Yüzey Dalgası Kaynaklı Bozucu Etkiler Altında Otonom Sualtı Araçlarının (AUVs) Hız Kararlılığının Uzun Kısa Süreli Bellek (LSTM) Tabanlı Model Öngörülü Kontrol (MPC) ile Sağlanması

Abstract

Otonom Sualtı Araçları (AUV’ler), sualtı keşfi, bilimsel araştırmalar ve açık deniz operasyonlarında hayati bir rol oynamaktadır. Ancak, yüzeye yakın sığ derinliklerde dalga kaynaklı bozucu etkiler nedeniyle hız kararlılığı bozulabilmektedir. Bu çalışma, karmaşık hidrodinamik modelleme veya sistem tanımlamasına ihtiyaç duymayan veri odaklı bir kontrol çerçevesi önermektedir. Dalga etkilerini içeren simülasyon verileriyle eğitilen Uzun Kısa Süreli Bellek (LSTM) tabanlı Tekrarlayan Sinir Ağı (RNN), AUV’nin hız tepkilerini tahmin etmektedir. Bu tahminler, harici bozucuları telafi eden optimal kontrol sinyalleri üretmek amacıyla Model Öngörülü Kontrol (MPC) yapısı içinde kullanılmaktadır. Yöntem, giriş–çıkış verilerinden sistemin doğrusal olmayan dinamiklerini ve çevresel etkileşimleri öğrenerek kontrol tasarımını sadeleştirirken öğrenilmiş deniz koşulları altında kararlılık sağlamaktadır. Simülasyonlar, LSTM-MPC çerçevesinin düşük izleme hataları sağladığını göstermektedir. Kontrol çerçevesi öğrenilmiş dalga koşullarında başarılı performans sergilerken, LSTM-MPC daha dinamik deniz durumlarına karşı da uyum yeteneği göstermektedir. Bu bulgular, yüzeye yakın AUV operasyonları için öğrenmeye dayalı öngörülü kontrol stratejilerinin potansiyelini vurgulamaktadır. Gelecek çalışmalarda, gerçek zamanlı uyum sağlamak için çevrimiçi öğrenmenin entegrasyonu ve önerilen yöntem performansının kontrollü bir dalga havuzunda deneysel olarak doğrulanması hedeflenmektedir.

Keywords

• Otonom Sualtı Aracı (AUV)
• Dalga Kaynaklı Bozucu Etkiler
• Uzun Kısa Süreli Bellek (LSTM)
• Model Öngörülü Kontrol (MPC)
• Dinamik Sistem Modellemesi
• Veri Tabanlı Kontrol

Velocity Stabilization of Autonomous Underwater Vehicles (AUVs) under Surface Wave Induced Disturbances Using Long Short-Term Memory (LSTM) Based Model Predictive Control (MPC)

Abstract

Autonomous Underwater Vehicles (AUVs) play a vital role in underwater exploration, scientific research, and offshore operations. However, at shallow depths near the surface, wave-induced disturbances can impair velocity stabilization. This study proposes a data-driven control framework that eliminates the need for complex hydrodynamic modeling or system identification. A Long Short-Term Memory (LSTM)-based Recurrent Neural Network (RNN), trained on simulation data incorporating wave effects, is used to predict the AUV’s velocity responses. These predictions are then utilized within a Model Predictive Control (MPC) framework to generate optimal control signals that compensate for external disturbances. By learning the nonlinear dynamics of the system and environmental interactions directly from input–output data, the proposed method simplifies controller design while ensuring stability under the wave conditions represented in the training data. Simulation results demonstrate that the LSTM-MPC framework yields low tracking errors. While the control scheme performs well under learned wave conditions, it also shows adaptability to more dynamic marine environments. These findings highlight the potential of learning-based predictive control strategies for near-surface AUV operations. Future work will focus on integrating online learning for real-time adaptation and experimentally validating the proposed method in a controlled wave basin.

