Automatic Identification System Research Articles

Ship-Route Prediction Based on a Long Short-Term Memory Network Using Port-to-Port Trajectory Data

In this study, a ship-route prediction model based on a long short-term memory network using port-to-port trajectory data is presented. Data from a traditional automatic identification system—often used for ship-route prediction—are limited by uneven sampling intervals and noise. To address these issues, equal-interval data collected every 10 s from a target ship, which is a liner container vessel, were employed. Our study focuses on predicting the entire trajectory between the Gunsan and Busan ports. The root mean square error (RMSE), mean absolute error (MAE), and average distance d¯ between two trajectories were used as the key evaluation metrics. The analysis yielded excellent predictive performance, with the values RMSE = 0.000999, MAE = 0.000672, and d¯ = 0.101 km. This study thus provides a foundation for predicting complete port-to-port routes and offers practical insights for managing vessel operations. Accurate route prediction contributes to reducing port congestion, improving fuel efficiency, and lowering carbon emissions.

Illegal ship identification system based on trajectory similarity measurement

The Automatic Identification System (AIS) relies on the identification code (MMSI) of a vessel to identify it and records its trajectory data, which is an important technical means for modern maritime management. However, there are some vessels that violate regulations by opening multiple AIS transponders, resulting in multiple location trajectory records for the same vessel. This behavior is difficult to distinguish through traditional means and has a negative impact on the accuracy and effectiveness of maritime management, thereby threatening the safety of navigation. This system uses big data processing tool Spark to efficiently process and analyze large amounts of AIS data and detects the similarity of vessel trajectories to automatically identify the violating vessels that open multiple AIS transponders. In this way, the system not only improves the detection efficiency of violations, but also provides more reliable evidence for maritime management departments to ensure the safety and order of navigation routes.

Robust deep-learning based refrigerator food recognition.

Automatic food identification utilizing artificial intelligence (AI) technology in smart refrigerators presents an innovative solution. However, existing studies exhibit significant limitations. Achieving consistent high performance in recognition across varying camera distances and diverse real-world conditions remain a formidable challenge. Current approaches often struggle to accurately recognize items in scenarios involving occlusions, variable distortions, and complex backgrounds, thereby limiting their practical applicability in household environments. This study addresses these deficiencies by enhancing the Feature Pyramid Network (FPN) of YOLACT with an additional layer designed to capture nuanced information. Furthermore, we propose a two-stage data augmentation method that simulates diverse conditions including distortion and occlusion, to generate images that reflect various backgrounds and handheld scenarios. Comparative analyses with previous research and evaluations on our original dataset demonstrate that our approach significantly improves recognition rates for both typical and challenging real-world images. These enhancements contribute to more effective food waste management in households and indicate broader applications for automatic identification systems.

Information Extraction from Automatic Identification System (AIS) Data for Identifying Shipping Line based on Maritime Mobile Service Identity (MMSI)

Abstract Automatic Identification System (AIS) is a technology used to track the movement of ships through the transmission of ship data using Very High Frequency (VHF) waves. Ships around the world transmit their information, which is then captured by AIS receivers. This AIS data collected and stored at the ground station. This study uses the data on the number of ships categorized based the Maritime Mobile Service Identity (MMSI) in the southern Indonesian Archipelago Sea Route II (ALKI II), Madura strait, and its surroundings during one month period of March 2023. The AIS raw data processed using The Jupyter Notebook to generate the information such as MMSI, longitude, latitude, and navigation status. The information then plotted on QGIS to see the difference pattern of shipping lanes based on the ship’s MMSI number. The aim of this research is to demonstrate how AIS data processing and analysis can generate sufficient information for maritime spatial planning, including depicting maritime traffic density and traffic constellations, shipping lanes and navigation status, the hierarchical network of maritime routes, fishing zones, and spatial-temporal interactions between shipping activities. The findings from maritime spatial planning can offer valuable insights for law enforcement at sea and offer guidance for developing port infrastructure based on specific needs through the utilization of AIS data.

Air quality monitoring and assessment based on AIS data

Abstract Indonesia is an archipelagic country, where sea transportation is highly needed to connect the islands among Indonesian regions. Indonesian waters are also passed by international shipping passing through the Indonesia Archipelagic Sea Lane. Number of ships tends to increase by increasing the economic activities in Indonesia. The increasing number of ships will potentially cause the increase of air pollution in Indonesia waters. AIS consortium has been established since 2020 focusing on conducting join research and development on several issues that can be solved using Automatic Identification System (AIS) data gathered from AIS devices that have been installed in several universities as members of the consortium. One potential use of AIS data is the estimation of capacity and spreading of air pollution caused by shipping activities in Indonesia waters. As AIS can provide ship’s navigation status time-to-time, the estimation of fuel consumption as well as the emission can be estimated in real time. For highly accusation fuel oil consumption prediction is done by using the Random Forest Model and ship emissions including CO2, NO2, SO2, PM2.5, and PM10 are estimated based on emission factors from ships. This analysis is mapping the contribution of air pollution based on ship type in the Tanjung Perak port area. Later, it is expected to become the basis for Indonesia’s maritime operation policy in order to monitor and control the emission contributed by the ships in Indonesian waters.

Trajectory reconstruction of buoys using Qingdao harbour data: Filling missing buoy trajectories using LSTM-Attention model

In maritime operations, instances occur where vessels flee after colliding with buoys. Automatic Identification System (AIS) data collected from these buoys play a vital role in identifying the responsible party and determining liability. However, data from buoys can be lost in rough sea conditions and equipment failure, which makes it difficult to identify accidents where ships collide with buoys. To address this issue, this study proposes a model that combines Long Short-Term Memory Network (LSTM) with Multiple Attention mechanisms to reconstruct buoy trajectories with high precision. In this paper, the buoy trajectory reconstruction test is carried out using Qingdao Harbour Buoy 304 a case study, and the validation results show that the mean squared error (MSE) of this model is 12.15. In addition, the LSTM-Attention model shows a significant improvement in all the metrics compared with other models: compared with the Random sampling model, the mean absolute error (MAE) is improved by 75.15%, and the root mean squared error (RMSE) by 71.02%; MSE by 77.31%, MAE by 36.70%, and RMSE by 52.36% compared to the Cubic spline model; and MSE by 49.92%, MAE by 27.35%, and RMSE by 29.52% compared to the Hermite model. These results show that the LSTM-Attention model significantly improves the accuracy and reliability of trajectory reconstruction.

Modeling of deep ocean ambient noise and sound-ray-arrival-grazing-angle estimation for shallow receivers.

The ambient noise model of isotropic fields is not applicable to shallow-located platforms in the deep ocean, and the conventional methods for estimating the sound-ray-arrival-grazing-angle of targets do not account for the effects of surface noise. Therefore, target parameter estimation methods based on a single vector hydrophone are explored in this study. The approach used in this study integrates the hydroacoustic physical model, signal processing method, and ocean ambient. A time-domain model of the vector field of deep ocean ambient noise was developed for receivers positioned at shallow depths, followed by derivation of the covariance matrix of the single vector hydrophone based on this model. Subsequently, a target signal-to-noise ratio (SNR) estimation method using the covariance matrix derived from the single vector hydrophone was formulated. This method effectively addresses the challenge of distinguishing between target signal and noise by transforming the power estimation problem into a covariance matrix solving the task. Finally, a refined approach for estimating the sound-ray-arrival-grazing-angle of target is proposed, aiming to theoretically mitigate the impact of surface noise on the target signal. The experimental data obtained from a deep ocean region in the South China Sea indicate that the findings obtained using the method proposed in this study are consistent with the reference values derived from automatic identification system information. The method also demonstrates reliable estimation results even when the SNR exceeds -5 dB. The conceptual framework developed for SNR and sound-ray-arrival-grazing-angle estimation in this study can be readily applied to other ambient models, indicating potential applications in engineering field. The primary objective of the study was to enhance and augment underwater acoustic signal processing methods for shallow receivers deployed in the deep ocean.

Vessel trajectory prediction in harbors: A deep learning approach with maritime-based data preprocessing and Berthing Side Integration

This study proposes a novel deep learning approach for vessel trajectory prediction in harbor environments, emphasizing the integration of berthing side information. Using Automatic Identification System (AIS) data from Busan Port, we compare Bidirectional Long Short-Term Memory (Bi-LSTM), Bidirectional Gated Recurrent Unit (Bi-GRU), and Transformer models. We identify the berthing side (port or starboard) as a crucial factor that influences vessel trajectories during berth approach, revealing significant differences in final approach paths. Port-side berthing vessels typically take the shortest course from the fairway endpoint, while starboard-side vessels approach parallel to the berth before final berthing. We employed advanced data preprocessing techniques, including DBSCAN clustering, cubic spline interpolation, and Gaussian noise-based data augmentation. Models were trained and evaluated on datasets with and without berthing side information. The Bi-LSTM model outperformed others in most scenarios, improving prediction accuracy by up to 23% when using berthing side-specific datasets. Paired t-tests confirmed statistically significant performance improvements for the Bi-LSTM model. This research enhances maritime safety and efficiency by providing more accurate trajectory predictions, enabling Vessel Traffic Services Operators (VTSOs) to respond to potential hazards swiftly and assisting pilots in developing safer berthing plans.

Analysis of Data Communication Networks for VTS Surabaya in Class I Type A Navigation District Tanjung Perak

A reliable data communication network is essential in supporting Vessel Traffic Service (VTS) operations, especially in managing vessel traffic in the Class I Navigation District of Tanjung Perak, Surabaya. However, challenges such as network reliability, bandwidth capacity, and potential interference are still obstacles in ensuring efficient operations. This research aims to analyze the condition of the data communication network at the Tanjung Perak Class I Navigation District VTS, focusing on the technology used, its performance, and the obstacles faced. This research uses a qualitative descriptive approach, with data collected through direct observation, interviews with VTS operators, and analysis of related documents. The results show that the communication network at VTS uses Very High Frequency (VHF) frequency-based radio communication technology to support operations, such as digital maps of ship traffic and Automatic Identification System (AIS) systems. In conclusion, the data communication network at the Tanjung Perak Class I Navigation District VTS has met basic operational needs, but still requires further development in terms of technology and infrastructure to improve its efficiency and reliability. This research provides recommendations for the integration of advanced technology, infrastructure improvement, and human resource training to optimize the performance of the communication network.

Development of HTC-DBSCAN: A Hierarchical Trajectory Clustering Algorithm with Automated Parameter Tuning

Existing route-clustering methods often fail to identify abnormal sections or similarities between routes, mainly when working with large or long datasets. While sub-route clustering can detect regional patterns, it struggles to accurately capture the overall route structure. The present study proposes a new ship route-clustering method that enhances computational efficiency and noise recognition while addressing these limitations. We refined Automatic Identification System data via four data-cleaning processes and applied a statistical distance measurement to assess ship trajectory similarity. Dimensionality reduction was then used to facilitate clustering. The clustering of ship route similarities is non-parametric and can be applied to datasets not separated based on density to find clusters of various densities. Density-Based Spatial Clustering of Applications (DBSCA) applies to many research fields; using the DBSCA with Noise (DBSCAN) algorithm, we propose an improved DBSCAN algorithm that automatically determines the parameters Epsilon and MinPts. In this study, as a core ship route-clustering process, we propose a sub-route clustering process by setting the distance and density of data points to clear standards for re-analysis and completion. The proposed approach demonstrates markedly enhanced clustering performance, offering a more sophisticated and efficient basis for ship route decision-making.

A machine learning method for the recognition of ship behavior using AIS data

The efficiency of maritime traffic management and the safety of ship navigation have become top priorities. This study introduces a ship behavior recognition method that utilizes the Extreme Gradient Boosting (XGBoost) classification model, in conjunction with the Sparrow Search Algorithm (SSA), to enhance proactive maritime traffic management. The method leverages Automatic Identification System (AIS) data as its primary source and involves several critical steps. Initially, the AIS data is preprocessed, and ship behaviors are encoded. Subsequently, the encoded behaviors are clustered using spectral clustering to create a labeled dataset. Then, this dataset is employed to train and validate the SSA-XGBoost classification algorithm for identifying ship behaviors. Finally, an example analysis is performed in the Yangtze River. The results indicate that the proposed method can accurately and swiftly identify ship behaviors, achieving an accuracy of 97.28%, precision of 96.97%, recall of 97.43%, and an F1 score of 97.19%, surpassing the performance of the existing algorithms. The findings have the potential to aid maritime supervision authorities in promptly assessing ship navigation statuses and provide a valuable reference for developing ship scheduling decisions.

Fine-scale behaviour and population estimates of endangered sooty albatross suggest low exposure (while not excluding high sensitivity) to bycatch

Recent developments in assessing species-specific seabird bycatch risks have demonstrated that fine-scale approaches are essential tools to quantify interactions with fishing vessels and to understand attraction and attendance behaviours. Matching boat movement with bird tracking data specifically allows us to investigate seabird-fisheries interactions for cryptic species for which on-board information is critically lacking. The sooty albatross Phoebetria fusca overlaps with fisheries throughout its range and is known to be vulnerable to incidental bycatch. Combining radar detectors, GPS and behavioural data from individuals from the Crozet Islands and boat locations during the incubation period, we investigated interactions of sooty albatrosses with fisheries in the southern Indian Ocean. Individuals foraged mostly in sub-tropical international waters, where they only encountered a small number of boats, all reporting to the Automatic Identification System (AIS). The low interaction rate during this period may suggest that sooty albatrosses are not strongly attracted to fishing vessels, and that attraction rates may vary between populations. However, this result should be interpreted with caution due to the low sample size and fishing effort during the study period, as these observations may conceal a higher bycatch risk during intense fishing effort and/or energetically demanding periods. The conservation status of this species requires further data to be collected throughout the annual cycle to provide an accurate assessment of the threat.

Spatiotemporal Analysis of Light Purse Seine Fishing Vessel Operations in the Arabian High Seas Based on Automatic Identification System Data

Understanding the dynamic spatial distribution and characteristics of fishing activities is crucial for fisheries management and sustainable development. In recent years, small pelagic fish and cephalopods in the Arabian Sea have become new targets for light purse seine fishing; however, there is a lack of publicly available reports. This study uses automatic identification system (AIS) data from January to May and October to December of 2021 to 2022 in the region between 58°–70° E and 10°–22° N to extract spatial distribution information through three methods. The results show that with a spatial resolution of 0.25° × 0.25°, the spatial similarity index between the fishing ground information extracted in 2022 and catch data was consistently above 0.60, reaching 0.76 in March 2021 and 0.79 in November 2022, while the spatial similarity index in March 2022 exceeded 0.71. The spatial distribution of fishing effort and kernel density was similar to that of the fishing grounds, and the fishing intensity information exhibited the highest spatiotemporal similarity with commercial catch data, making it more suitable as a substitute for fishery data. Therefore, effective international cooperation and efficient joint management mechanisms for fishing vessels are needed to enhance the regulatory oversight of fishing vessels in this region. Integrating AIS data with other technological methods is crucial for more effective monitoring and management of fishing vessels. The findings presented in this paper provide both quantitative and qualitative scientific support for resource conservation and sustainable development in the region.

Classification of Ship Type from Combination of HMM–DNN–CNN Models Based on Ship Trajectory Features

This study proposes an enhanced ship-type classification model that employs a sequential processing methodology integrating hidden Markov model (HMM), deep neural network (DNN), and convolutional neural network (CNN) techniques. Four different ship types—fishing boat, passenger, container, and other ship—were classified using multiple ship trajectory features extracted from the automatic identification system (AIS) and small fishing vessel tracking system. For model optimization, both ship datasets were transformed into various formats corresponding to multiple models, incorporating data enhancement and augmentation approaches. Speed over ground, course over ground, rate of turn, rate of turn in speed, berth distance, latitude/longitude, and heading were used as input parameters. The HMM–DNN–CNN combination was obtained as the optimal model (average F-1 score: 97.54%), achieving individual classification performances of 99.03%, 97.46%, and 95.83% for fishing boats, passenger ships, and container ships, respectively. The proposed approach outperformed previous approaches in prediction accuracy, with further improvements anticipated when implemented on a large-scale real-time data collection system.

Shedding light on trawl fishing activity in the Mediterranean Sea with remote sensing data

Abstract This study uses Synthetic Aperture Radar (SAR) vessel detections and Automatic Identification System (AIS) to predict trawl fishing intensity and distribution of fishing activity in areas where public AIS data are not available. By processing SAR data, considering spatial and temporal autocorrelation, and building a General Additive Model, a statistical relationship between SAR vessel detections and AIS fishing activity was established. The study provides spatially explicit estimates of trawler fishing activity, compared with official fleet records published by the General Fisheries Commission of the Mediterranean, revealing the distribution and intensity of trawl fishing activity not previously publicly tracked. Fishing grounds in the Strait of Sicily along the coast of Tunisia and North of Egypt showed an intensity of trawl fishing activity similar to the Adriatic Sea. This area is historically known to be subject to the highest trawling pressure in the Mediterranean, and also as one of the most heavily trawled regions in the world. The study shows that the integration of remote sensing data, such as SAR, offers a promising avenue to overcome data gaps and improve fisheries management in the Mediterranean where only a portion of the fishing fleet is publicly tracked.

Betta Fish Identification System Based On Convolutional Neural Network

This study developed an automated identification system based on the Convolutional Neural Network (CNN) to classify Betta Splendens, a fish species with high economic value in Indonesia. The system aims to improve accuracy and efficiency in the identification process. The research was divided into several experiments, where the data was split into 320 images for training, 80 for validation, and 100 for testing. We used two optimizers, Adam and RMSprop. The Adam optimizer experiments conducted two stages with learning rates of 0.0001 and 0.001, each with 100 and 200 epochs. The results showed that a lower learning rate (0.0001) with 200 epochs yielded the best test accuracy of 71%, while a learning rate of 0.001 caused accuracy to stagnate at 66%, indicating potential overfitting. The RMSprop optimizer with a learning rate of 0.00001 demonstrated good stability, though with slightly lower accuracy than Adam. This study highlights the importance of selecting the appropriate learning rate and number of epochs to achieve an optimal balance between training, validation, and testing accuracy, ensuring the model generalizes well to new data.

Application of simulated AIS data to study the collision risk system of ships

Abstract In previous research, several computational methods have been proposed to analyse the navigation, transportation safety and collision risks of maritime vessels. The objective of this study is to use Automatic Identification System (AIS) data to assess the collision risk between two vessels before an actual collision occurs. We introduce the concept of an angle interval in the model to enable real-time response to vessel collision risks. When predicting collision risks, we consider factors such as relative distance, relative velocity and phase between the vessels. Lastly, the collision risk is divided into different regions and represented by different colours. The green region represents a low-risk area, the yellow region serves as a cautionary zone and the red region indicates a high-alert zone. If a signal enters the red region, the vessel's control system will automatically intervene and initiate evasive manoeuvres. This reactive mechanism enhances the safety of vessel operations, ensuring the implementation of effective collision avoidance measures.

Environmental and Cost Assessments of Marine Alternative Fuels for Fully Autonomous Short-Sea Shipping Vessels Based on the Global Warming Potential Approach

This research paper presents an effective approach to reducing marine pollution and costs by determining the optimal marine alternative fuels framework for short-sea shipping vessels, with a focus on energy efficiency. Employing mathematical models in a Python environment, the analyses are tailored specifically for conventional and fully autonomous high-speed passenger ferries (HSPFs) and tugboats, utilizing bottom-up methodologies, ship operating phases, and the global warming potential approach. The study aims to identify the optimal marine fuel that offers the highest Net Present Value (NPV) and minimal emissions, aligning with International Maritime Organization (IMO) regulations and environmental objectives. Data from the ship’s Automatic Identification System (AIS), along with specifications and port information, were integrated to assess power, energy, and fuel consumption, incorporating parameters of proposed marine alternative fuels. This study examines key performance indicators (KPIs) for marine alternative fuels used in both conventional and autonomous vessels, specifically analyzing total mass emission rate (TMER), total global warming potential (TGWP), total environmental impact (TEI), total environmental damage cost (TEDC), and NPV. The results show that hydrogen (H2-Ren, H2-F) fuels and electric options produce zero emissions, while traditional fuels like HFO and MDO exhibit the highest TMER. Sensitivity and stochastic analyses identify critical input variables affecting NPV, such as fuel costs, emission costs, and vessel speed. