Ocean Engineering, Год журнала: 2024, Номер 318, С. 120144 - 120144
Опубликована: Дек. 24, 2024
Язык: Английский
Ocean Engineering, Год журнала: 2024, Номер 312, С. 119280 - 119280
Опубликована: Сен. 18, 2024
Язык: Английский
Процитировано
5
Ocean Engineering, Год журнала: 2024, Номер 311, С. 118927 - 118927
Опубликована: Авг. 19, 2024
Ship Time Headway (STH) is the time interval between two consecutive ships arriving in same water area. It serves as a crucial indicator for visually measuring probability of ship congestion and frequency passage busy waterways. Accurately predicting STH effective maritime traffic management. In this paper, we propose deep learning method aimed at simultaneously multiple areas (multi-STH). This integrates Variational Mode Decomposition (VMD) algorithm with Spatial-Temporal Attention Graph Convolution Network (STAGCN) to deeply capture complex spatial-temporal features STHs each areas. sequences were obtained from Automatic Identification System (AIS) reach, ensuring that these remained numerically continuous on timeline. The VMD was employed decompose into multi-feature inputs STAGCN, training model conjunction inland waterway network patterns variation Extensive experiments demonstrate proposed prediction surpasses accuracy robustness other existing methods, exhibiting excellent performance various multi-STH study accounts inherent correlation waterways, substantially improving efficiency compared single-waterway prediction. may have potential provide useful support be practical significance enhancing safety waterways navigation.
Язык: Английский
Процитировано
4
Journal of Marine Science and Engineering, Год журнала: 2025, Номер 13(1), С. 147 - 147
Опубликована: Янв. 16, 2025
The dry bulk shipping network is an important carrier of global commodity flow. To better understand the structural characteristics and future development trends (GDBSN), this study proposes a framework for analysis link prediction based on complex theory. integrates large-scale heterogeneous data, including automatic identification system data port geographic information, to construct GDBSN. findings reveal that exhibits small-world properties, with Port Singapore identified as most influential node. Link results indicate many potential new routes exist within regions or between neighboring countries, exhibiting clear regional clustering characteristics. added links mainly influence local structure, minimal impact overall topology. This provides valuable insights companies in route planning authorities developing strategic plans.
Язык: Английский
Процитировано
0
Applied Sciences, Год журнала: 2025, Номер 15(7), С. 3489 - 3489
Опубликована: Март 22, 2025
To improve the feature extraction method for ship trajectories and enhance trajectory classification performance, this paper proposes a model that combines one-dimensional residual network (ResNet1D) an attention-based Long short-term memory (AttLSTM). The aims to address limitations of traditional methods in extracting patterns jointly represented by non-adjacent local regions trajectories, optimized through introduction self-attention mechanism. Specifically, first utilizes ResNet1D module progressively extract implicit motion pattern features from global levels, while AttLSTM captures temporal sequence trajectories. Finally, fusion these two types generates more comprehensive rich spatiotemporal representation, enabling accurate five including towing vessels, fishing sailing passenger ships, tankers. Experimental results show excels on extensive real-world datasets, achieving accuracy 89.7%, significantly outperforming models relying solely single sets or lacking integrated attention mechanisms. This not only validates model’s superior performance tasks but also demonstrates its potential effectiveness practical applications.
Язык: Английский
Процитировано
0
Journal of Marine Science and Engineering, Год журнала: 2024, Номер 12(9), С. 1672 - 1672
Опубликована: Сен. 18, 2024
This study addresses the challenges of maritime traffic management in western waters Taiwan, a region characterized by substantial commercial shipping activity and ongoing environmental development. Using 2023 Automatic Identification System (AIS) data, this develops robust feature extraction framework involving data cleaning, anomaly trajectory point detection, compression, advanced processing techniques. Dynamic Time Warping (DTW) Hierarchical Density-Based Spatial Clustering Applications with Noise (HDBSCAN) algorithms are applied to cluster revealing 16 distinct patterns, key navigation routes, intersections. The findings provide fresh perspectives on analyzing traffic, identifying high-risk areas, informing safety spatial planning. In practical applications, results help navigators optimize route planning, improve resource allocation for authorities, inform development infrastructure navigational aids. Furthermore, these outcomes essential detecting abnormal ship behavior, they highlight potential surveillance.
Язык: Английский
Процитировано
2
Ocean Engineering, Год журнала: 2024, Номер 318, С. 120126 - 120126
Опубликована: Дек. 20, 2024
Язык: Английский
Процитировано
2
Опубликована: Июнь 26, 2024
A methodology based on deep recurrent models for maritime surveillance, over publicly available Automatic Identification System (AIS) data, is presented in this paper. The setup employs a Recurrent Neural Network (RNN)-based model, encoding and reconstructing the observed ships' motion patterns. Our approach thresholding mechanism, calculated errors between reconstructed patterns of vessels. Specifically, deep-learning framework, i.e. an encoder-decoder architecture, trained using patterns, enabling to learn predict expected trajectory, which will be compared effective ones. models, particularly bidirectional GRU with dropouts, showcased superior performance capturing temporal dynamics illustrating potential learning enhance surveillance capabilities. work lays solid foundation future research domain, highlighting path toward improved safety through innovative application technology.
Язык: Английский
Процитировано
0
Lecture notes in networks and systems, Год журнала: 2024, Номер unknown, С. 11 - 22
Опубликована: Янв. 1, 2024
Язык: Английский
Процитировано
0
Applied Sciences, Год журнала: 2024, Номер 14(24), С. 11751 - 11751
Опубликована: Дек. 16, 2024
The maritime industry is undergoing a paradigm shift driven by rapid advancements in wireless communication and an increase traffic data. However, the existing automatic identification system (AIS) struggles to accommodate increasing data, leading introduction of very-high-frequency (VHF) data exchange (VDES). While VDES increases bandwidth rates, ensuring stable transmission IoT (MIoT) application congested coastal areas remains challenge due frequent collisions AIS messages. This paper presents slot occupancy-based collision avoidance algorithm (SOCA) for network MIoT. SOCA designed mitigate impact interference caused transmissions messages on VDE-Terrestrial (VDE-TER) areas. To this end, provides four steps: (1) construction neighbor information table (NIT) frame maps, (2) candidate list, (3) TDMA channel selection, (4) selection avoidance. operates constructing NIT based estimate intervals updating maps upon receiving monitor usage dynamically. After that, it generates list VDE-TER channels, classifying slots into non-interference categories. then selects that minimizes allocates with low expected occupancy probabilities avoid collisions. evaluate performance SOCA, we conducted experimental simulations under static dynamic ship scenarios. In scenario, outperforms VDES, achieving improvements 13.58% aggregate throughput, 11.50% average latency, 33.60% ratio, 22.64% packet delivery ratio. Similarly, demonstrates 7.30%, 11.99%, 39.27%, 11.82% same metrics, respectively.
Язык: Английский
Процитировано
0
Ocean Engineering, Год журнала: 2024, Номер 318, С. 120144 - 120144
Опубликована: Дек. 24, 2024
Язык: Английский
Процитировано
0