studies-rov-concepts

Studies of ROV concepts 2024 - Part B

Click on the octopus to return to the top of the page

01 - Development of a Control System for Underwater Vehicles with Multilink Manipulators Performing Contact Manipulation Operations

Authors: Alexander Konoplin, Nikita Krasavin, Alexander Yurmanov, Pavel Piatavin, Roman Vasilenko, and Maxim Panchuk This paper describes a new method for controlling autonomous underwater vehicles (AUVs) equipped with multilink manipulators, enabling AUVs to perform contact manipulation operations in a stabilized hovering mode near or above target objects. It outlines the process of calculating the force vector needed for the manipulator's tool to exert the desired force on an object's surface, generating control signals for additional movements, and adjusting the AUV's thrusters to maintain the necessary force. It also discusses compensating for stabilization errors to ensure continuous contact with the object. The effectiveness of this system is supported by numerical modeling and experimental data.

Authors: Can Wang, Chensheng Cheng, Chun Cao, Xinyu Guo, Guang Pan and Feihu Zhang This document discusses the limitations of existing methods, such as the Extended Kalman Filter (EKF) and Invariant EKF (IEKF), and proposes an alternative fusion filtering approach that addresses these limitations to highlight the innovation and effectiveness of this new methodology through simulations and real datasets, emphasizing its robustness and optimization capabilities in the context of underwater multi-sensor fusion.

Authors: Yueming Li, Mingquan Ma, Jian Cao, Guobin Luo, Depeng Wang and Weiqiang Chen. This document outlines AUVs' challenges in unknown and uncertain underwater environments, such as limited detection capabilities and the inability to obtain real-time global state information. It explains how multi-agent reinforcement learning (MARL) is used to improve autonomous decision-making and collaboration among AUVs by modeling the search task as a decentralized partial observation decision process and establishing a multi-AUV cooperative area search system (MACASS), allowing AUVs to perform more efficiently in both civilian and military applications.

Authors: Xiaokai Mu, Yuanhang Yi, Zhongben Zhu, Lili Zhu, Zhuo Wang, and Hongde Qin This study aims to improve speed estimation accuracy by utilizing a Self-Attention mechanism with Long Short-Term Memory (LSTM) to process various input variables. It also seeks to enhance the model's generalization capability by incorporating water-track velocity information. It highlights the effectiveness of this model in comparison to traditional methods, particularly in scenarios where persistent bottom-track velocity failures occur, as demonstrated by sea trial experiments.

Authors: Mohammad Afi f Kasno, Izzat Nadzmi Yahaya, and Jin- Woo Jung This study highlights the challenges faced by ROV operators due to the reliance on 2D camera feeds, which affect orientation awareness and accuracy. It presents a solution using nine-degrees-of-freedom sensors to capture real-time orientation data, which is then used to create a 3D model for better visualisation. It also mentions the constraints of testing due to operator agreements and presents the effectiveness of the system, noting a low mean absolute error, indicating its potential as a cost- effective solution.

Authors: Patrick Ng, and Michael Krieg This paper intends to describe the improvements made to the ArduSub system for the BlueROV2 Heavy, focusing on enhancing simulation accuracy and autonomous controller design. It details the implementation and validation of these improvements, including updates to the simulation model, replacement of the manual control algorithm, and the development of a proportional–derivative (PD) controller within a Robot Operating System (ROS) to aid augmented reality pilots, also outlining the testing processes and results.

Authors: Guangwu Song , Wei Chen, Qilong Zhou and Chenkai Guo This document describes a proposed enhancement to underwater target identification techniques using an improved version of the YOLOv8 model. It explains the modifications made to the model, such as using deformable convolution and replacing the loss function with another to improve identification accuracy and picture quality in underwater environments. This document also highlights the improved performance metrics achieved by the enhanced algorithm, demonstrating its effectiveness for the real-time detection needs of underwater robots.

Authors: Hailin Wang, Yiping Li, Shuo Li, and Gaopeng Xu This study addresses optimizing task allocation for multiple Autonomous Underwater Vehicles (AUVs) under energy constraints by conceptualizing it as a Capacitated Vehicle Routing Problem (CVRP) and using the Solving Constraint Integer Programs to find effective strategies. It highlights the significance of this research for marine exploration, environmental monitoring, and underwater construction and emphasizes the theoretical and practical contributions of the study in improving operational efficiency and mission duration in energy-restricted underwater environments.

