Stringent engine emission standards are a widely adopted regulatory approach to mitigating pollution. However, the statutory ship engine tests often fail to adequately reflect real-world operational conditions. This paper introduces a method for developing representative ship engine tests. A Similarity-based Navigational Cycle Construction (SNCC) method is proposed to generate representative engine tests from actual operational data, including engine speed and power output per second during the tests. The method is applied to an inland ship operating continuously over an eight-month period. Based on real-ship measurements, real-world emission levels are calculated and compared with those obtained under existing statutory engine test conditions. The results demonstrate that the navigational cycle constructed using the proposed method exhibits a substantially closer correspondence to the ship's real operational profile, thereby yielding emission evaluation results that more accurately reflect actual conditions. As a result, the discrepancy in the CO2 emission factor between real-world emissions and the statutory engine emission test is reduced from 38.6% to 4.1%. This method can provide scientific support for the development of next-generation regulatory test cycles for ship engines, offering a pathway to update certification benchmarks and enhance the effectiveness of maritime emission reduction policies.

The complex ocean disturbances in ocean engineering have long constrained the precise autonomous navigation of intelligent marine vehicles, such as surface vessels and underwater vehicles. Nevertheless, the unpredictable wind-wave-current coupling effects pose severe challenges to the safety of autonomous navigation for marine vehicles. Here we introduce a domain knowledge embedded disturbance observation-control framework, fusing real-time observation and compensation for composite environmental disturbances using model-free control. This framework embeds specialized basis functions from domain knowledge into a specialized Kolmogorov-Arnold network and extracts control knowledge therefrom to train a machine learning controller. Our approach achieves better adaptability and robustness, surpassing conventional model-based controllers. It enables more accurate path-following and safer operations under complex ocean disturbances. It is worth noting that this method has been validated to be effective for both surface vessels and underwater vehicles through offshore wind farms inspection task scenarios. It fundamentally extends the adaptive control theory for marine cyber-physical systems and has potential applications in multi-domain oceanographic operations.

Existing deep reinforcement learning (DRL) methods for autonomous underwater vehicle (AUV) path planning face two practical challenges: 1) dependency on manual reward engineering and 2) hyperparameter sensitivity in dynamic marine environments. This article presents a novel AUV path planning framework incorporating generative adversarial imitation learning (GAIL) and DRL algorithm that automates reward function synthesis through adversarial learning from expert demonstrations. The proposed architecture introduces a hierarchical reward mechanism that concurrently optimizes global trajectory planning and local motion constraints. By eliminating manual reward engineering, our approach reduces training complexity while maintaining policy convergence stability. Extensive experimentation demonstrates superior performance with 93.7% faster training convergence and 72.7% higher path convergence optimality compared to conventional DRL baselines. Two-tier validation confirms operational effectiveness: 1) Gazebo simulations achieve maximum 100% success rate in dynamic scenarios and 2) field deployments for submarine pipeline inspection attain 1 m average tracking accuracy. The results demonstrate that GAIL-DRL trained AUVs exhibit enhanced path planning stability while satisfying real-time planning requirements for marine transportation systems.

From the perspective of fishermen surveillance, this study investigates fishermen's behaviour and operational patterns to enhance workplace safety during fishing activities. This paper introduces an early-fusion, multimodal approach to monitoring individual fishermen's behaviour using wrist-worn inertial sensors. To characterise behaviour during fishing operations, accelerometer and gyroscope signals were collected onboard from 17 subjects across four behaviour classes. The proposed model performs channel-level early fusion to jointly process the heterogeneous six-channel time series, enhancing the extraction of complementary and discriminative representations from multi-sensor inputs. An experimental analysis was conducted, including a hyperparameter study, to assess the effects of class imbalance on model performance, compare the proposed approach with baseline and state-of-the-art models, and examine how different sensor configurations influence attention distribution within convolutional layers. Class imbalance affects model performance, with balanced training yielding 2.25 % higher accuracy than imbalanced training. A single accelerometer or gyroscope produces 8.43 % and 28.57 % lower accuracy than the fused accelerometer-gyroscope setup. The proposed model achieved the highest accuracy of 90.11 % with an inference time of 0.14 ms per sample compared with baseline and state-ofthe-art models. Integration of wearable devices with human activity recognition algorithms is highly applicable to maritime scenarios and has strong potential to accurately distinguish fishermen's behaviours, thereby providing a practical tool for monitoring human-related risk factors in maritime transportation.

