Smart and Low Carbon Emission-Oriented Maritime Traffic Management and Controlling

As the global shipping industry faces mounting pressure to reduce carbon emissions, AI-powered intelligent shipping systems are playing a pivotal role in optimizing vessel routes, fuel consumption, and port operations. AI-driven systems process real-time data from IoT sensors, weather forecasts, and Automatic Identification System (AIS) signals to recommend optimal fuel-efficient sailing routes, while circumventing congested waterways and adverse weather conditions. Machine learning algorithms bolster predictive maintenance capabilities, mitigating excessive emissions caused by engine inefficiency. Furthermore, AI-guided autonomous and remotely controlled vessels enable precise maneuvering, curtailing port idle time and ensuring just-in-time arrivals, which significantly reduces unnecessary fuel consumption [1]. Ports equipped with intelligent traffic management systems can dynamically allocate berths and optimize cargo handling processes, yielding further energy savings. AI-driven predictive maintenance helps to maintain engine efficiency, thereby further reducing carbon emissions [2]. The rise in autonomous and remotely controlled vessels also holds promise for more precise control, enabling vessels to achieve just-in-time arrivals, reduce idle time at ports, and minimize fuel waste. By leveraging AI-driven decision support systems, the shipping industry can transition to a low-carbon, high-efficiency future aligned with global sustainability goals, ensuring that shipping management systems become smarter and greener.

AI-driven intelligent shipping systems play a crucial role in optimizing vessel routes, fuel consumption, and port operations. Machine learning algorithms not only enhance predictive maintenance capabilities, reducing the excessive emissions caused by engine inefficiencies, but can also integrate circular economy principles to optimize resource flows through digital monitoring and reverse logistics [3]. The convergence of blockchain and AI technologies further strengthens the traceability and decision-making efficiency of maritime supply chains, ensuring that route optimization and emission monitoring comply with sustainability standards [4]. For example, dynamic path planning algorithms can efficiently generate low-carbon optimal routes by controlling the search space scale and restricting search directions, complementing the precise control of AI-guided autonomous and remotely operated vessels to reduce port idle time and ensure punctual arrivals, thereby minimizing fuel waste [5].

Additionally, ports equipped with intelligent traffic management systems can adopt centralized logistics and spatial planning strategies to dynamically allocate berths and optimize cargo handling processes, achieving more efficient energy utilization [6,7]. By integrating AI-driven decision support systems with predictive maintenance technologies, the shipping industry can transition toward a low-carbon, high-efficiency future aligned with global sustainable development goals, ensuring smarter and greener shipping management systems. The comprehensive application of this interdisciplinary approach not only improves maritime operational efficiency but also provides practical solutions for achieving environmentally friendly maritime traffic management.

Recent advances in maritime technology demonstrate significant progress across five interconnected domains: (1) vessel systems optimization, (2) intelligent navigation, (3) emission control, (4) port operations, and (5) safety management. The research trajectory begins with Zhang et al. (Contribution 1) addressing bandwidth limitations in BeiDou satellite communications through an innovative attention-averaging compression algorithm. This technical solution employs a three-stage processing pipeline (quantization → attention-averaging-based predictive coding → arithmetic coding) to achieve the efficient real-time transmission of ship status data while maintaining data integrity, providing a cost-effective monitoring solution for ocean-going vessels. Yang et al. (Contribution 2) enhanced propulsion system stability through neural network control, proposing a second-order sliding-mode observer integrating the Super-Twisting Algorithm (STA) to address speed estimation errors and torque vibration in permanent magnet synchronous motors (PMSMs). By dynamically adjusting control parameters via fuzzy neural networks and implementing online compensation through Iterative Learning Control (ILC), they significantly improved operational stability and the energy efficiency of propulsion systems. These foundational achievements enabled Zhao et al. (Contribution 3) to develop a multi-objective route optimization model. Their enhanced Non-dominated Sorting Genetic Algorithm II (NSGA-II) incorporates ship stability constraints and complex sea condition interference factors, establishing a dual-objective optimization model (energy consumption vs. time) that achieves an optimal balance between safety and economy in transoceanic route planning. Complementing these developments, Shao et al. (Contribution 4) and Chen et al. (Contribution 5) strengthened maritime safety through rescue vessel deployment optimization and VHF radio identification technology, respectively. Shao’s team designed a multi-level coverage strategy to significantly improve emergency response efficiency, while Chen’s group developed a hybrid deep learning model (CNN-BiLSTM) that achieves ship radio identification in non-cooperative conditions through CAM feature extraction and bidirectional temporal feature fusion, pioneering new approaches for maritime traffic awareness.