Authors: Xuejiao Yang, Yunxiu Zhang, Rongrong Li, Xinhui Zheng, and Qifeng Zhang This paper describes the limitations of current methods, such as real-time teleoperation and preprogramming. It introduces a new method using learning from demonstrations (LfD) and dynamic movement primitives (DMP) to enhance task reproduction and generalization in complex underwater environments. The proposed underwater DMP (UDMP) method aims to address the challenges posed by the stochastic nature of underwater environments and improve operational efficiency by reducing the cognitive burden on operators. It also mentions using Gaussian mixture models for feature extraction and the transition from homomorphic to heteromorphic teleoperation modes.

15 - Adaptive Control for Underwater Simultaneous Lightwave Information and Power Transfer: A Hierarchical Deep- Reinforcement Approach

Authors: Huicheol Shin, Sangki Jeong, Seungjae Baek, and Yujae Song This study focuses on developing an advanced algorithm for optimizing energy harvesting and communication performance in an underwater optical wireless communication system. That involves a scenario where an underwater sensor communicates with a remotely operated vehicle, and the goal is to maximize energy harvesting at the sensor while maintaining communication standards. The study also outlines a hierarchical deep reinforcement learning approach and deep deterministic policy gradient-based online algorithm to achieve this objective.

Authors: Gang Sun, Shengtao Chen, Hongkun Zhou, and Fei Wan This study analyzes dual floating bodies' interaction dynamics and hydrodynamic coefficients to optimize salvage strategies. It discusses specific parameters and phases of the salvage process, such as the berthing stage, side-by-side phase, and towing phase, to enhance stability and efficiency in these operations.

17 - Autonomous Underwater Vehicle Navigation Enhancement by Optimized Side-Scan Sonar Registration and Improved Post- Processing Model Based on Factor Graph Optimization

Authors: Lin Zhang, Lianwu Guan, Jianhui Zeng, and Yanbin Gao This study aims to address the challenge of improving the accuracy of navigation data for Autonomous Underwater Vehicles (AUVs) equipped with Side-Scan Sonar (SSS) used in seabed mapping. It introduces a post-processing navigation method using Factor Graph Optimization (FGO) to enhance the accuracy of AUV trajectories by integrating various data sources and validating the method through simulations and experiments.

Authors: Jaime Alonso Restrepo-Carmona, Elkin A. Taborda, Esteban Paniagua-García, Carlos A. Escobar, Julián Sierra- Pérez, and Rafael E. Vásquez This paper presents a research study that integrates Systems Engineering (SE) methodologies with Industry technologies to enhance the design and performance of underwater robotic systems, specifically focusing on Remotely Operated Vehicles (ROVs). It aims to demonstrate how SE can systematically incorporate tools to improve mission performance and meet stakeholder expectations using a case study of an underwater exploration vehicle. It also seeks to address design challenges and provide a structured methodology for advancing the design and functionality of complex systems.

21 - Robust H∞ Control for Autonomous Underwater Vehicle’s Time- Varying Delay Systems under Unknown Random Parameter Uncertainties and Cyber-Attacks

Authors: Soundararajan Vimal Kumar and Jonghoek Kim This study addresses challenges such as time-varying delays, model uncertainties, and cyber-attacks that can affect the performance and stability of AUV systems. It aims to enhance the robustness of these systems by using advanced mathematical techniques, including Lyapunov–Krasovskii functional and linear matrix inequality, to derive new stability criteria. The research is positioned as a novel due to its comprehensive approach to handling multiple complexities simultaneously, and the effectiveness of the proposed solutions is validated through examples and simulations.

Authors: Xinglong Feng, Yuzhong Zhang, Ang Gao, and Qiao Hu This document presents research findings on improving underwater communication for autonomous underwater vehicles (AUVs) by addressing issues such as high attenuation, weak reception signals, and channel blockage. That involves using a multi-node networking communication system based on the CSMA/CA protocol, comparing the performance of OLSR and AODV protocols, and proposing a dynamic backoff method to enhance communication reliability and efficiency.

Authors: Zhiyuan Xu, Yong Shen, Zhexue Xie, and Yihua Liu This paper is a study on developing and enhancing path- planning algorithms for Autonomous Underwater Vehicles (AUVs). It introduces the Field Theory Augmented algorithm, which integrates field theory principles with the classical algorithm to address macro- global and micro-local path planning challenges. It also highlights the improvements in navigation safety and efficiency achieved by this algorithm, demonstrating its practical value for AUV operations in complex marine environments.

Authors: Guanghao Yang, Weidong Liu, Le Li, Jingming Xu, Liwei Guo, and Kang Zhang This paper explains a new approach for 3D trajectory tracking in remotely operated vehicles (ROVs) that combines deep neural networks with event-triggered nonlinear model predictive control. The text aims to highlight the advantages of this double closed-loop control system over traditional single-loop model predictive control, particularly in terms of handling external disturbances, reducing computational burden, and improving tracking performance. Additionally, it seeks to demonstrate the effectiveness of this method through simulation results.

Authors: Yuliang Wang, Han Bao, Yiping Li, and Hongbin Zhang This paper describes a technical approach for improving the control and stability of underactuated Autonomous Underwater Vehicles (AUVs) in the vertical plane that outlines the use of mathematical and control theory methods, such as SO(3) representation, exponential mapping, interval type II fuzzy systems, backstepping, and the small gain theorem, to address challenges like Euler angle singularity, quaternion ambiguity, model changes, ocean currents, and actuator saturation. It aims to convey the effectiveness of the proposed controller through simulation results.

Authors: Fushen Ren and Zhongyang Wang This document outlines the challenges addressed in the robot's design, such as modular structure, motion modes, and control systems, and highlights the use of a sliding mode control algorithm to enhance the robot's stability and efficiency underwater. It also compares this algorithm with a Proportional-Integral-Derivative (PID) control algorithm, emphasizing the advantages of the sliding mode control in terms of response speed and stability..

Authors: Raphaël Garin, Pierre-Jean Bouvet, Beatrice Tomasi, Philippe Forjonel, and Charles Vanwynsberghe This paper discusses a novel low-cost algorithm for simultaneous communication and localization of autonomous underwater vehicles (AUVs). It outlines the challenges of AUV navigation due to the limitations of radio-frequency signals underwater and the absence of GNSS, proposing a solution that involves using acoustic signals to estimate the location of an AUV by analyzing Doppler shift and range data exchanged with a fixed beacon. It also mentions evaluating this algorithm through numerical simulations and at-sea experiments.

02 - A Method for Multi-AUV Cooperative Area Search in Unknown Environment Based on Reinforcement Learning.

Authors: Yueming Li, Mingquan Ma, Jian Cao, Guobin Luo, Depeng Wang and Weiqiang Chen Multiple autonomous underwater vehicle (multi-AUV) cooperative area search technology is important for both civilian and military applications. AUVs face challenges in unknown environments and limited detection capabilities. To improve search efficiency, multi-agent reinforcement learning (MARL) is used, allowing AUVs to make better decisions by sharing information and using a coordinated search strategy.

07 - Model Reference Adaptive Control based on Neural Network for Depth of an AUV.

Authors: Fatima Zohra Kadri, Rachida Ghoul Hadiby, Kadda Boumediene, Yasmine Derkaoui, Bachir Daaou, Aicha Djellouli This study proposes a Model Reference Adaptive Control (MRAC) using multilayer perceptron neural networks for controlling the depth of a REMUS AUV amid range challenges. It aims for adaptive control through hyperstability applied to a linear vertical AUV model. A new approach, neural network model reference adaptive control (NNMRAC), combines classic MRAC with multilayer perceptron neural networks to enhance performance. Stability is analyzed using a Lyapunov function, and the effectiveness of both adaptive control strategies is evaluated through MATLAB/Simulink, providing insights on their advantages and limitations.

10 - Autonomous Visual Fish Pen Inspections for Estimating the State of Biofouling Buildup Using ROV - Extended Abstract.

Authors: Matej Fabijanic, Nadir Kapetanovic, & Nikola Miškovic The process of fish cage inspections at fish farms can be fully automated by using autonomous marine vehicles instead of trained divers. This would reduce costs and risks. The work proposes a complete solution involving developing a control algorithm for a remotely operated vehicle (ROV) and an image processing framework to estimate biofouling buildup. Experimental results demonstrate the effectiveness of the system for autonomous inspections and accurate biofouling assessments, showing promise for the aquaculture industry.

11 - AUV path planning based on improved IFDS and deep reinforcement learning.

Authors: Fan Yiqun, Li Hongna, Xie Jiaqi, and Zhou Yunfu Existing autonomous underwater vehicle (AUV) path planning algorithms are improving but perform poorly in real environments due to currents. This paper proposes an algorithm that combines perturbed flow field and deep reinforcement learning, improving obstacle avoidance and adaptability to real environments. The study compares four algorithms in various scenarios, showing that the proposed method allows AUVs to plan paths dynamically in unknown environments.

12 - Modelling of ROV-Operated Disconnectable Mooring Systems in Wave Dominated Environments.

Authors: Jonathan Glasspool, Francesco Giorgio-Serchi, Ajit C. Pillai Existing autonomous underwater vehicle (AUV) path planning algorithms are improving but perform poorly in real environments due to currents. This paper proposes an algorithm that combines perturbed flow field and deep reinforcement learning, improving obstacle avoidance and adaptability to real environments. The study compares four algorithms in various scenarios, showing that the proposed method allows AUVs to plan paths dynamically in unknown environments.

19 - Quantitative Comparison of ROV and Diver-Based Photogrammetry to Reconstruct Maerl Bed Ecosystems.

Authors: Ona L. R. Paterson, Kathryn E. Dawson, Andrew O. M. Mogg, Martin D. J. Sayer, Heidi L. Burdett Floating offshore wind turbines (FOWTs) have great potential for offshore wind resources. This paper analyzes mooring line disconnection using ROV technology, focusing on mooring connector depth at Arven Offshore Wind Farm. Results show deeper installations improve connector availability, enhancing operational efficiency. The study highlights the need to integrate O&M strategies with early mooring design considerations.

20 - Comparison of a tower geodetic micro-network optimization results obtained using the MABAC, MAIRCA, COCOSO and ROV methods with those obtained applying the VIKOR method.

Author: Darko Andie In this study, the author optimizes a tower geodetic micro- network using the MABAC (Multi-Attributive Border Approximation Area Comparison), MAIRCA (Multi- Attributive Ideal-Real Comparative Analysis), COCOSO (Combined Compromise Solution), and ROV methods. These methods were compared with the VIKOR (VIseKriterijumska Optimizacija I Kompromisno Resenje) method, developed to solve multi-criteria decision-making problems, which is proven effective. Criteria for precision and reliability were established, and four acceptable network solutions were ranked.

Authors: Nikolaos Manos, Ergina Kavallieratou, Nikos Vasilopoulos Kalypso is a 3D-printed underwater robot designed to improve aquaculture management in the Mediterranean Sea for species like sea bream and catfish. It operates in two modes: autonomously for inspections and remotely for complex tasks. It helps detect problems and allows direct intervention. Kalypso enhances maintenance, fish welfare, and resource use, making it a valuable tool for Mediterranean fish farms.

31 - Kalypso: An inspection AUV for aquaculture.

33 - USV-AUV Collaboration Framework for Underwater Tasks under Extreme Sea Conditions.

Authors: Jingzehua Xu, Guanwen Xie, Xinqi Wang, Yimian Ding, Shuai Zhang Autonomous underwater vehicles (AUVs) are useful for exploring oceans but struggle in tough sea conditions. This study presents a framework for AUV and unmanned surface vehicle (USV) collaboration, improving multi-AUV positioning and coordination. Extensive simulations confirm the framework's effectiveness in extreme conditions, and the simulation code is open-source.

28 - A Hierarchical Planning Method for AUV Search Tasks Based on the Snake Optimization Algorithm.

Authors: Zhiwen Wen, Zhong Wang, Xiangdong Wen, Chenxi Niu, Pei Wang, and Daming Zhou In a complex battlefield, enabling autonomous underwater vehicles (AUVs) to quickly reach dynamic targets is crucial for search tasks. This study proposes a hierarchical AUV task planning method combining hierarchical programming and a snake optimization algorithm. The method includes three planning levels focused on minimizing encounter and navigation times while optimizing trajectories. Simulation experiments verified its effectiveness, demonstrating that it can plan AUV searches efficiently. This method holds practical significance for similar underwater task planning issues.

29 - Design and Computational Modelling of AUV Tunnel Thruster Covers for Effi cient Operation

Authors: Christopher McNeill, Zachary Cooper-Baldock, and Karl Sammut Autonomous underwater vehicles are widely used for mapping, surveillance, and inspections. Hovering AUVs have limited energy due to their design, requiring efficient hydrodynamic designs. This work suggests adding a retractable thruster cover to improve propulsion efficiency. A comparison shows that this cover reduces drag by 9. 51%, which enhances the AUV's operational efficiency, making it suitable for energy-limited vehicles.

30 - Allocation Strategy Optimization Using Repulsion-Enhanced Quantum Particle Swarm Optimization for Multi-AUV Systems.

Authors: Changjian Lin, Dan Yu, and Shibo Lin In multi-autonomous underwater vehicle (multi-AUV) operations, target assignment is a multi-objective allocation problem. Strategy selection depends on the non-cooperative environment. This paper creates a system to evaluate this environment and develops a target allocation model using bi-matrix game theory. The model includes a payoff matrix based on various factors. A simulation experiment validates the model and shows the effectiveness of the repulsion process quantum particle swarm optimization (RPQPSO) algorithm in multi-AUV scenarios.

Authors: Jefferson Osowsky, Tin Muskardin, Leonardo M. Barreira This work presents a project to convert a commercial Remote Operated Vehicle (ROV) into a Remotely Piloted ROV using two approaches. The first approach developed a serial/ethernet converter with an ESP32 microcontroller to transfer data between the hand controller and the ROV control console. The second approach replaced the ROV remote control with a Single Board Computer (SBC) to emulate the remote control signals. The development and results of these methods will be presented.

32 - Remote Piloting Development of an ROV