This paper focuses on the three-dimensional trajectory tracking of autonomous underwater vehicles (AUVs) and proposes a nonlinear disturbance observer-based sliding mode control (NDO-SMC) scheme that operates without linear velocity and acceleration measurements. First, a four-degree-of-freedom AUV model accounting for uncertainties and external disturbances is established. Then, to tackle the issue of unmeasurable linear velocity, a second-order differentiator is proposed for velocity estimation, which guarantees bounded velocity estimation errors. Next, an NDO-SMC scheme is developed that eliminates the requirement for linear velocity and acceleration feedback, with theoretical analysis verifying bounded trajectory tracking errors under uncertainties and external disturbances. Finally, comprehensive simulations and experiments are conducted to validate the proposed method, confirming its strong robustness across multiple scenarios.

To address the problems of unknown target states, difficult terminal alignment, and safety control in complex environments during the stern-ramp recovery of a waterjetpropelled underactuated USV onto a moving mothership, this paper proposes a dynamic autonomous recovery method integrating state estimation, distance-based switching guidance, and finite-time control. The proposed method uses relative pose information to achieve online estimation of the mothership motion states, employs distance-based switching guidance and control barrier function to balance recovery accuracy and process safety, and incorporates finite-time observation and control to improve system robustness against uncertainties and environmental disturbances. Simulation results show that the proposed method achieves higher recovery efficiency and terminal accuracy in the obstacle-free scenario, and can achieve safe and stable recovery in complex scenarios with obstacles, measurement noise, and environmental disturbances.

Due to the lack of side thrusters, three-dimensional trajectory tracking control of underactuated autonomous underwater vehicles (AUVs) presents a significant challenge. current solutions, such as angular guidance and feedback linearization, often suffer from delayed responses or overly complex models. to overcome these limitations, this paper proposes a novel velocity guidance law (VGL) derived by the backstepping method. compared to existing guidance approaches, the proposed method reduces complexity while ensuring accurate trajectory tracking. A rigorous theoretical analysis is conducted to guarantee the stabilization of both position and orientation, effectively addressing the backward motion phenomenon observed in prior studies. for operations in constrained environments, the proposed method is extended with prescribed performance to maintain tracking errors within predefined time-varying bounds. additionally, an adaptive velocity control strategy based on nonlinear disturbance observer is proposed without assuming bounded uncertainties. simulation and experimental results demonstrate the effectiveness and superior performance of the proposed methods compared to existing techniques (c) 2026 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.

Most existing trajectory-tracking methods for underactuated unmanned surface vehicles (USVs) rely on pre-planned and sufficiently smooth desired trajectories. However, when obstacles appear online, the original trajectory may become locally unsafe or difficult to track. This paper investigates obstacle-aware trajectory tracking of underactuated USVs by introducing an online reference-point generation framework. A trajectory-progressive reference-point generator with a predictive repulsive field (TPG-PRF) is first developed to generate a safe and trackable reference point before lower-layer guidance and control are executed. The proposed generator combines a look-ahead attractive field, a predictive repulsive field, and a direction-speed decoupled update law to improve local obstacle-response smoothness while preserving trajectory progression. Then, a prescribed-performance guidance law and a prescribed-time dynamic controller are designed to ensure constrained tracking under lumped disturbances and input saturation. Theoretical analysis shows the boundedness of the cascaded closed-loop system. Simulations under Sea State 3 disturbances, including straight-line, circular, and sinusoidal trajectories with static and dynamic obstacles, demonstrate improved reference smoothness and tracking performance, reduced input variation, and low computational cost. Lake experiments further verify the onboard feasibility of the proposed generator under real-time LiDAR-based perception.

