original abstract

The control of marine robots has long relied on model-based methods grounded in classical and modern control theory. However, the nonlinearity and uncertainties inherent in robot dynamics, coupled with the complexity of marine environments, have revealed the limitations of conventional control methods. The rapid evolution of machine learning has opened new avenues for incorporating data-driven intelligence into control strategies, prompting a paradigm shift in the control of marine robots. This paper provides a review of recent progress in marine robot control through the lens of this emerging paradigm. The review covers both individual and cooperative marine robotic systems, highlighting notable achievements in data-driven control of marine robots and summarizing open-source resources that support the development and validation of advanced control methods. Finally, several future perspectives are outlined to guide research toward achieving high-level autonomy for marine robots in real-world applications. This paper aims to serve as a roadmap toward the next-generation control framework of marine robots in the era of data-driven intelligence.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

149 extracted references · 77 canonical work pages

[1]

Zhang F, Marani G, Smith RN, Choi HT. 2015. Future trends in marine robotics [tc spotlight]. IEEE Robotics & Automation Magazine22(1):14–122

2015
[2]

Campos DF, Gon¸ calves EP, Campos HJ, Pereira MI, Pinto AM. 2024. Nautilus: An au- tonomous surface vehicle with a multilayer software architecture for offshore inspection.Jour- nal of Field Robotics41(4):966–990

2024
[3]

Katzschmann RK, DelPreto J, MacCurdy R, Rus D. 2018. Exploration of underwater life with an acoustically controlled soft robotic fish.Science Robotics3(16):eaar3449

2018
[4]

Wang T, Joo HJ, Song S, Hu W, Keplinger C, Sitti M. 2023. A versatile jellyfish-like robotic platform for effective underwater propulsion and manipulation.Science Advances 9(15):eadg0292

2023
[5]

Liu Y, Li C, Li J, Lin Z, Meng W, Zhang F. 2025. Wukong: Design, modeling and control of a compact flexible hybrid aerial-aquatic vehicle.IEEE Robotics and Automation Letters 10(2):1417–1424

2025
[6]

Leonard NE, Paley DA, Davis RE, Fratantoni DM, Lekien F, Zhang F. 2010. Coordinated con- trol of an underwater glider fleet in an adaptive ocean sampling field experiment in monterey bay.Journal of Field Robotics27(6):718–740

2010
[7]

Paull L, Seto M, Leonard JJ, Li H. 2018. Probabilistic cooperative mobile robot area coverage and its application to autonomous seabed mapping.The International Journal of Robotics Research37(1):21–45

2018
[8]

Wang Y, Garcia E, Casbeer D, Zhang F. 2017. Cooperative control of multi-agent systems: Theory and applications

2017

Show all 149 references

[9]

Chang D, Wu W, Edwards CR, Zhang F. 2017. Motion tomography: Mapping flow fields using autonomous underwater vehicles.The International Journal of Robotics Research36(3):320– 336

2017
[10]

Heshmati-Alamdari S, Bechlioulis CP, Karras GC, Kyriakopoulos KJ. 2020. Cooperative impedance control for multiple underwater vehicle manipulator systems under lean communi- cation.IEEE Journal of Oceanic Engineering46(2):447–465

2020
[11]

Degorre L, Delaleau E, Chocron O. 2023. A survey on model-based control and guidance prin- ciples for autonomous marine vehicles.Journal of Marine Science and Engineering11(2):430

2023
[12]

He Y, Wang DB, Ali ZA. 2020. A review of different designs and control models of remotely operated underwater vehicle.Measurement and Control53(9-10):1561–1570

2020
[13]

Wang L, Wu Q, Liu J, Li S, Negenborn RR. 2019. State-of-the-art research on motion control of maritime autonomous surface ships.Journal of Marine Science and Engineering7(12):438

2019
[14]

Er MJ, Ma C, Liu T, Gong H. 2023. Intelligent motion control of unmanned surface vehicles: A critical review.Ocean Engineering280:114562

2023
[15]

Wang J, Wu Z, Dong H, Tan M, Yu J. 2022. Development and control of underwater gliding robots: A review.IEEE/CAA Journal of Automatica Sinica9(9):1543–1560

2022
[16]

Wang R, Wang S, Wang Y, Cheng L, Tan M. 2020. Development and motion control of 22 Hong et al. biomimetic underwater robots: A survey.IEEE Transactions on Systems, Man, and Cyber- netics: Systems52(2):833–844

2020
[17]

Sun B, Li W, Wang Z, Zhu Y, He Q, et al. 2022. Recent progress in modeling and control of bio-inspired fish robots.Journal of Marine Science and Engineering10(6):773

2022
[18]

Xiang X, Yu C, Lapierre L, Zhang J, Zhang Q. 2018. Survey on fuzzy-logic-based guidance and control of marine surface vehicles and underwater vehicles.International Journal of Fuzzy Systems20:572–586

2018
[19]

Wei H, Shi Y. 2022. Mpc-based motion planning and control enables smarter and safer au- tonomous marine vehicles: Perspectives and a tutorial survey.IEEE/CAA Journal of Auto- matica Sinica10(1):8–24

2022
[20]

Hassani V, Pascoal AM, Onstein TF. 2018. Data-driven control in marine systems.Annual Reviews in Control46:343–349

2018
[21]

Tijjani AS, Chemori A, Creuze V. 2022. A survey on tracking control of unmanned underwater vehicles: Experiments-based approach.Annual Reviews in Control54:125–147

2022
[22]

Das B, Subudhi B, Pati BB. 2016. Cooperative formation control of autonomous underwater vehicles: An overview.International Journal of Automation and computing13:199–225

2016
[23]

Peng Z, Wang J, Wang D, Han QL. 2020. An overview of recent advances in coordinated control of multiple autonomous surface vehicles.IEEE Transactions on Industrial Informatics 17(2):732–745

2020
[24]

Yang Y, Xiao Y, Li T. 2021. A survey of autonomous underwater vehicle formation: Perfor- mance, formation control, and communication capability.IEEE Communications Surveys & Tutorials23(2):815–841

2021
[25]

Jiang T, Yan Y, Wu D, Yu S, Li T. 2022. Neural network based adaptive sliding mode track- ing control of autonomous surface vehicles with input quantization and saturation.Ocean Engineering265:112505

2022
[26]

Chen L, Cui R, Yang C, Yan W. 2020. Adaptive neural network control of underactuated surface vessels with guaranteed transient performance: Theory and experimental results.IEEE Transactions on Industrial Electronics67(5):4024–4035

2020
[27]

Zhang JX, Yang T, Chai T. 2022. Neural network control of underactuated surface vehicles with prescribed trajectory tracking performance.IEEE Transactions on Neural Networks and Learning Systems

2022
[28]

Li F, Li H, Wu C. 2024. Gaussian process-based learning model predictive control with appli- cation to usv.IEEE Transactions on Industrial Electronics

2024
[29]

Lima GS, Trimpe S, Bessa WM. 2020. Sliding mode control with gaussian process regression for underwater robots.Journal of Intelligent & Robotic Systems99(3):487–498

2020
[30]

Dang Y, Huang Y, Shen X, Zhu D, Chu Z. 2025. Incremental sparse gaussian process-based model predictive control for trajectory tracking of unmanned underwater vehicles.IEEE Robotics and Automation Letters

2025
[31]

Amer A, Mehndiratta M, Brodskiy Y, Kayacan E. 2025. Empowering autonomous underwa- ter vehicles using learning-based model predictive control with dynamic forgetting gaussian processes.IEEE Transactions on Control Systems Technology

2025
[32]

Cui Y, Peng L, Li H. 2022. Filtered probabilistic model predictive control-based reinforce- ment learning for unmanned surface vehicles.IEEE Transactions on Industrial Informatics 18(10):6950–6961

2022
[33]

Rahmani M, Redkar S. 2024. Enhanced koopman operator-based robust data-driven control for 3 degree of freedom autonomous underwater vehicles: A novel approach.Ocean Engineering 307:118227

2024
[34]

Li J, Park H, Hao W, Xin L, Chavez-Galaviz J, et al. 2024. C3d: Cascade control with change point detection and deep koopman learning for autonomous surface vehicles.arXiv preprint arXiv:2403.05972

2024 arXiv
[35]

Mamakoukas G, Castano ML, Tan X, Murphey TD. 2021. Derivative-based koopman operators www.annualreviews.org • Control of Marine Robots in the Era of Data-Driven Intelligence 23 for real-time control of robotic systems.IEEE Transactions on Robotics37(6):2173–2192

2021
[36]

2025.Koopman Operator and Model Predictive Control Integration for Au- tonomous Underwater Vehicle Speed Regulation

Cong B, Liu Z. 2025.Koopman Operator and Model Predictive Control Integration for Au- tonomous Underwater Vehicle Speed Regulation. In2025 Australian & New Zealand Control Conference (ANZCC), pp. 81–86. IEEE

2025
[37]

Raissi M, Perdikaris P, Karniadakis GE. 2019. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial dif- ferential equations.Journal of Computational physics378:686–707

2019
[38]

Liu T, Zhao J, Huang J, Li Z, Xu L, Zhao B. 2024. Research on model predictive control of autonomous underwater vehicle based on physics informed neural network modeling.Ocean Engineering304:117844

2024
[39]

2024.Safe navigation of autonomous underwater vehicles using physics-informed neural networks

Majumder R, Makam R, Mane P, KS B, Sundaram S. 2024.Safe navigation of autonomous underwater vehicles using physics-informed neural networks. InOCEANS 2024-Singapore, pp. 1–6. IEEE

2024
[40]

Algar ´ ın-Pinto JA, Garza-Casta˜ n´ on LE, Vargas-Mart ´ ınez A, Minchala-´Avila LI, Payeur P. 2025. Intelligent motion control to enhance the swimming performance of a biomimetic underwater vehicle using reinforcement learning approach.IEEE Access

2025
[41]

Higo Y, Sakano M, Nobe H, Hashimoto H. 2023. Development of trajectory-tracking ma- neuvering system for automatic berthing/unberthing based on double deep q-network and experimental validation with an actual large ferry.Ocean Engineering287:115750

2023
[42]

Wu X, Chen H, Chen C, Zhong M, Xie S, et al. 2020. The autonomous navigation and obstacle avoidance for usvs with anoa deep reinforcement learning method.Knowledge-Based Systems 196:105201

2020
[43]

Chen G, Zhao Z, Lu Y, Yang C, Hu H. 2024. Deep reinforcement learning-based pitch attitude control of a beaver-like underwater robot.Ocean Engineering307:118163

2024
[44]

Wang X, Wang S, Liang X, Zhao D, Huang J, et al. 2022. Deep reinforcement learning: A survey.IEEE Transactions on Neural Networks and Learning Systems35(4):5064–5078

2022
[45]

Øvereng SS, Nguyen DT, Hamre G. 2021. Dynamic positioning using deep reinforcement learning.Ocean Engineering235:109433

2021
[46]

Hasankhani A, Tang Y, VanZwieten J. 2023. Integrated path planning and control through proximal policy optimization for a marine current turbine.Applied Ocean Research137:103591

2023
[47]

Wu C, Yu W, Li G, Liao W. 2023. Deep reinforcement learning with dynamic window ap- proach based collision avoidance path planning for maritime autonomous surface ships.Ocean Engineering284:115208

2023
[48]

Huang H, Jiang T, Zhang Z, Sun Y, Qin H, et al. 2024. Learning strategies for underwater robot autonomous manipulation control.Journal of the Franklin Institute361(7):106773

2024
[49]

2024.Target Tracking Control of Underactuated Unmanned Boats Using an Improved PPO Algorithm

Hao Y, Wang Q, Shen K. 2024.Target Tracking Control of Underactuated Unmanned Boats Using an Improved PPO Algorithm. InInternational Conference on Guidance, Navigation and Control, pp. 330–340. Springer

2024
[50]

Zhang T, Miao X, Li Y, Jia L, Wei Z, et al. 2023. Auv 3d docking control using deep rein- forcement learning.Ocean engineering283:115021

2023
[51]

Chu S, Lin M, Li D, Lin R, Xiao S. 2025. Adaptive reward shaping based reinforcement learning for docking control of autonomous underwater vehicles.Ocean Engineering318:120139

2025
[52]

Wang N, Gao Y, Zhang X. 2021. Data-driven performance-prescribed reinforcement learning control of an unmanned surface vehicle.IEEE Transactions on Neural Networks and Learning Systems32(12):5456–5467

2021
[53]

Deng Y, Liu T, Zhao D. 2021. Event-triggered output-feedback adaptive tracking control of autonomous underwater vehicles using reinforcement learning.Applied Ocean Research 113:102676

2021
[54]

Wang Y, Gao J. 2024. Reinforcement-learning-based visual servoing of underwater vehicle dual-manipulator system.Journal of Marine Science and Engineering12(6):940

2024
[55]

Wang Y, Chu H, Ma R, Bai X, Cheng L, et al. 2024. Learning-based discontinuous path 24 Hong et al. following control for a biomimetic underwater vehicle.Research7:0299

2024
[56]

Ma R, Wang Y, Tang C, Wang S, Wang R. 2023. Position and attitude tracking control of a biomimetic underwater vehicle via deep reinforcement learning.IEEE/ASME Transactions on Mechatronics28(5):2810–2819

2023
[57]

Wang Y, Hou Y, Lai Z, Cao L, Hong W, Wu D. 2025. An expert-demonstrated soft actor– critic based adaptive trajectory tracking control of autonomous underwater vehicle with long short-term memory.Ocean Engineering321:120405

2025
[58]

Yuan W, Rui X. 2023. Deep reinforcement learning-based controller for dynamic positioning of an unmanned surface vehicle.Computers and Electrical Engineering110:108858

2023
[59]

Dong N, Liu S, Ip A W, Yung KL, Gao Z, et al. 2025. End-to-end autonomous underwater vehicle path following control method based on improved soft actor-critic for deep space ex- ploration.Journal of Industrial Information Integration:100792