Keywords

• Autonomous Underwater Vehicle (AUV)
• Wave Induced Disturbance
• Long Short-Term Memory (LSTM)
• Model Predictive Control (MPC)
• Dynamic System Modeling
• Data-Driven Control

References

1. Fossen, T.I. Handbook of Marine Craft Hydrodynamics and Motion Control; John Wiley & Sons: Chichester, İngiltere, 2011; pp. 1–227, 417–451. https://doi.org/10.1002/9781119994138
2. Giorgi, G.; Ringwood, J.V. Analytical Formulation of Nonlinear Froude-Krylov Forces for Surging-Heaving-Pitching Point Absorbers. In Proceedings of the ASME 37th International Conference on Ocean, Offshore and Arctic Engineering (OMAE 2018), Madrid, İspanya, 17–22 Haziran 2018. https://doi.org/10.1115/OMAE2018-77072
3. Guerinel, M.; Alves, M.; Sarmento, A. Nonlinear Modelling of the Dynamics of a Free Floating Body. In Proceedings of the 9th European Wave and Tidal Energy Conference (EWTEC), Southampton, İngiltere, 5–9 Eylül 2011.
4. Newman, J.N. Marine Hydrodynamics, 40th ed.; MIT Press: London, İngiltere, 2017; pp. 247–340.
5. Sarpkaya, T. Wave Forces on Offshore Structures; Cambridge University Press: Cambridge, İngiltere, 2010; pp. 109–171.
6. Park, J.; Kim, N.; Yoon, H.K.; Cho, H. Adaptive Depth Controller Design for a Submerged Body Moving Near Free Surface. Appl. Ocean Res. 2016, 58, 83–94. https://doi.org/10.1016/j.apor.2016.04.001
7. Zarzycki, K.; Ławryńczuk, M. Advanced Predictive Control for GRU and LSTM Networks. Inf. Sci. 2022, 616, 229–254. https://doi.org/10.1016/j.ins.2022.10.078
8. Ławryńczuk, M.; Zarzycki, K. LSTM and GRU Type Recurrent Neural Networks in Model Predictive Control: A Review. Neurocomputing 2025, 632, 129712. https://doi.org/10.1016/j.neucom.2025.129712