To enable real-time detection and continuous tracking for short-range recovery of Autonomous Underwater Vehicles (AUVs) in complex underwater environments, this paper proposes a lightweight detection-and-tracking framework comprising AUV-DETR and TD-ByteTrack. A split-before-augmentation protocol with perceptual similarity screening is adopted to prevent data leakage and ensure training data diversity. For detection, AUVDETR builds upon RT-DETRv1 by incorporating an ultra-lightweight HWDNet backbone, a wavelet-based HWD downsampling module, a lightweight L-HIFA feature fusion module, and a ShapeIoU regression strategy, achieving 98.74% precision, 95.83% recall, and 91.85% mAP0.5:0.95 at 60.6 FPS with only 9.8M parameters and 23.85 GFLOPs. For tracking, TD-ByteTrack introduces a Time-Distance Matching Module and an ID Delayed Assignment mechanism to leverage temporal and spatial information for robust re-association after occlusion and suppress premature identity creation, significantly reducing ID switches and improving tracking stability. When integrated on underwater video datasets, the framework achieves 86.10% MOTA, 81.95% HOTA, and 82.68% IDF1 with only 4 ID switches at 34 FPS on a CPU-only desktop platform. Extensive experiments including ablation studies, comparative evaluations, and cross-sequence generalization validation confirm the effectiveness and robustness of the proposed framework in both controlled laboratory and challenging real underwater environments.

This paper presents a novel state-dependent. 8 function-based adaptive sliding mode control (ASMC) scheme for systems with unknown parameters and unbounded uncertainties, guaranteeing predefined performance specifications. The proposed method fundamentally resolves the stability breakdown problem inherent in conventional barrier function-based control approaches under such challenging conditions. Unlike existing ASMC approaches that only guarantee globally uniformly ultimately bounded (GUUB) stability, our method incorporates an adaptive parameter adjustment mechanism to prevent gain over-increasing while strictly maintaining predefined performance bounds. The proposed control scheme is extended to systems with unmeasured states and applied to the control of autonomous underwater vehicle (AUV). Theoretical analysis further supports the validity of the approach. Comparative simulation results demonstrate its superior performance in handling system uncertainties and precisely satisfying predefined control requirements. Copyright (c) 2025 The Authors. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)

The chattering and model uncertainty are two main challenges for the application of sliding mode control (SMC). To address these issues, a novel class infinity function-based SMC combined with disturbance observer is proposed for the control of remotely operated vehicles (ROVs). First, a novel class infinity function-based SMC is proposed, ensuring the global uniform ultimate boundedness (GUUB) of ROV without model parameters. Then, to improve performance under strong external disturbances, a disturbance observer is designed to estimate and compensate for system unknows, with a rigorous stability analysis of the controller-observer structure. At last, the proposed method is extended to operate without acceleration measurement. Simulations and experiments are conducted to verify the effectiveness of the proposed control scheme.

The swarm reclamation control strategy is designed to decouple the reclamation task into three parts: target allocation, swarm control, and reclamation scheduling. On this basis, the Hungarian algorithm is used to assign virtual berths to each unmanned surface vehicle (USV) in the swarm. Based on the approach planning of Dubins curve, the requirements of fast approach and smooth transition for USV are met, and the genetic algorithm is used to optimize swarms scheduling and generate the recovery scheduling schedule. To minimize swarm recovery time and maximize resource utilization, swarm recovery task allocation, navigation, scheduling, and other processes are realized. The simulation results show that the proposed method allows the unmanned surface swarms to exhibit a high degree of robustness and flexibility.

In this article, a full-order sliding mode controller with echo state network (ESN-FOSMC) disturbance compensation is proposed for the trajectory tracking and vibration suppression of underwater flexible manipulators (UFM). To improve the robustness under lumped disturbances and reduce the computational complexity of traditional recurrent neural networks method, ESN, a continuous recurrent neural network, is used to approximate and compensate model uncertainties and hydrodynamic disturbances. A FOSMC is designed to ensure accurate tracking of joints and end-effectors, and proportional-derivative (PD) control method is utilized to further suppress flexible vibration. The adaptive law of ESN is formulated by using Lyapunov method, integrating the ESN method with sliding mode control method. Then, Lyapunov method is used to prove the stability of the control system. Finally, the virtual prototype system of the UFM is established to validate the effectiveness of the proposed control method. Simulation results present that, compared with the nonsingular fast terminal sliding mode controller and a FOSMC with radical basis functions (RBF-FOSMC) neural network disturbance compensation, the ESN-FOSMC achieves superior tracking accuracy with reduced vibration.