Multi-bridge inland waterways constitute some of the most constrained and high-risk navigation environments. The dense spacing of bridges and the resulting fragmentation of navigable channels create complex spatiotemporal variability that existing prediction models struggle to capture, which limits their applicability in such areas. This study proposes a spatiotemporal variational model, Spatio-Temporal Variational Autoencoder (ST-VAE), for multi-step vessel traffic and density prediction in multi-bridge inland waterways. The framework integrates an embedded graph-enhanced Long Short-Term Memory (LSTM) encoder, which exploits bridge-to-bridge connectivity, with a variational latent space and a distributed sampling network to represent uncertainty in traffic evolution. Using Automatic Identification System (AIS) data from 13 consecutive bridge areas in Wuhan, two datasets are constructed for high-flow (V-high) and low-flow (V-low) seasons, and 12 similar to 48-h prediction tasks are evaluated against ten baseline models. On the V-high dataset with a 24-h horizon, ST-VAE achieves a mean absolute error (MAE) of 2.57 and a mean absolute percentage error (MAPE) of 3.83%, which are 24% and 61% lower than the best baseline, respectively, and reaches a Pearson correlation coefficient (PCC) of 98.0%. For 48-h forecasts, MAE and MAPE are reduced by 34% and 50% compared with the strongest competitor. One-week rolling experiments further show that ST-VAE maintains a PCC above 90% while achieving approximately 38% lower MAE than the best ablation model. Finally, K-means clustering is applied to convert predicted flows into three density levels, revealing persistent bottlenecks near key bridges and providing actionable support for congestion monitoring and traffic management in bridge-dense inland waterways.

With the rapid development of inland waterway transport, vessel traffic flow becomes increasingly busy. Ship-induced waves result in significant hydrodynamic effects on inland waterway infrastructures and ship navigation safety. The characteristics of ship-induced waves could not only provide a reference for ship hull form optimization but also support the scientific maintenance of waterway infrastructure. In the context of extensive field ship wave observations under complex circumstances, an optimized estimation formula has been proposed considering the influence of ship dimensions. Different dimensionless variables were introduced, and the estimation formula was optimized through the pi theorem. Its coefficients were determined by the least square fitting method. A detailed comparison with the classic estimation formulae demonstrates that the proposed formula would provide accurate estimation of ship waves under various conditions in inland waterways. The root mean square error is reduced by approximately 49% in contrast with the traditional estimation formulae. Additionally, a high coefficient of determination was derived as 0.88, highlighting its versatility in maximum ship wave height estimation of inland waterways. The present study shares valuable insights into ship wave height estimation which would be beneficial for inland waterways planning and design.

With the growing emphasis on maritime safety, human behavioral errors on fishing vessels have become a leading cause of navigation accidents. However, conventional monitoring systems focus primarily on vessel-level indicators, overlooking crew behaviors and cognitive states. This study introduces a real-time behavior prediction framework that combines wearable multimodal sensing with an online deep learning architecture to improve safety supervision in dynamic maritime operations. A twenty-one day field experiment on working trawlers employed wrist-worn sensors that record acceleration, angular velocity and photoplethysmography, generating labeled data for seven categories of operational behaviors performed by six fishermen. The proposed online Transformer model incorporates an adaptive attention mechanism to represent multi-scale temporal patterns and uses a hedged parameter update strategy that maintains stability under shifts in data distribution. The method attains a mean F1 score of 93.5 percent for predicting behaviors 30 s in advance and achieves prediction latency below 1 s. Additional evaluations involving cross-individual testing and controlled noise perturbation with a standard deviation up to 0.05 verify its robustness. The results demonstrate that integrating wearable sensing with deep online learning can support timely identification of unsafe crew states and offers a practical basis for human-centered risk warning and maritime safety supervision.

