To address the altitude-constrained path planning problem for Unmanned Underwater Vehicles (UUVs) in complex terrain environments, this paper proposes an improved hybrid A^* path planning algorithm based on UUV motion characteristics, termed UUVA^*. Firstly, the traditional hybrid A^* algorithm is extended by expanding the state space and incorporating the nonholonomic motion constraints of UUVs. Secondly, motion primitives tailored to UUV navigation characteristics are designed, and the cost function is optimized to generate smoother and more navigable paths. Finally, to enhance search efficiency, a multi-resolution hierarchical search strategy and branch pruning mechanism are introduced, significantly reducing the number of node expansions. Simulation tests conducted in two representative complex terrains—submarine hills, and submarine canyons—verify the altitude-constrained planning capability of the proposed algorithm. Furthermore, comparative experiments against classical A^* algorithm, hybrid A^*, and RRT^* algorithms are conducted. The result shows that the generated paths are smoother and yield smaller tracking errors, making the proposed algorithm more suitable for practical UUV applications, which verifies the superior performance of the proposed method in terms of trajectory tracking error and path navigability.

This study aims to improve the accuracy of ship motion prediction by proposing a gray-box modeling method for ship maneuvering motion, termed AE-LS-SVM, which combines the Alpha Evolution (AE) algorithm to optimize the Least Squares Support Vector Machine (LS-SVM). The AE algorithm's superior global search capability was employed to optimize the hyperparameters of LS-SVM. The dataset in this study was obtained through ship maneuvering tests conducted using the Marine Systems Simulator (MSS). Based on the Abkowitz model, a three-Degree-of-Freedom (3-DOF) gray-box model of ship maneuvering motion was constructed, incorporating maneuvering variables and ship responses. The gray-box model established in this study was then subjected to comparative analysis with gray-box models developed using GWO-LS-SVM (Grey Wolf Optimizer) and PSO-LS-SVM (Particle Swarm Optimization) through maneuvering tests. Experimental results demonstrate that the AE-LS-SVM method achieves reductions in Mean Squared Error (MSE) of approximately 6% and 4%, reductions in Mean Absolute Percentage Error (MAPE) of 3.84% and 1.96%, and improvements in correlation coefficient (CC) of 0.13% and 0.04%, respectively, compared to the two benchmark methods. Additionally, the anti-interference capability of the proposed method was validated by introducing varying levels of noise into the training dataset to simulate measurement errors. The findings indicate that the AE-LS-SVM method exhibits high modeling accuracy and robust noise resistance, offering valuable support for ship motion prediction.

With the development of intelligent shipping and autonomous ship maneuvering technologies, higher requirements were imposed on the parameter identification accuracy of ship motion models in practical navigation environments. However, under limited sample conditions, traditional parameter identification methods for nonlinear ship motion models were difficult to achieve both accuracy and stability, which restricted their engineering applications. To address this problem, a ship maneuvering parameter identification method based on Physics-Informed Neural Networks (PINN) was proposed. During network training, a second-order nonlinear ship maneuvering response model was deeply introduced as a physical prior constraint, and the dynamic information of the model was deeply integrated with limited observation data to improve the effectiveness of parameter learning under limited sample conditions. Numerical simulations of 20°/20° zigzag maneuvering tests were carried out based on the Mariner model ship, and training and validation datasets were constructed to identify ship maneuverability indices. The identification results were compared with those obtained by the Extended Kalman Filter (EKF) method and the Least-Squares Support Vector Machine (LS-SVM) method. The results show that, under limited sample conditions, the proposed method improves the identification accuracy of nonlinear ship maneuvering model parameters by 14.75% and 7.74% compared with the EKF method and the LS-SVM method, respectively. In robustness tests with Gaussian white noise contamination, stable parameter identification performance is maintained by the PINN-based method, and the minimum identification error reaches 0.05%. The results indicate that the proposed method effectively solves the problem of high-precision parameter identification for nonlinear ship maneuvering models under limited data conditions and provides a feasible technical approach for ship motion modeling and maneuvering performance analysis.

Reliable maritime situational awareness is a key requirement for the development of intelligent ships. Radar and the Automatic Identification System (AIS) are among the most widely deployed navigation sensors on board ships, and the association and fusion of their data provide the foundation for intelligent perception. However, ship state measurement data are often characterized by asynchronous sampling, missing observations, and uncertainty, which make track association challenging. To address this problem, this paper proposes a multi-target track association algorithm based on Dynamic Time Warping (DTW) and Dempster-Shafer Theory (DST). First, multi-source track data are temporally aligned and represented using the DTW method. Then, multidimensional track features are fused based on DST. A two-stage association decision strategy is further constructed to obtain the globally optimal matching of heterogeneous sensor data under constrained conditions. A multi-target track simulation dataset is established according to the observation characteristics of radar and AIS targets, and the applicability of the proposed algorithm is evaluated under different scenarios. The algorithm is further validated using sea-trial data collected from the intelligent research and training vessel Xin Hong Zhuan of Dalian Maritime University. The results show that the F₁-score of the proposed algorithm exceeds 0.9 in the track association task, while the performance fluctuation across different scenarios is only 9.03%. These results demonstrate that the proposed method has high accuracy and robustness, providing technical support for multi-source perception data fusion and maritime situational awareness in intelligent ships.

In the busy waters of nearshore port channels, physical Aids to Navigation (AtoN) are highly susceptible to collisions, which may cause structural damage and prevent them from operating normally. To address the research gap in autonomous navigation of AtoN, including the inability of existing methods to achieve autonomous navigation and collision avoidance and their insufficient integration with artificial intelligence technologies for intelligent navigation, this paper proposes an autonomous navigation method for AtoN based on improved Proximal Policy Optimization (PPO) combined with the Velocity Obstacle (VO) algorithm. The PPO algorithm was improved by introducing an offline learning strategy to enhance network update efficiency. By designing the autonomous navigation action space for AtoN, the model can achieve faster convergence. However, the model is prone to falling into local optima during convergence. To avoid this problem, the VO algorithm is incorporated to represent the state space, and a reward function is designed using velocity vectors. On the basis of calculating the distance-based reward function, an additional reward function based on position and velocity information is introduced to improve convergence speed and enable AtoN to adaptively select the optimal velocity in nearshore port channels. To further enhance algorithm robustness, an entropy reward term is added to the PPO algorithm to encourage greater exploration. Experimental results show that, compared with the method without the entropy reward term, the proposed method improves the average navigation step length and average navigation velocity by 9.3% and 10.2%, respectively, while maintaining a model test success rate of over 99%. The proposed method enables a safe operational mode featuring unmanned control, autonomous navigation, and collision avoidance for AtoN. It can reduce the collision damage risk of AtoN in busy waters and lower the costs of manual inspection and maintenance. The study is of significance for accelerating the construction of a modern maritime navigation support system and ensuring maritime navigation safety.

To address the challenge of coordinating and scheduling construction vessels in large-scale water projects such as waterway regulation, this problem is modeled for the first time as a Dual-Resource Berth Allocation Problem (DR-BAP). A mathematical model is established with the objective of minimizing vessel waiting costs and resource preference mismatch costs. To solve this Non-deterministic Polynomial-time hard (NP-hard) problem, a Two-Stage Hybrid Cuckoo Search Algorithm (TSH-CSA) is proposed. In the first stage, the algorithm employs dynamic priorities, a load balancing guidance strategy, and a historical congestion feedback penalty mechanism for macroscopic resource allocation. It integrates domain knowledge and feedback information to optimize fitness evaluation. In the second stage, the algorithm integrates an adaptive dual-population mechanism and an intelligent restart mechanism to achieve a microscopic, refined solution. Overall, the algorithm establishes a global iterative framework of "allocation-scheduling-feedback" to enhance its global optimization capability. Simulation experiments based on a waterway regulation project show that, compared with the standard Cuckoo Search Algorithm, TSH-CSA exhibits significant improvements in solution quality, convergence speed, and robustness. These results verify the effectiveness and advancement of the proposed model and algorithm in solving complex engineering scheduling problems.

With the continuous expansion of global trade and shipping, container throughput at terminals increases exponentially. Higher requirements are posed on port logistics efficiency, particularly the collaborative operation efficiency of automated guided vehicles (AGVs). Existing centralized algorithms and distributed algorithms exhibit limitations in balancing computational complexity and optimization quality. The demands of automated container terminals for shortening makespan and improving efficiency cannot be effectively accommodated by existing algorithms. To address these problems, a Hybrid task Allocation Algorithm (HAA) combining centralized algorithms and distributed bid was proposed. First, a multi-objective optimization model was constructed from the perspectives of the scheduling system, AGV fleet, and task set to minimize makespan, total AGV travel distance, and task execution delay. Second, a genetic algorithm was utilized to solve this problem. The task set was divided into ordered task subsets equal to the number of AGVs. Finally, bid values for the task subsets were distributively computed by each AGV. The cost matrix constructed from the bid values was solved using the Hungarian algorithm. Through simulation experiments, the proposed HAA method was validated under scenarios with various task scales and task densities. Comparative analyses with traditional centralized algorithm and distributed algorithms were performed. The results demonstrate that a solution quality close to that of centralized algorithms is achieved by the proposed method. Meanwhile, the computation time is reduced by 39.09% on average. Good and stable optimization performance is exhibited. Furthermore, the method can be effectively applied to periodic scheduling scenarios.