9. Wang, D.; Wan, J.; Shen, Y.; Qin, P.; He, B. Hyperparameter Optimization for the LSTM Method of AUV Model Identification Based on Q-Learning. J. Mar. Sci. Eng. 2022, 10, 1002. https://doi.org/10.3390/jmse10081002
10. Arcos-Legarda, J.; Gutiérrez, Á. Robust Model Predictive Control Based on Active Disturbance Rejection Control for a Robotic Autonomous Underwater Vehicle. J. Mar. Sci. Eng. 2023, 11, 929. https://doi.org/10.3390/jmse11050929
11. Zimmerman, S.; Nagamune, R.; Rogak, S. Wind Estimation by Multirotor Drone State Using Machine Learning with data rotation and reduction. Measurement 2022, 199, 111491. https://doi.org/10.1016/j.measurement.2022.111491
12. Mendez, A.P.; Whidborne, J.F.; Chen, L. Wind Preview-Based Model Predictive Control of Multi-Rotor UAVs Using LiDAR. Sensors 2023, 23, 3711. https://doi.org/10.3390/s23073711
13. Kim, K.H.; Jeong, C.; Kim, J.; Lee, S.; Kang, C.M. Data-Driven LSTM Model and Predictive Control for Vehicle Lateral Motion. J. Electr. Eng. Technol. 2024, 19, 3635–3644. https://doi.org/10.1007/s42835-024-01982-w.
14. Yu, Y.; Zhang, J.; Zhang, T. AUV Drift Track Prediction Method Based on a Modified Neural Network. Appl. Sci. 2022, 12, 12169. https://doi.org/10.3390/app122312169
15. Wang, G.; Zhang, D.; Wu, Z. Reinforcement Learning-Based Tracking Control for AUVs Subject to Disturbances. In Proceedings of the 34th Chinese Control and Decision Conference (CCDC), Hefei, Çin, 15–17 Ağustos 2022; pp. 2222–2227. https://doi.org/10.1109/CCDC55256.2022.10033544
16. Duan, Y.; Xiang, X.; Liu, C.; Yang, L. Double-loop LQR Depth Tracking Control of Underactuated AUV: Methodology and Comparative Experiments. Ocean Eng. 2024, 300, 117410. https://doi.org/10.1016/j.oceaneng.2024.117410.
17. Muhssin, M.T.; Ajaweed, M.N.; Khalaf, S.K. Optimal Control of Underwater Vehicle Using LQR Controller Driven by New Matrix Decision Control Algorithm. Int. J. Dyn. Control 2023, 11, 2911–2923. http://dx.doi.org/10.1007/s40435-023-01186-6
18. Turner, T.M.; Klamo, J.T.; Kwon, Y. Comparison of Wave-Induced Loads on a Near Surface Slender Body from Inviscid Flow Linear Solution and an Experimental Model Test. In Proceedings of the ASME 37th International Conference on Ocean, Offshore and Arctic Engineering (OMAE 2018), Madrid, İspanya, 17–22 Haziran 2018. https://doi.org/10.1115/OMAE2018-77760
19. Massé, C.; Gougeon, O.; Nguyen, D.; Saussié, D. Modeling and Control of a Quadcopter Flying in a Wind Field: A Comparison Between LQR and Structured H∞ Control Techniques. In Proceedings of the 2018 International Conference on Unmanned Aircraft Systems (ICUAS), Dallas, TX, ABD, 12–15 Haziran 2018; pp. 1408–1417. https://doi.org/10.1109/ICUAS.2018.8453402
20. Cummins, W.E. Hydrodynamic Forces and Moments Acting on a Slender Body of Revolution Moving under a Regular Train of Waves; Report No. 910, Navy Department, The David W. Taylor Model Basin: Washington, DC, ABD, 1954.
21. Kristia, S. Study of Wave-Induced Load Effects on a Submerged Body Near the Surface. Yüksek Lisans Tezi, Naval Postgraduate School, Monterey, CA, ABD, 2019.
22. Willy, C.J. Attitude Control of an Underwater Vehicle Subjected to Waves. Yüksek Lisans Tezi, Massachusetts Institute of Technology (MIT), Cambridge, MA, ABD, 1994.
23. Thu, A.M.; Htwe, E.E.; Win, H.H. Mathematical Modeling of a Ship Motion in Waves under Coupled Motions. Int. J. Eng. Appl. Sci. 2015, 2, 97–102.
24. Faltinsen, O.M. Sea Loads on Ships and Offshore Structures; Cambridge University Press: Cambridge, İngiltere, 1990; pp. 13–170.
25. Fossen, T.I. How to Incorporate Wind, Waves and Ocean Currents in the Marine Craft Equations of Motion. In Proceedings of the 9th IFAC Conference on Manoeuvring and Control of Marine Craft (MCMC), Arenzano, İtalyay, 19–21 Eylül 2012. https://doi.org/10.3182/20120919-3-IT-2046.00022
26. Bergdahl, L. Wave-Induced Loads and Ship Motions; Report No. 2009:1, Department of Civil and Environmental Engineering, Division of Water Environment Technology, Chalmers University of Technology: Gothenburg, İsveç, 2009; pp. 113–163.
27. Lima, F.T.; Souza, V.M.A. A Large Comparison of Normalization Methods on Time Series. Big Data Res. 2023, 34, 100407. https://doi.org/10.1016/j.bdr.2023.100407.
28. Huang, K.; Wei, K.; Li, F.; Yang, C.; Gui, W. LSTM-MPC: A Deep Learning Based Predictive Control Method for Multimode Process Control. IEEE Trans. Ind. Electron. 2023, 70, 11544–11554. https://doi.org/10.1109/TIE.2022.3229323.
29. Silva, K.M.; Maki, K.J. Data-Driven System Identification of 6-DoF Ship Motion for Enhanced Navigation of Remotely Operated Vehicles. Appl. Ocean Res. 2022, 125, 103222. https://doi.org/10.1016/j.apor.2022.103222.
30. Jung, M.; Mendes, P.R.D.C.; Önnheim, M.; Gustavsson, E. Model Predictive Control When Utilizing LSTM as Dynamic Models. Eng. Appl. Artif. Intell. 2023, 123, 106226. https://doi.org/10.1016/j.engappai.2023.106226.
31. Zamarreno, J.M. Neural Predictive Control Toolbox for CACSD in MATLAB Environment. In Proceedings of the 15th IFAC Triennial World Congress, Barcelona, İspanya, 21–26 Temmuz 2002; pp. 61–66. https://doi.org/10.3182/20020721-6-ES-1901.00916
32. Chin, C.; Lau, M.; Low, G.; Seet, G.G.L. Software for Modelling and Simulation of a Remotely-Operated Vehicle. Int. J. Simul. Model. 2006, 5, 114–125. https://doi.org/10.2507/IJSIMM05(3)3.077
33. Gunel, O.; Ankaralı, A. Modeling of Basic Propeller Thrust Test System and Thrust Control Using PID Method. In Proceedings of the 4th International Symposium on Innovative Technologies in Engineering and Science (ISITES), Alanya, Türkiye, 3–5 Kasım 2016; pp. 1663–1672.
34. Ardhuin, F.; Orfila, A. Wind Waves. In New Frontiers in Operational Oceanography; Chapter 14; GOV Institutional Publication: Fransa, 2018; pp. 396. https://doi.org/10.17125/gov2018.ch14