As a transportation hub for the import and export transshipment of bulk cargo, the manual operation mode of bulk terminals can no longer meet the developmental demand for efficient unloading of bulk carriers. Therefore, this article designs a whole-process intelligent unloading (WPIU) system for bulk carriers, including a series of intelligent unloading equipment and a robot-shore cooperative perception system. Second, a novel unloading strategy of bulk carriers that incorporates a sliding window detection algorithm rooted in 3-D feature distribution enables the creation of a real-time cargo hold and material model, facilitating the prediction of safer and more efficient grasping areas according to the materials' dynamic distribution characteristics. Subsequently, by utilizing the structural features of the cargo hold and multisensor information, a robot-shore cooperative localization method integrating adaptive dual observation is proposed to obtain the relative position of the clearing robot and the grabber, achieving collision avoidance during the unloading operation. Finally, the WPIU system and its integrated intelligent algorithms were experimentally validated at two typical bulk ports handling grain and ore. The results demonstrate the system's capability for intelligent and efficient unloading operations of bulk carriers.

Inland water vessels are a substantial but poorly quantified source of transport-sector CO2 emissions. Using approximately 20 billion automatic identification system records and a trajectory completion algorithm, we provide a global high-resolution assessment of inland vessel CO2 emissions. These vessels emitted 164.0 ± 1.58 TgCO2 in 2022, equivalent to 24.7% of global waterborne vessel emissions, despite operating on only 0.4% of the global navigable marine and inland water surface area. Emissions are concentrated within a few hotspot river corridors and exhibit baseline-dependent amplification (2019–2022) together with seasonal variability. These patterns reflect the combined effects of transport demand and hydroclimatic constraints on navigability. Because inland water vessel emissions occur within river corridors where riverine CO2 evasion is also active, they represent an important anthropogenic carbon source and may influence river–atmosphere CO2 exchange. These results provide a basis for improving inland transport emission inventories and informing targeted mitigation strategies under climate change. The magnitude and distribution of CO2 emissions from river transport are poorly constrained. This study finds that global CO2 emissions from inland vessels were about one-quarter of total vessel emissions, and that emissions were concentrated in the Northern Hemisphere, related to economic activity.

Enhancing navigational safety of hazardous cargo vessels constitutes a critical imperative for sustaining maritime transportation system stability and fostering sustainable industry development. Based on the developed database containing 106 accident reports involving hazardous cargo vessels collected from the International Maritime Organization (IMO), this study aims to analyze the key risk influencing factors (RIFs) contributing to the maritime traffic accidents. Utilizing text analysis, the research first identifies critical RIFs across five primary domains, which are human, vessel, cargo, environment, and management. A Bayesian network model is subsequently developed to map out the interrelationships among these identified navigational safety RIFs. The findings suggest that factors such as "insufficient personnel training," "inadequate safety inspections," "flammable and explosive cargo," "inadequate hazardous goods management," and "pollutant and toxic cargo" exert the most pronounced influence on maritime traffic accidents. Based on these pivotal RIFs and their evolutionary trajectories, this paper can offer theoretical support for enhancing the navigational safety of hazardous cargo vessels.

Intelligent navigation decision support systems are crucial for maritime safety, yet these systems frequently exhibit limited adaptability and reliability in novel, non-predefined scenarios, constituting a persistent challenge. This study proposes Navigation-GPT, a dual-core large language model (LLM) agent designed for intelligent marine navigation. The framework leverages the strong generalization capability of LLMs in unfamiliar situations. It employs a large-scale LLM with ReAct prompting as its control core, responsible for task parsing, planning, and orchestrating external tools to mitigate hallucinations. Furthermore, we fine-tune a lightweight LLM in two stages: using LoRA and a novel rule-controlled GRPO (RC-GRPO) method to develop a specialized agent decision core. This core generates COLREGs-compliant high-level collision avoidance decisions, which are translated into dynamically feasible reference trajectories using a ship dynamics model formulated according to Fossen's equations. A PID-based controller then tracks these trajectories to guide the ship through the resulting avoidance maneuver. Experimental results show that Navigation-GPT completes the process from task reception to decision output in 11.13 s, remaining within the critical safety window for collision avoidance, though longer than the 0.73 s of traditional methods. In complex scenarios, it achieves an 86% collision avoidance success rate and a 90% behavioral compliance rate, outperforming its base model Qwen2.5-7B by 38% and surpassing benchmarks including the dynamic window approach, artificial potential field, and other LLMs (Qwen2.5-0.5B, Qwen2.5-14B, DeepSeek, GPT-4o). This work integrates LLM technology with traditional navigation systems, offering a comprehensive solution that enhances both safety and operational efficiency across diverse maritime scenarios.