As the lifeblood of global trade, the international shipping industry accounts for approximately 3% of global carbon emissions, posing a significant challenge to the achievement of climate targets. Onboard carbon capture and storage (OCCS) technology is widely regarded as an important pathway for the shipping industry to move toward net-zero emissions because of its substantial emission-reduction potential and compatibility with existing vessels. This study focuses on the economic feasibility of OCCS technology and systematically analyzes its cost structure. The results reveal that initial capital expenditure and the additional fuel consumption caused by the capture system are the major economic barriers limiting its large-scale adoption. A quantitative case study of a medium-range (MR) tanker shows that, under specific operating and market conditions, the investment payback period of the OCCS system is 8.38 years, preliminarily demonstrating the economic potential of this technology for commercial application. Furthermore, the sensitivity analysis indicates that the carbon tax rate and the CO₂ market price are the decisive variables affecting the payback period, as well as key uncertainties that should be emphasized in future techno-economic assessments. Based on these findings, this study suggests that the commercialization of OCCS cannot rely solely on technological maturity, but requires a coordinated support framework integrating policy guidance, technological innovation, and market mechanisms. By providing empirical data, this study offers an analytical basis for translating the shipping industry's macro-level decarbonization strategy into micro-level economic decision-making and helps fill the gap in quantitative assessments of OCCS applications for specific ship types.

With the continuous enhancement of global shipping's requirements for energy conservation and emission reduction, accurate prediction of ship fuel consumption has become crucial for optimizing shipping efficiency and mitigating environmental impacts. However, the issues of feature redundancy and noise interference in multi-source data restrict the accuracy and stability of fuel consumption prediction models. To address these problems, a comprehensive framework for precise fuel consumption prediction is proposed. First, ship noon report data, ERA5 reanalysis data, global ocean physical analysis and prediction data, and Automatic Identification System (AIS) data are integrated, and a multi-dimensional feature space dataset is constructed. Second, to address feature redundancy and noise interference, a genetic algorithm is introduced for feature selection, and the original 32 features are reduced to 15. Third, nine machine learning regression models are systematically evaluated, and the results indicate that the Random Forest model exhibits significant advantages in capturing the nonlinear characteristics of fuel consumption. Finally, feature importance analysis based on the RF model reveals that time-related variables (such as daily operating hours and total operating hours) play a dominant role in fuel consumption prediction, providing quantitative decision support for energy efficiency management. The results show that the genetic algorithm-based feature selection effectively reduces computational complexity by approximately 53.1%, and the Random Forest model performs best in capturing the nonlinear characteristics of fuel consumption, significantly improving prediction accuracy. The research findings provide theoretical support for fuel optimization and emission reduction strategies in the shipping industry, and are of great significance in promoting the green transformation of maritime transport.

With the green transformation of port equipment, tugboat scheduling faces new opportunities and challenges. On the one hand, the deployment of new energy tugboats helps reduce port operating costs and carbon emissions; on the other hand, research on new energy tugboats remains relatively scarce. To address this research gap, a mixed-integer linear programming model for heterogeneous tugboat scheduling under the context of multiple port areas is established based on the characteristics of hybrid tugboats, aiming to minimize port costs, tugboat operation costs, and carbon emission costs. To solve the model, a Multi-Verse Optimization algorithm embedded with heuristic rules is applied, and its effectiveness and accuracy are verified through comparison with the Gurobi solver. In large-scale instances, the proposed algorithm achieved objective values that were, on average, 4% and 3.25% better than those obtained by the GA and IGA, respectively. The research results show that configuring high-power tugboats can effectively reduce ship port costs and tugboat operation costs. However, when the number of high-power tugboats is too large, it may increase operation costs, and at this point, appropriately configuring lower-power tugboats can leverage their low operation cost advantage. An appropriate configuration of a certain number of hybrid tugboats can effectively reduce the total scheduling cost as well as the carbon emission cost of tugboat operations. Meanwhile, the proportion of carbon emission costs in the overall operational cost can be continuously reduced from 8% to approximately 1.51%.