2025
[60]

Ma R, Wang Y, Wang S, Cheng L, Wang R, Tan M. 2023. Sample-observed soft actor-critic learning for path following of a biomimetic underwater vehicle.IEEE Transactions on Au- tomation Science and Engineering

2023
[61]

Sun Y, Ran X, Zhang G, Wang X, Xu H. 2020. Auv path following controlled by modified deep deterministic policy gradient.Ocean Engineering210:107360

2020
[62]

Wang Y, Tang C, Wang S, Cheng L, Wang R, et al. 2022. Target tracking control of a biomimetic underwater vehicle through deep reinforcement learning.IEEE Transactions on Neural Networks and Learning Systems33(8):3741–3752

2022
[63]

Zhang C, Cheng P, Du B, Dong B, Zhang W. 2022. Auv path tracking with real-time ob- stacle avoidance via reinforcement learning under adaptive constraints.Ocean Engineering 256:111453

2022
[64]

Masmitja I, Martin M, O’Reilly T, Kieft B, Palomeras N, et al. 2023. Dynamic robotic tracking of underwater targets using reinforcement learning.Science robotics8(80):eade7811

2023
[66]

Zhu P, Liu S, Jiang T, Liu Y, Zhuang X, Zhang Z. 2022. Autonomous reinforcement control of visual underwater vehicles: Real-time experiments using computer vision.IEEE Transactions on Vehicular Technology71(8):8237–8250

2022
[67]

2021.Deep reinforcement learning for mapless navigation of a hybrid aerial underwater vehicle with medium transition

Grando RB, de Jesus JC, Kich V A, Kolling AH, Bortoluzzi NP, et al. 2021.Deep reinforcement learning for mapless navigation of a hybrid aerial underwater vehicle with medium transition. In2021 IEEE International Conference on Robotics and Automation (ICRA), pp. 1088–1094. IEEE

2021
[68]

Hadi B, Khosravi A, Sarhadi P. 2022. Deep reinforcement learning for adaptive path planning and control of an autonomous underwater vehicle.Applied Ocean Research129:103326

2022
[69]

Fan Y, Dong H, Zhao X, Denissenko P. 2024. Path-following control of unmanned underwater vehicle based on an improved td3 deep reinforcement learning.IEEE Transactions on Control Systems Technology

2024
[70]

Zhou Y, Gong C, Chen K. 2025. Adaptive control scheme for usv trajectory-tracking under complex environmental disturbances via deep reinforcement learning.IEEE Internet of Things Journal

2025
[71]

Sivaraj S, Dubey A, Rajendran S. 2023. On the performance of different deep reinforcement learning based controllers for the path-following of a ship.Ocean Engineering286:115607

2023
[72]

2020.Vision-Based Goal-Conditioned Policies for Underwater Navigation in the Presence of Obstacles

Manderson T, Higuera JCG, Wapnick S, Tremblay JF, Shkurti F, et al. 2020.Vision-Based Goal-Conditioned Policies for Underwater Navigation in the Presence of Obstacles. InPro- ceedings of Robotics: Science and Systems. Corvalis, Oregon, USA

2020
[73]

2024.Uivnav: Underwater information- driven vision-based navigation via imitation learning

Lin X, Karapetyan N, Joshi K, Liu T, Chopra N, et al. 2024.Uivnav: Underwater information- driven vision-based navigation via imitation learning. In2024 IEEE International Conference on Robotics and Automation (ICRA), pp. 5250–5256. IEEE

2024
[74]

Liu R, Ha H, Hou M, Song S, Vondrick C. 2024. Self-improving autonomous underwater www.annualreviews.org • Control of Marine Robots in the Era of Data-Driven Intelligence 25 manipulation.arXiv preprint arXiv:2410.18969

2024 arXiv
[75]

Zhang T, Yue L, Wang C, Sun J, Zhang S, et al. 2022. Leveraging imitation learning on pose regulation problem of a robotic fish.IEEE Transactions on Neural Networks and Learning Systems35(3):4232–4245

2022
[76]

Goodfellow I, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, et al. 2020. Generative adversarial networks.Communications of the ACM63(11):139–144

2020
[77]

Chaysri P, Spatharis C, Blekas K, Vlachos K. 2023. Unmanned surface vehicle navigation through generative adversarial imitation learning.Ocean Engineering282:114989

2023
[78]

Jiang D, Huang J, Fang Z, Cheng C, Sha Q, et al. 2022. Generative adversarial interactive imitation learning for path following of autonomous underwater vehicle.Ocean Engineering 260:111971

2022
[79]