[Objective]Intelligent ship navigation has become a core technological enabler for accelerating the digital,intelligent and low-carbon transformation of the global shipping industry.Driven by the Interna-tional Maritime Organization's greenhouse gas emission reduction targets and China's strategic policies for in-telligent shipping development,ship intelligent navigation systems are evolving from experimental validation toward engineering applications.However,several key challenges remain in practical implementation.In par-ticular,the coordination mechanism among navigation decision-making,propulsion response,and energy-efficiency constraints lacks a unified architectural representation.In addition,the adaptive mapping logic be-tween shipboard intelligent capability levels and shore-based hierarchical operation modes has not yet been systematically established.To address these gaps,this study focuses on the architectural design of a bridge-engine integrated ship intelligent navigation system and investigates the adaptive mechanisms of ship-shore collaborative operation.[Method]In terms of research methods,this paper firstly defines the functional connotation,system boundaries,and essential characteristics of bridge-engine integration based on existing regulations for intelligent ships and relevant domestic and international research outcomes.Secondly,a hierar-chical intelligent navigation architecture is established.A three-stage evolution pathway for shipboard autono-my is proposed,including enhanced navigation,assisted navigation and autonomous navigation.In parallel,three shore-based operation modes are defined,namely monitoring,remote control,and supervisory naviga-tion.On this basis,an asymmetric ship-shore functional adaptation matrix is developed to systematically clari-fy recommended combinations,restricted feasible combinations,and inapplicable combinations,with typical application scenarios clearly specified.Finally,key enabling technologies are comprehensively analyzed,in-cluding environmental perception and situational awareness,navigation decision-making and path planning control,intelligent engine room operation and energy efficiency management,system testing and evaluation,as well as ship-shore human-machine collaborative control.[Results]The research results systematically present the overall hierarchical architecture of a bridge-engine integrated ship intelligent navigation system under ship-shore collaboration.They clarify the functional connotation and operational boundaries of each shipboard capability level and shore-based operation mode,and reveal the asymmetric adaptation rules and principles of control authority allocation between shipboard and shore-end systems.[Conclusion]It is con-cluded that the proposed system architecture and asymmetric adaptation matrix can provide a solid theoretical foundation and technical reference for the engineering implementation of intelligent ship navigation systems and the development of relevant industrial standards.Furthermore,the findings offer important theoretical guidance for improving ship-shore collaborative operation mechanisms,standardizing control authority switching,and promoting the development and application of a new-generation intelligent and green maritime transportation system.

Effective control of the health operating condition of multi-support, ultra-long shaft system water-lubricated stern bearings is crucial for supporting the intelligent maintenance and health management of ships. This study investigates the failure modes of water-lubricated stern bearings and focuses on the critical failure modes of abnormal wear and high-temperature meltdown to analyze the mechanisms and influencing factors of these failures. It discusses the conditions for healthy operation of water-lubricated stern bearings, as well as methods for controlling lubrication and temperature rise. Based on this, controllable parameters for the healthy operation of water-lubricated stern bearings were selected, an experimental rig was constructed, and experiments were conducted using SF-2A material water-lubricated bearings. The experimental results indicate that by controlling parameters such as shaft rotational speed, inlet lubrication water temperature, clear-water lubrication, sediment-laden-water lubrication, bearing specific pressure, and the surface morphology of the bearing liner, the velocity characteristics, lubrication characteristics, and temperature rise characteristics of the bearings can be effectively altered. The sensitivity of the lubrication and temperature rise characteristics of SF-2A material water-lubricated stern bearings to controllable parameters varies under different environmental conditions. The study finds that precise control of these parameters can improve the operating condition and reliability of water-lubricated bearings.

