Accurate and continuous wave-direction perception is an essential prerequisite for shipborne marine environmental awareness, intelligent navigation, route optimization, and safe offshore operations. However, vision-based wave-direction estimation onboard a moving vessel remains challenging, as the observed sea-surface texture is highly susceptible to nighttime low illumination, partial rain–fog occlusion, strong specular reflections, and dynamic vessel motions such as roll, pitch, and yaw. These factors may lead to texture degradation, viewpoint disturbances, image blur, and unstable wave-direction regression. To address these challenges, this paper proposes a multispectral monocular-vision and inertial measurement unit (IMU)-assisted method for shipborne wave-direction estimation to meet all-weather sensing requirements. Dominant wave-direction estimation is treated as the core task, and the feasibility of the proposed method is validated using limited field data collected in real shipborne scenarios.

A visible-light and long-wave infrared multispectral imaging system with synchronized attitude acquisition is first constructed. The image streams and IMU measurements are temporally aligned using timestamps, and sequential samples are generated via a sliding-window strategy. To enhance the visibility and stability of sea-surface texture under nonuniform illumination and low-contrast conditions, contrast limited adaptive histogram equalization (CLAHE) is applied to the input images. A U-Net-based segmentation network is then employed to extract the effective sea-surface region while suppressing interference from the sky, vessel structures, wake, and localized high-intensity reflections. On this basis, a ResNet-18 backbone initialized with transfer learning is used to encode visual features from each frame. The synchronized vessel attitude angles are represented using sine–cosine encoding to avoid angular discontinuities and are further embedded into the same feature space as the visual representations via a multilayer perceptron. The visual and attitude features are fused and fed into a lightweight multi-head self-attention module, which models cross-frame temporal dependencies and learns stable directional representations from evolving wave textures. The final wave direction is represented as a two-dimensional unit vector on the angular circle and is recovered using the arctangent function, thereby mitigating discontinuities near the 0°/360° boundary. In a subset of system-level validation experiments, three orthogonally arranged cameras perform independent inference. Their outputs are verified using geometric consistency constraints, and valid results are fused in post-processing to improve output stability.

Field experiments were conducted in the Guangzhou–Zhuhai coastal waters. Under visible-light conditions, the proposed method achieved a mean absolute error of 0.41°, a standard deviation of 0.28°, and a success rate of 99.90% for wave-direction estimation. Under long-wave infrared conditions, the mean absolute error was 1.14°, the standard deviation was 0.89°, and the success rate was 98.94%. These results show that visible-light imagery provides clear wave-crest and wave-trough texture cues under daytime illumination, whereas long-wave infrared imaging still preserves effective sea-surface texture information under nighttime and low-illumination conditions included in this study. The multispectral configuration therefore improves the continuity of shipborne wave-direction estimation across varying illumination conditions. In addition, the incorporation of IMU-based attitude information and temporal attention modeling helps mitigate the effects of vessel motion and single-frame fluctuations, resulting in more stable wave-direction estimation over time.

The proposed multispectral monocular-vision–IMU multimodal temporal estimation method shows good stability on the collected visible-light daytime samples, long-wave infrared low-light/nighttime samples, and a limited set of complex-environment samples. The results indicate that the integration of multispectral imaging, attitude assistance, effective sea-surface extraction, and temporal correlation modeling is feasible for shipborne dominant wave-direction estimation. It should be noted that the current dataset does not systematically cover varying rain intensities, fog conditions, or different wind–wave coupling regimes. Therefore, the results should be regarded as a feasibility validation based on limited field data rather than a comprehensive demonstration of adaptability to all-weather operating conditions. Future work will focus on expanding the field dataset, improving model robustness under severe weather and complex sea states, and extending the framework from dominant wave-direction estimation to more comprehensive wave-parameter perception.

The rapid integration of artificial intelligence (AI) into maritime technology has driven unprecedented advancements in unmanned surface vehicles (USVs), positioning them as a crucial force in future maritime operations and military transformations. The intelligent navigation system is the core of USVs, responsible for environmental perception, decision-making, and motion control, which collectively enable autonomous mission execution and integration into systematic operations. This study provides a comprehensive review of the fundamental technologies underpinning USV intelligent navigation, critically evaluates existing challenges, and proposes future research directions to advance and expand the application of intelligent navigation technologies for USVs. The research aims to bridge existing knowledge gaps, providing a foundation for the further development of autonomous maritime systems.