Chu Z, Sun B, Zhu D, Zhang M, Luo C. 2020. Motion control of unmanned underwater vehi- cles via deep imitation reinforcement learning algorithm.IET Intelligent Transport Systems 14(7):764–774

2020
[80]

Yan S, Wu Z, Wang J, Huang Y, Tan M, Yu J. 2022. Real-world learning control for au- tonomous exploration of a biomimetic robotic shark.IEEE Transactions on Industrial Elec- tronics70(4):3966–3974

2022
[81]

Martinsen AB, Lekkas AM, Gros S. 2022. Reinforcement learning-based nmpc for tracking control of asvs: Theory and experiments.Control Engineering Practice120:105024

2022
[82]

2024.Deep Reinforcement Learning with Model Predictive Control for Path Following of Autonomous Underwater Vehicle

Zhang Z, Pan X, Chen T, Jiang D, Fang Z, Li G. 2024.Deep Reinforcement Learning with Model Predictive Control for Path Following of Autonomous Underwater Vehicle. In2024 43rd Chinese Control Conference (CCC), pp. 2516–2523. IEEE

2024
[83]

Ma D, Chen X, Ma W, Zheng H, Qu F. 2023. Neural network model-based reinforcement learn- ing control for auv 3-d path following.IEEE Transactions on Intelligent Vehicles9(1):893–904

2023
[84]

Cui Y, Osaki S, Matsubara T. 2021. Autonomous boat driving system using sample-efficient model predictive control-based reinforcement learning approach.Journal of Field Robotics 38(3):331–354

2021
[85]

Pan J, Zhang P, Wang J, Liu M, Yu J. 2022. Learning for depth control of a robotic penguin: A data-driven model predictive control approach.IEEE Transactions on Industrial Electronics 70(11):11422–11432

2022
[86]

2019.Adaptive DDPG Design-Based Sliding- Mode Control for Autonomous Underwater Vehicles at Different Speeds

Wang D, Shen Y, Sha Q, Li G, Kong X, et al. 2019.Adaptive DDPG Design-Based Sliding- Mode Control for Autonomous Underwater Vehicles at Different Speeds. In2019 IEEE Un- derwater Technology (UT), pp. 1–5

2019
[87]

Zhu S, Zhang G, Wang Q, Li Z. 2025. Sliding mode control for variable-speed trajectory tracking of underactuated vessels with td3 algorithm optimization.Journal of Marine Science and Engineering13(1):99

2025
[88]

Wang D, Shen Y, Wan J, Sha Q, Li G, et al. 2022. Sliding mode heading control for auv based on continuous hybrid model-free and model-based reinforcement learning.Applied Ocean Research118:102960

2022
[89]

Peng Z, Wang D, Li T, Han M. 2020. Output-feedback cooperative formation maneuvering of autonomous surface vehicles with connectivity preservation and collision avoidance.IEEE Transactions on Cybernetics50(6):2527–2535

2020
[90]

Jiang Y, Liu Z, Chen F. 2024. Adaptive output-constrained finite-time formation control for multiple unmanned surface vessels with directed communication topology.Ocean Engineering 292:116552

2024
[91]

Guo Q, Zhang X, Ma D. 2024. Adaptive fixed-time path following cooperative control for autonomous surface vehicles based on path-dependent constraints.Applied Ocean Research 142:103826

2024
[92]

Ma L, Wang YL, Han QL. 2023. Cooperative target tracking of multiple autonomous surface vehicles under switching interaction topologies.IEEE/CAA Journal of Automatica Sinica 10(3):673–684 26 Hong et al

2023
[93]

Sahu BK, Subudhi B. 2018. Flocking control of multiple auvs based on fuzzy potential func- tions.IEEE Transactions on Fuzzy Systems26(5):2539–2551

2018
[94]

Peng Z, Jiang Y, Liu L, Shi Y. 2023. Path-guided model-free flocking control of unmanned surface vehicles based on concurrent learning extended state observers.IEEE Transactions on Systems, Man, and Cybernetics: Systems53(8):4729–4739

2023
[95]

Peng Z, Liu L, Wang J. 2021. Output-feedback flocking control of multiple autonomous sur- face vehicles based on data-driven adaptive extended state observers.IEEE Transactions on Cybernetics51(9):4611–4622

2021
[96]

Zhang C, Zeng R, Lin B, Zhang Y, Xie W, Zhang W. 2025. Multi-usv cooperative target en- circlement through learning-based distributed transferable policy and experimental validation. Ocean Engineering318:120124

2025
[97]

Jiang Y, Peng Z, Liu L, Wang D, Zhang F. 2024. Safety-critical cooperative target enclosing control of autonomous surface vehicles based on finite-time fuzzy predictors and input-to-state safe high-order control barrier functions.IEEE Transactions on Fuzzy Systems32(3):816–830