The environmentally friendly water-lubricated bearing is one of the most important constituents of the marine propulsion system. However, the water-lubricated bearings still face the challenge of poor tribological properties under harsh working conditions such as dry frictional and heavy load, resulting in unreliable working performance. In this case, the Lignum vitae wood biomaterial, which is the earliest water-lubricated bearing material due to its excellent self-lubricating properties, is used as the bionic object. Three different kinds of bionic fibers, including thermoplastic polyurethanes (TPU) fibers, silicone oil/TPU core-shell fibers, and polydimethylsiloxane (PDMS)/TPU core-shell fibers, are designed, and the bionic materials are manufactured further. The Lignum vitae wood sample (LVS) is used as a contrast for evaluation. The tribological experimental results indicate that the bionic material samples present far better tribological properties than the LVS under dry frictional and heavy-load water-lubricated working conditions. Among all the samples, the PDMS/TPU fibrous membranes bionic material samples (PTM) under parallel to fibers sliding direction present the best tribological properties due to their unique soft-hard gradient structure. Specifically, the PTM achieves the maximum reduction of 73.0% and 98.7% in friction coefficient and wear volume compared with the LVS. The knowledge gained in this study provides a novel route to improve the tribological properties of water-lubricated bearing materials.

Autonomous Transportation Systems (ATS) represents a transformative advancement in modern transportation, surpassing traditional transportation systems (TTS), autonomous driving technologies (ADT), and intelligent transportation systems (ITS). Unlike vehicle-centric autonomy or information-driven support systems, ATS emphasises system-level autonomy, multi-vehicle collaboration, and dynamic optimisation across road, maritime, air and rail transport. This paper provides a comprehensive review of the current state and future trends of ATS. It first examines the theoretical foundations and the evolution of ATS concept, including system engineering methodologies, the closed-loop architecture of perception-awareness-decision-control-collaboration-improvement, and the integration of model-driven and data-driven approaches. It then defines the concept, components, and characteristics of ATS, systematically analysing autonomous vehicle, digital infrastructure, information platform, and traffic environment, as well as the core features required to achieve autonomy. Key technologies are reviewed, covering information perception, interoperability, collaborative computing, and digital twin. The standards and applications are discussed across different modes of transportation in facilitating industrial deployment. Finally, future research trends are outlined, with emphasis on theoretical integration, cross-modal collaboration, verifiable safety, and application scenarios. Overall, ATS is regarded as the ultimate form of ITS, offering significant potential in enhancing safety, efficiency, and sustainability, and promoting a paradigm shift towards the next generation of transportation systems.

Despite advancements in science and technology, ship collisions and groundings remain the most prevalent types of maritime accidents. Recent developments in accident prevention and mitigation methods have been bolstered by the rise of autonomous shipping, digital technologies, and Artificial Intelligence (AI). This paper provides an exhaustive review of the characteristics of fleets at risk over the past two decades, emphasizing the societal impacts of preventing collisions and groundings. It also delves into the key components of decision support systems from a ship's perspective and undertakes a systematic literature review on the foundations and applications of systems-driven decision support methods for ship collision and grounding prevention. The study covers risk analysis, damage evaluation, and ship motion prediction methods from 2002 to 2023. The conclusions indicate that modern ship science methods are increasingly valuable in ship design and maritime operations. Emerging multi-physics systems and AI-enabled predictive analytics show potential for future integration into intelligent decision support systems. The strategic research challenges include (1) underestimating the impacts of real operational conditions on ship safety, (2) the inherent limitations of static risk analysis and finite numerical methods, and (3) the need for rapid, probabilistic assessments of damage extents. The demands and trends suggest that leveraging big data analytics and rapid prediction methods, underpinned by digitalization and AI technologies, represents the most feasible way forward.

Onboard Carbon Capture System (OCCS) technology is one of the most promising solutions for reducing greenhouse gas emissions from maritime transportation. To address the challenges of low carbon capture efficiency and high energy consumption associated with OCCS systems, a dedicated experimental platform was developed. In this study, monoethanolamine (MEA) was used as the carbon capture solvent to investigate the effects of spray structures in the absorber tower, liquid-to-gas ratios (L/G), and gas flow rates on system performance. It was found that atomized spray nozzles could extend the gas-liquid contact time, thereby improving CO2 absorption efficiency. Experiments conducted at a gas flow rate of 17 m3/h under different L/G showed that increasing the L/G could enhance carbon capture efficiency by approximately 4%. Further experiments at a reduced gas flow rate of 13 m3/h indicated that lowering the gas flow rate significantly improved the carbon capture rate, achieving an optimal carbon capture efficiency of 61% at an L/G ratio of 3:1. Moreover, absorption efficiency exhibited distinct trends under varying L/G and gas flow rates, highlighting the importance of rich solution loading in determining capture efficiency.