In the field of cooperative trajectory tracking control for unmanned surface vehicle (USV) swarms, model predictive control (MPC) offers notable advantages in fault tolerance and robustness. However, existing studies have overlooked collision avoidance challenges inherent to practical maritime applications. Addressing these critical issues in USV swarms control, this paper innovatively proposes a dual-loop MPC algorithm for trajectory tracking and swarms collision avoidance, whose core components comprise: 1) Based on USV models, a dual-loop MPC algorithm is proposed. The outer-loop NMPC optimizes reference velocities, while the inner-loop LMPC computes optimal control inputs for trajectory tracking. 2) A collision avoidance strategy integrating obstacle-USV and USV-USV repulsive constraints into the NMPC constraint design, ensuring collision-free navigation amid marine obstacles and swarms formations. Numerical simulations validate the efficacy of the proposed method. Results demonstrate that the algorithm exhibits exceptional performance in both trajectory tracking accuracy and collision avoidance reliability.

This paper presents a composite controller for trajectory tracking of moving-base underwater flexible manipulators (UFM). Firstly, a dynamics model of the moving-base UFM is established, and the model is decomposed into a slow-varying subsystem and a fast-varying subsystem by using the singular perturbation method. Then, an adaptive non-singular fixed-time sliding mode controller based on a high-order sliding mode observer is proposed for the slow-varying subsystem. In this controller, the high-order sliding mode (HOSM) observers are used to estimate and compensate for lumped disturbances, and adaptive super-twisting algorithm is used to reduce sliding mode chattering, which overcomes the disadvantage that the traditional adaptive method is prone to overestimation. To further suppress the system chattering, a HOSM observer is used to obtain the flexible mode derivatives for the fast-varying subsystem to achieve the suppression of vibration modes. The main advantages of this controller are its non-singularity, fast finite-time convergence and good vibration suppression. Extensive simulation results have validated the effectiveness of the proposed control method.

A novel control approach merging deep reinforcement learning (DRL) and linear active disturbance rejection control (LADRC) is proposed to improve the recovery success rate of underactuated unmanned surface vehicles (USVs) under environmental disturbances. Firstly, the USV’s three degree-of-freedom model and models of environmental disturbances are established. Aiming at recovery guidance for underactuated USVs, an improved line-of-sight (LOS) guidance law is adopted to calculate the desired heading angle and surge speed. Then, LADRC is used to stabilize the tracking errors of heading angle and surge speed, with the twin delayed deep deterministic policy gradient (TD3) algorithm adjusting its parameters in real time. Simulation results validate that the proposed approach is efficient in the recovery control of the USV and superior to conventional LADRC, even in the presence of complex environmental disturbances.

It is of great significance to expand the functions of submarines by carrying underwater manipulators with a large working space. To suppress the flexible vibration of underwater manipulators, an improved sparrow search algorithm (ISSA) combining an elite strategy and a sine algorithm is proposed for the trajectory planning of underwater flexible manipulators. In this method, the vibration evaluation function is established based on the precise dynamic model of the underwater flexible manipulator and considering complex motion and vibration constraints. Simulation results show that the ISSA algorithm requires only 1/3.68 of the time of PSO. Compared to PSO, SSA and the opposition-based learning sparrow search algorithm (OBLSSA), the optimization performance is improved by 17.3%, 13.1% and 9.7%, respectively. However, because the complex dynamics model of the underwater flexible manipulator leads to large computational effort and a long optimization time, ISSA is difficult to apply directly in practice. To obtain a large number of optimization results in a shorter time, an incremental Kriging-assisted ISSA (IKA-ISSA) is proposed in this paper. Simulation results show that IKA-ISSA has good nonlinear approximation ability and the optimization time is only 3% of that of the ISSA.

An adaptive fixed-time backstepping control is proposed to achieve the three-dimensional trajectory tracking control of an underactuated autonomous underwater vehicle (AUV) in the presence of model uncertainty and external disturbances. In this paper, the dynamics of the AUV in terms of five degrees of freedom (DOFs) are discussed. Considering it is an underactuated AUV, a virtual velocity guidance law is derived using the backstepping method. For velocity convergence, an adaptive fixed-time control is derived without model parameters, with adaptive adjusting law tackling system unknows. Theoretical analyses demonstrate that the tracking error converges to a small bounded field within a fixed time in the proposed control scheme. The effectiveness and superiority of the proposed method are verified by simulation results.