2024
[98]

Yan Z, Zheng H, Jiang Z, Xu W. 2025. Distributed control of unmanned marine vehicles for target circumnavigation in communication-denied environments.IEEE/ASME Transactions on Mechatronics30(1):345–356

2025
[99]

Luo J, Su Y. 2024. Path planning for multi-usv target coverage in complex environments. Ocean Engineering312:119090

2024
[100]

Jiao S, Liu L, Peng Z, Li T, Zhang W. 2025. Collision-free dynamic coverage of autonomous surface vehicles with anisotropic sensing.Ocean Engineering319:120163

2025
[101]

Liu L, Jiao S, Han B, Li T, Peng Z. 2025. Swarm-based dynamic coverage of multi-asv systems in the presence of measurement noises.IEEE Transactions on Vehicular Technology:1–15

2025
[102]

Li F, Yin M, Wang T, Huang T, Yang C, Gui W. 2024. Distributed pursuit-evasion game of limited perception USV swarm based on multiagent proximal policy optimization.IEEE Transactions on Systems, Man, and Cybernetics: Systems

2024
[103]

Meng Y, Liu C, Wang Q, Tan L. 2024. Cooperative advantage actor-critic reinforcement learn- ing for multi-agent pursuit-evasion games on communication graphs.IEEE Transactions on Artificial Intelligence

2024
[104]

Jiang Y, Li Z. 2024. Fully distributed target encircling control of autonomous surface vehicles based on noncooperative games.IEEE Transactions on Intelligent Vehicles9(4):4769–4779

2024
[105]

Nantogma S, Zhang S, Yu X, An X, Xu Y. 2023. Multi-usv dynamic navigation and target capture: A guided multi-agent reinforcement learning approach.Electronics12(7):1523

2023
[106]

Kang T, Gu N, Wang D, Liu L, Hu Q, Peng Z. 2024. Neurodynamics-based attack-defense guidance of autonomous surface vehicles against multiple attackers for domain protection. IEEE Transactions on Industrial Electronics71(10):12655–12663

2024
[107]

Du B, Lin B, Zhang C, Dong B, Zhang W. 2022. Safe deep reinforcement learning-based adaptive control for usv interception mission.Ocean Engineering246:110477

2022
[108]

Feng D, Yang J, Zhang N, Xiao J, Dai S, et al. 2025. Study on key technologies for air– water surface collaboration of observation unmanned aircraft vehicle.Electronics Letters 61(1):e70164

2025
[109]

Jiang B, Wen G, Zhou J, Zheng D. 2024. Cross-domain cooperative technology of intelligent unmanned swarm systems: Current status and prospects.Strategic Study of Chinese Academy of Engineering26(1):117–126

2024
[110]

Cao X, Liu W, Ren L. 2024. Underwater target capture based on heterogeneous unmanned system collaboration.IEEE Transactions on Intelligent Vehicles

2024
[111]

Liu Y, Chen C, Qu D, Zhong Y, Pu H, et al. 2023. Multi-USV system antidisturbance co- operative searching based on the reinforcement learning method.IEEE Journal of Oceanic Engineering48(4):1019–1047

2023
[112]

Wang Z, Zhang L, Zhu Z. 2023. Game-based distributed optimal formation tracking control of underactuated AUVs based on reinforcement learning.Ocean Engineering287:115879 www.annualreviews.org • Control of Marine Robots in the Era of Data-Driven Intelligence 27

2023
[113]

Cao W, Yan J, Yang X, Luo X, Guan X. 2023. Communication-aware formation control of AUVs with model uncertainty and fading channel via integral reinforcement learning. IEEE/CAA Journal of Automatica Sinica10(1):159–176

2023
[114]

Wang P, Yu C, Lv M, Cao J. 2023. Adaptive fixed-time optimal formation control for uncertain nonlinear multiagent systems using reinforcement learning.IEEE Transactions on Network Science and Engineering11(2):1729–1743

2023
[115]

Ding TF, Ge MF, Liu ZW, Wang L, Liu J. 2023. Reinforcement learning formation tracking of networked autonomous surface vehicles with bounded inputs via cloud-supported commu- nication.IEEE Transactions on Intelligent Vehicles9(1):469–480

2023
[116]

Ding TF, Zhang HY, Ge MF, Liu ZW. 2024. Predefined-time fuzzy reinforcement learning control for secure surrounding formation of nmsvs with dos attacks.IEEE Transactions on Fuzzy Systems

2024
[117]

Zhang J, Ren J, Cui Y, Fu D, Cong J. 2024. Multi-usv task planning method based on improved deep reinforcement learning.IEEE Internet of Things Journal11(10):18549–18567

2024
[118]

Gan W, Qu X, Song D, Yao P. 2023. Multi-usv cooperative chasing strategy based on obstacles assistance and deep reinforcement learning.IEEE Transactions on Automation Science and Engineering

2023
[119]

Wang Z, Chen P, Chen L, Mou J. 2025. Collaborative collision avoidance approach for USVs based on multi-agent deep reinforcement learning.IEEE Transactions on Intelligent Trans- portation Systems

2025
[120]

Xu J, Huang F, Wu D, Cui Y, Yan Z, Zhang K. 2021. Deep reinforcement learning based multi-auvs cooperative decision-making for attack–defense confrontation missions.Ocean En- gineering239:109794

2021
[121]

Peng Z, Wang J, Wang D. 2017. Distributed maneuvering of autonomous surface vehicles based on neurodynamic optimization and fuzzy approximation.IEEE Transactions on Control Systems Technology26(3):1083–1090

2017
[122]

Peng Z, Jiang Y, Liu L, Wang D. 2022. Distributed optimization for coordinated dynamic positioning of multiple surface vessels based on asymptotically stable esos.Ocean Engineering 246:110507

2022
[123]

Lyu G, Peng Z, Wang J. 2024. Safety-critical receding-horizon planning and formation control of autonomous surface vehicles via collaborative neurodynamic optimization.IEEE Transac- tions on Cybernetics54(12):7236–7247

2024
[124]

Sun Y, Du Y, Qin H. 2022. Distributed adaptive neural network constraint containment control for the benthic autonomous underwater vehicles.Neurocomputing484:89–98

2022
[125]

Fan Y, Li Z, Li J, Ma G, Bu H. 2025. Fixed-time event-triggered distributed formation control for underactuated usvs considering actuator saturation.Ocean Engineering316:119829

2025
[126]

Liu L, Zhang J, Sun R, Peng Z, Wang D, Shi Y. 2024. Cloud-based self-triggered coopera- tive path following of underactuated usvs with multimodel extended state observers.IEEE Transactions on Industrial Electronics

2024
[127]

Wang H, Tan H, Peng Z. 2023. Quantized communications in containment maneuvering for output constrained marine surface vehicles: Theory and experiment.IEEE Transactions on Industrial Electronics71(1):880–889

2023
[128]

Suryendu C, Subudhi B. 2020. Formation control of multiple autonomous underwater vehicles under communication delays.IEEE Transactions on Circuits and Systems II: Express Briefs 67(12):3182–3186

2020
[129]

Yan J, Gao J, Yang X, Luo X, Guan X. 2019. Position tracking control of remotely oper- ated underwater vehicles with communication delay.IEEE Transactions on Control Systems Technology28(6):2506–2514

2019
[130]

Jiang Y, Wang L, Sun J, Yu H. 2024. Attack-resistant distributed formation control for multiple unmanned surface vessels subject to output constraints.Ocean Engineering314:119712

2024
[131]

Gao S, Peng Z, Liu L, Wang D, Han QL. 2022. Fixed-time resilient edge-triggered estimation 28 Hong et al. and control of surface vehicles for cooperative target tracking under attacks.IEEE Transac- tions on Intelligent Vehicles8(1):547–556

2022
[132]

Gu N, Wang D, Peng Z, Liu L. 2020. Adaptive bounded neural network control for coordinated path-following of networked underactuated autonomous surface vehicles under time-varying state-dependent cyber-attack.ISA Transactions104:212–221

2020
[133]

Jiang X, Xia G. 2022. Nonfragile formation control of leaderless unmanned surface vehicles with memory sampling data and packet loss.IEEE Systems Journal17(2):3026–3035

2022
[134]

Zhou X, Huang B, Zhou B, Zhu C, Qin H, Miao J. 2025. Affine formation maneuver con- trol for NUSVs: An anti-competing interaction solution with random packet losses.IEEE Transactions on Automation Science and Engineering22:5916–5932

2025
[135]

2013.Glider CT: Reconstructing flow fields from predicted motion of underwater gliders

Wu W, Chang D, Zhang F. 2013.Glider CT: Reconstructing flow fields from predicted motion of underwater gliders. InProceedings of the 8th International Conference on Underwater Networks & Systems, pp. 1–8

2013
[136]

2016.Glider CT: Analysis and experimental validation

Chang D, Wu W, Zhang F. 2016.Glider CT: Analysis and experimental validation. InDis- tributed Autonomous Robotic Systems: The 12th International Symposium, pp. 285–298. Springer

2016
[137]

2016.Distributed motion tomography for time-varying flow fields

Chang D, Zhang F. 2016.Distributed motion tomography for time-varying flow fields. In OCEANS 2016-Shanghai, pp. 1–7. IEEE

2016
[138]