This research provides a comprehensive review of the current state of intelligent navigation technologies for USVs, focusing on three critical areas: environmental perception, decision-making and planning, and motion control. (a) In the domain of environmental perception, the primary sensing modalities include visible light, infrared, sonar, electromagnetic signals, navigation radar, and LiDAR. With advancements in multi-source information fusion technology, perception techniques have evolved from relying on single sensors to utilizing multi-sensor fusion, transitioning from object-level fusion to feature-level fusion. Despite these advancements, achieving accurate and efficient environmental perception remains a key challenge. The ability to provide real-time, comprehensive environmental awareness is essential for USVs to navigate autonomously in complex maritime conditions. (b) For decision-making and planning, a variety of methodologies, including operations research, optimization algorithms, and AI-based approaches, have been employed to generate optimal decisions under multiple constraints, such as mission parameters, payload configurations, and environmental factors. Existing technologies facilitate global path optimization, target tracking, and emergency collision avoidance under predefined conditions. However, challenges remain in multi-objective adversarial decision-making and path planning in highly dynamic and adversarial environments, especially under strong external interferences. The ability to enhance decision-making robustness in these scenarios is crucial for advancing autonomous USV capabilities. (c) In motion control, various algorithms such as proportional-integral-derivative (PID) control, model predictive control (MPC), model-free adaptive control (MFAC), linear quadratic regulators (LQR), robust control, and sliding mode control have been applied to achieve accurate trajectory tracking, course keeping, and speed regulation. Current advancements allow for precise control under design conditions; however, adaptive control remains a challenge in scenarios with extreme environmental variations. Moreover, effective control of roll and pitch motions remains underdeveloped, limiting USV stability in high sea states. Motion control techniques serve as the foundation of USV intelligent navigation, ensuring the successful implementation of autonomous navigation systems.

The study identifies key limitations. In environmental perception, while current technologies allow for target detection and identification in open seas, their accuracy significantly decreases under adverse weather conditions, such as fog, heavy rain, and high sea states. Real-time wave field perception remains inadequate, further compromising navigation safety in dynamic operational conditions. Decision-making and planning algorithms, though effective in structured mission scenarios, struggle with the complexity of dynamic constraints, adversarial interactions, and unexpected environmental disturbances in real operations. Motion control strategies, while efficient under nominal operating conditions, require enhanced adaptability to handle sudden environmental shifts and complex vessel dynamics. The inability to manage roll and pitch movements effectively limits the operational capability of USVs in high sea states.

To address these challenges, the study proposes four key technological advancements and research directions. First, the development of high-precision six-degree-of-freedom (6-DoF) motion modeling for USVs will provide a robust framework for intelligent navigation algorithms. Second, the integration of large-scale AI models for multi-modal perception and decision-making will enhance autonomous situational awareness and adaptive response capabilities. Third, the advancement of high-sea-state adaptive navigation technologies through real-time wave observation and predictive motion modeling will significantly improve USV stability and safety in complex maritime environments. Additionally, the incorporation of real-time optimization techniques will enhance navigation efficiency under operational constraints. These technological developments will not only expand the application scope of USVs but also significantly improve their autonomous navigation capabilities. The study emphasizes the necessity of interdisciplinary research efforts that integrate AI-driven models, control theory, and maritime engineering to accelerate the development of fully autonomous USVs capable of performing in diverse operational conditions.

This paper presents a novel approach to the precise control of variable mass unmanned surface vehicles (USVs) during payload deployment tasks, addressing the control challenges caused by unpredictable variations in both mass and draught. The primary objective is to propose an adaptive control method that can effectively adapt to these unknown variations in mass and draught, thereby ensuring the stable and reliable operation of the USV under complex and dynamic mass conditions.