Aiming at the attitude control of unmanned underwater vehicle (UUV), a control moment gyros (CMGs) control system is designed based on ARM stm32f407, which owns many advantages such as large output torque and does not depend on fluid motion. Firstly, this paper briefly describes the system attitude control mechanism. Then the hardware design scheme is proposed with detailed design and type selection. Considering the requirements of the control task, the lower computer control program and the upper computer control interface are designed. Finally, the system is tested and verified by experiment. The results show that the designed system can effectively control the speed of execution motors, so as to output the required torque and meet the attitude control demand of UUV.

This article renders the trajectory tracking of underactuated autonomous underwater vehicles (AUVs) in horizon plane. Considering it is hard to obtain dynamic model, a model-free controller is proposed based on deep reinforcement learning (DRL). To accelerate the deployment, backstepping method is adopted to simplify the model. Then a DRL agent is trained by deep deterministic policy gradient (DDPG) for the convergence of velocity. Simulation and comparison to another model-free control, i.e. backstepping PID, and a model-parameter-free control are conducted, verifying the effectiveness of the proposed control scheme.

ObjectivesAiming at the high-precision recovery guidance control requirements of current stern ramp recovery technology, a self-adaptive cascade tracking control method for unmanned surface vessels (USVs) is proposed specifically for stern ramp recovery. MethodsBased on the technical requirements of stern ramp recovery, a motion model of an underactuated USV is established, and the generalized Kalman filter (GKF) algorithm is used to predict the navigation state and recovery position of the mother ship. Introducing the idea of constant bearing guidance combined with the sliding mode variable structure control theory, a stable cascade control system is constructed to solve tracking control problems during the recovery process. ResultsIt is proven that the USV can stably track the target, by analyzing the stability of the system through the Lyapunov theory and cascade theorem. ConclusionsThe simulation results show that the proposed control method gives the USV stable tracking performance and strong robustness against uncertain disturbances.

该文基于模块化设计理念,对一款水下清洗机器人控制系统软件进行了设计.分析了软件系统结构,并在PyQt5界面开发框架下对上位机软件中的系统通信、运动控制、界面视频显示、位姿展示等关键功能的实现方式进行了详细阐述.针对清洗机器人水下控制系统存在的可能离线"失控"、模拟量信号干扰和噪声、推进器正反转状态不能直接判断等问题分别给出了解决方案.水池软件测试结果表明,所设计的水下清洗机器人控制系统软件具有人机交互性好、数据获取实时性高、控制稳定等特点.

This paper proposes a composite controller (CC) to improve the accuracy of trajectory tracking and suppress the vibration of two-link underwater flexible manipulators. A dynamic model of the flexible manipulators considering hydrodynamic force is established by combining the Lagrange equation and Morison formula. Then, the dynamic model is divided into a flexible dynamic subsystem and rigid dynamic subsystem, and a decomposed dynamic control strategy is presented for the two subsystems. In particular, an adaptive fuzzy sliding mode control scheme (AFSMC) with good robustness to compensate for uncertain factors is designed to track the joint trajectory and suppress vibration. Next, the trajectory tracking control of two-link underwater flexible manipulators is simulated to investigate the performance of the framework. The results show that the hydrodynamic force and flexible deformation markedly affect the input torque of the joint, and the traditional sliding mode controller (SMC) is superior to proportional integral derivative (PID) control in managing hydrodynamic force disturbance and inferior in suppressing flexible vibration. The proposed composite controller based on adaptive fuzzy sliding mode control CC(AFSMC) is more effective in restraining the vibration of flexible manipulators and resisting hydrodynamic force disturbance than PID and CC(SMC).

A new underwater two-part towed system is proposed and studied for the test of near-surface towed vehicles. The near-surface towed vehicle is towed by submarines or the large displacement unmanned undersea vehicle, and plays the role of the communication relay with the onshore station. To deal with the influence of the harsh wave environment and create a stable test platform, a fuzzy-PID control method and an active heave compensation device are applied in the towed system. The mathematical model of the two-part towed system is built. In the model, the nonlinear aspects are considered, such as the interaction of the multi-body, the flexibility of the cable, the wave force, the heave compensation. Based on the model, the simulation results indicate that the Fuzzy-PID control method can significantly reduce the depth fluctuation of the towed vehicle under different sea state. Besides, it is also demonstrated that the active heave compensation device can obviously improve the stability performance of the two-part towed system by greatly reducing the depth fluctuation of the ballast, and create a very stable test environment similar to the submarine or large underwater vehicle.