2019.Distributed motion tomography for reconstruction of flow fields

Chang D, Zhang F, Sun J. 2019.Distributed motion tomography for reconstruction of flow fields. In2019 international conference on robotics and automation (ICRA), pp. 8048–8054. IEEE

2019
[139]

Zuo W, Zhang F, Chen Z. 2023. Bio-inspired robotic fish enabled motion tomography.Inter- national Journal of Intelligent Robotics and Applications7(3):474–484

2023
[140]

2019.Residual reinforcement learning for robot control

Johannink T, Bahl S, Nair A, Luo J, Kumar A, et al. 2019.Residual reinforcement learning for robot control. In2019 international conference on robotics and automation (ICRA), pp. 6023–6029. IEEE

2019
[141]

2022.Deep residual reinforcement learning based autonomous blimp control

Liu YT, Price E, Black MJ, Ahmad A. 2022.Deep residual reinforcement learning based autonomous blimp control. In2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 12566–12573. IEEE

2022
[142]

Li X, Geng L, Liu K, Zhao Y, Du W. 2024. Motion control of autonomous underwater vehicle based on physics-informed offline reinforcement learning.Ocean Engineering313:119432

2024
[143]

Rodwell C, Tallapragada P. 2023. Physics-informed reinforcement learning for motion control of a fish-like swimming robot.Scientific Reports13(1):10754

2023
[144]

Achiam J, Adler S, Agarwal S, Ahmad L, Akkaya I, et al. 2023. Gpt-4 technical report.arXiv preprint arXiv:2303.08774

2023 arXiv
[145]

Liu A, Feng B, Xue B, Wang B, Wu B, et al. 2024. Deepseek-v3 technical report.arXiv preprint arXiv:2412.19437

2024 arXiv
[146]

Chen W, Li G, Li M, Wang W, Li P, et al. 2025. Llm-enabled incremental learning framework for hand exoskeleton control.IEEE Transactions on Automation Science and Engineering 22:2617–2626

2025
[147]

Yang R, Hou M, Wang J, Zhang F. 2023. Oceanchat: Piloting autonomous underwater vehicles in natural language.arXiv preprint arXiv:2309.16052

2023 arXiv
[148]

Xiao-long L, Qiang S, Zhong-hai Y, Ya-li W, Ping-ni L. 2015. Review on large-scale unmanned system swarm intelligence control method.Application Research of Computers/Jisuanji Yingyong Yanjiu32(1)

2015
[149]

Huang Z, Yuan L, Cai W. 2022. Research of key technology in marine unmanned swarm commnunication network.Application Research of Computers44(14):127–132

2022
[150]

Potokar E, Lay K, Norman K, Benham D, Ashford S, et al. 2024. Holoocean: A full-featured marine robotics simulator for perception and autonomy.IEEE Journal of Oceanic Engineering www.annualreviews.org • Control of Marine Robots in the Era of Data-Driven Intelligence 29

2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Zhang F, Marani G, Smith RN, Choi HT. 2015. Future trends in marine robotics [tc spotlight]. IEEE Robotics & Automation Magazine22(1):14–122

2015

[2] [2]

Campos DF, Gon¸ calves EP, Campos HJ, Pereira MI, Pinto AM. 2024. Nautilus: An au- tonomous surface vehicle with a multilayer software architecture for offshore inspection.Jour- nal of Field Robotics41(4):966–990

2024

[3] [3]

Katzschmann RK, DelPreto J, MacCurdy R, Rus D. 2018. Exploration of underwater life with an acoustically controlled soft robotic fish.Science Robotics3(16):eaar3449

2018

[4] [4]

Wang T, Joo HJ, Song S, Hu W, Keplinger C, Sitti M. 2023. A versatile jellyfish-like robotic platform for effective underwater propulsion and manipulation.Science Advances 9(15):eadg0292

2023

[5] [5]

Liu Y, Li C, Li J, Lin Z, Meng W, Zhang F. 2025. Wukong: Design, modeling and control of a compact flexible hybrid aerial-aquatic vehicle.IEEE Robotics and Automation Letters 10(2):1417–1424

2025

[6] [6]

Leonard NE, Paley DA, Davis RE, Fratantoni DM, Lekien F, Zhang F. 2010. Coordinated con- trol of an underwater glider fleet in an adaptive ocean sampling field experiment in monterey bay.Journal of Field Robotics27(6):718–740

2010

[7] [7]

Paull L, Seto M, Leonard JJ, Li H. 2018. Probabilistic cooperative mobile robot area coverage and its application to autonomous seabed mapping.The International Journal of Robotics Research37(1):21–45

2018

[8] [8]

Wang Y, Garcia E, Casbeer D, Zhang F. 2017. Cooperative control of multi-agent systems: Theory and applications

2017