First, regarding to the motion modeling of variable-mass USVs, this study analyzes the impact mechanism of mass variations on the hydrodynamic characteristics of the vehicle. It also analyzes how these variations, through changes in the parameters of the dynamic model, affect the vehicle's motion state. To address the issue that current controller design models are insufficient in analytically and intuitively representing this coupling influencing process, we use the draught term as the reference variable. The progressive coupling relationships among draught and the mass term, added mass term, added moment of inertia term, and various hydrodynamic derivatives are systematically analyzed. Based on this analysis, the mathematical model for the maneuvering motion of the variable-mass USV is then constructed. Secondly, to design an effective estimation method, a super-twisting sliding mode observer is proposed for estimating the unknown draught and mass of the variable-mass USV. This method is based on an analysis of the coupling relationships between mass variations and the vehicle's motion state and control inputs, as described in the maneuvering model of the USV. Subsequently, addressing the motion control problem of variable-mass USVs under unknown mass variations, we propose an adaptive speed control strategy based on the sliding mode observer. Specifically, leveraging the maneuvering motion mathematical model of the variable-mass USV and the draught observations from the sliding mode observer, a feedback linearization method is used to design the adaptive speed control algorithm. The asymptotic stability of the proposed control algorithm is proved using the Lyapunov theory.

A series of simulation experiments are conducted to validate the proposed method. In the mass step-change observation experiment, the super-twisting sliding mode observer demonstrates satisfactory performance. Compared to the traditional sliding mode observer, the average observation errors of the draught and mass are significantly reduced by 43.75% and 43.76%, respectively. Furthermore, it shows rapid convergence when mass changes occur suddenly. In the continuous mass change observation experiment, the observer also performs excellently, exhibiting fast convergence and high accuracy, thus demonstrating significant advantages compared to the traditional observer. The speed control experiments reveal that the designed adaptive speed control algorithm can stably track the target speed under both mass step-change and continuous-change conditions. Although it may require slightly more adjustment time compared to the traditional Backstepping controller, it offers significant advantages in handling variations in mass and draught, achieving superior control performance. In the environmental disturbance experiment, while the adaptive control algorithm maintains stable speed control, demonstrating a certain degree of robustness, it also highlights the need for further improvement in the draught observation method to enhance its disturbance rejection capabilities.

The control algorithm proposed in this paper is well-suited for control scenarios involving unknown mass variations, such as payload launch or agricultural dispensing operations. Future research should focus on mitigating the impact of external environmental disturbances on observation accuracy and enhancing the robustness of the observation algorithm to better handle such disturbances.

To solve the problem of the autonomous and safe landing of multi-rotor unmanned aerial vehicles (UAVs) in the application of unmanned platforms at sea, this paper studies multi-rotor UAV landing devices and their safety boundary assessment methods.

Based on the idea of the "Harpoon-Grid" rapid securing system, a new type of "V-foot–net" landing device is designed for multi-rotor UAVs. Multi-rigid body dynamic models of the UAV and deck are established, and simulations of the landing process are carried out using ADAMS under different rolling amplitudes and period conditions to obtain the attitude change process, landing critical period and attitude stabilization time of the UAV.

The simulation results show that the proposed "V-foot–net" device has higher docking redundancy under a smaller rolling period and larger rolling amplitude, and achieves attitude stability faster than ordinary landing devices.

The results show that this device has superior safety compared to ordinary devices, and its safety boundary obtained through simulation can provide valuable references for improving the reliability of the autonomous landing of UAVs on offshore platforms.

In order to solve the multi-objective collision avoidance problem of unmanned surface vehicles (USVs) in open waters, this paper takes a medium-sized USV as the object and carries out the research and application exploration of a local path planning algorithm based on the Convention on the International Regulations for Preventing Collisions at Sea (COLREGS).

A virtual obstacle line method is proposed to load the constraints of COLREGS based on the RRT algorithm for meeting the practical requirements of collision avoidance path planning in open water. In light of the problem that the RRT algorithm does not consider the speed dimension, a velocity obstacle (VO) algorithm is introduced. Next, a VO-RRT fusion algorithm is proposed and given the most dangerous obstacle strategy in order to solve the problem of real-time collision avoidance under multi-objective conditions.

The simulation and real ship test results show that the proposed algorithm has better real-time performance and takes less than 50 ms to undertake path re-planning. The collision avoidance path planned by the algorithm meets the relevant requirements of Articles 6, 8 and 13–18 of COLREGS. The proposed method can effectively deal with the multi-target collision avoidance problem in open waters.

The proposed method has good rule compliance, real-time and safety performance, and high practical value.