Building upon these achievements, Bao et al. (Contribution 6) developed an emission reduction decision prediction model, pioneering the application of the XGBoost algorithm to analyze shipowners’ preferences between Very Low-Sulfur Fuel Oil (VLSFO) and scrubbers. Their predictive model incorporates key features such as deadweight tonnage and annual sailing distance, establishing a data-driven reference framework for post-pandemic emission reduction decisions. This work has technical complementarity with Huang et al.’s (Contribution 7) innovative drone monitoring system, which employs a genetic algorithm based on a Sequential Insertion Heuristic (SIH). The proposed DroneByDrone strategy significantly enhances emission detection efficiency through optimized UAV speed parameters, creating a new paradigm for mobile maritime surveillance. Navigation capabilities were strengthened through Huang et al.’s (Contribution 8) collision avoidance algorithm and Li et al.’s (Contribution 9) high-precision ship segmentation technology. By incorporating the COLREGs rules constraints and Collision Risk Index (CRI) evaluation, their approach generates kinematically feasible safe paths for both static obstacles and dynamic encounter scenarios. The CA²HRNET model innovatively integrates channel attention, multi-scale spatial attention, and weight self-adjustment mechanisms, maintaining a precise segmentation capability for small target vessels even under strong noise interference through cross-scale attention fusion in feature pyramid networks. Liu et al. (Contribution 10) improved detection efficiency through cloud-edge collaboration, migrating computationally intensive tasks like semantic segmentation to the cloud while retaining only mask-based detection at the front end. Their optimization of coordinate attention modules and confluence algorithms significantly boosts detection speed without compromising accuracy. Wang et al. (Contribution 11) systematically mapped the technological landscape, conducting an in-depth analysis of key technical bottlenecks including SLAM and 3D reconstruction. They first proposed the “perception-decision-control” technical roadmap for visual navigation systems. Wei et al. (Contribution 12) achieved automation of critical draft measurement through innovative dynamically adjusted measurement tube positioning, addressing the pain points of traditional visual observation being susceptible to human factors.

Energy optimization emerged as another crucial pillar. Zhou et al. (Contribution 13) realized accurate fuel prediction by pioneering an interval prediction framework combining the Gaussian Process and Quantile Regression. By capturing fuel consumption volatility characteristics, their model generates probability-based prediction intervals, providing uncertainty quantification support for ship energy efficiency management and significantly improving bunkering planning reliability. Lu et al. (Contribution 14) ensured the robust management of hybrid power systems through a Distributionally Robust Optimization (DRO)-based bi-level scheduling model addressing dual uncertainties of wind–wave and PV conditions. By employing linear decision rules and column-and-constraint generation algorithms, their approach achieves the coordinated optimization of operational costs and carbon emissions while guaranteeing on-time arrival. Xu et al. (Contribution 15) provided critical emission insights through a bottom-up emission inventory model based on AIS trajectories and meteorological data. Zhen et al. (Contribution 16) optimized port traffic flow through intelligent scheduling, solving the NP-hard problem of one-way channel scheduling by designing state–action space encoding rules. Their real-time learning of vessel arrival dynamics and channel occupancy status enables the intelligent optimization of inbound/outbound sequences, effectively alleviating port congestion with improved computational efficiency compared to traditional algorithms. Xu et al.’s (Contribution 17) breakthroughs in materials science further enhanced vessel reliability through comparative studies of vibration stress relief and tempering methods for marine-grade steel components.

The refinement of the autonomous navigation ecosystem owes much to Guo et al.’s (Contribution 18) adaptive control system, which designed an ANFIS (Adaptive Neuro-Fuzzy Inference System) fuzzy neural network controller based on the MMG (Maneuvering Model Group) maneuvering model. By employing a backpropagation–least squares hybrid training strategy, the system achieved reduced response times under normal sea conditions and maintained rudder effectiveness stability in harsh sea conditions, thereby addressing the challenge of rigid parameter settings in traditional PID (Proportional–Integral–Derivative) control. Li et al.’s (Contribution 19) hydrogen safety framework integrated the Bow-tie model with fuzzy set theory to construct a Bayesian network incorporating Noisy–OR gates. Yang et al. (Contribution 20) incorporated carbon intensity indicators (CIIs) into route decision-making, establishing a profit–CII dual-objective model. They utilized the Gale–Shapley algorithm for the stable matching of cargo pallets and vessels, combined with a genetic algorithm for route optimization, achieving increased profits while meeting IMO (International Maritime Organization) rating requirements. The comprehensive emission analysis by Khayenzeli et al. (Contribution 21) and the advanced speed prediction model by Chen et al. (Contribution 22) provided support for these efforts. They pioneered a GAN-LSTM (Generative Adversarial Network–Long Short-Term Memory) hybrid architecture, leveraging Generative Adversarial Networks to reduce error accumulation in long-term LSTM predictions. Through spatiotemporal feature adversarial training, they offered forward-looking speed estimates for intelligent collision avoidance systems. Sun et al. (Contribution 23) integrated genetic algorithm-simulated annealing with deep Q-learning, driving innovation in port logistics through a hybrid intelligent algorithm-powered underground container system. This completed the construction of a comprehensive technological value chain from ship to port.

In the era of carbon neutrality and digital transformation, Smart and Low-Carbon-Emission-Oriented Maritime Traffic Management and Controlling has emerged as a pivotal direction for the sustainable development of the shipping industry. Leveraging the advancements in artificial intelligence (AI), big data analytics, and intelligent sensing, future maritime traffic systems aim to reduce carbon emissions while enhancing the safety, efficiency, and resilience of maritime operations. One of the core challenges lies in integrating green energy strategies with intelligent traffic control mechanisms, such as adaptive ship routing, energy-efficient voyage planning, and dynamic speed optimization. These solutions rely heavily on real-time data streams from AISs (Automatic Identification Systems), remote sensing, and onboard IoT devices, which enable intelligent decision-making through data-driven predictive models. Moreover, computer vision techniques applied to maritime surveillance videos allow for the real-time recognition of vessel behaviors, collision risk prediction, and automated traffic anomaly detection, thus supporting smart navigation in congested or low-visibility conditions. Outlier suppression in AIS data, multi-source data fusion, and the development of digital twins of ship traffic systems further contribute to the realization of intelligent emission-reducing strategies. As AI continues to evolve, its integration into maritime traffic situation awareness, autonomous ship control, and environmentally adaptive traffic scheduling is expected to play a transformative role in achieving low-carbon maritime logistics.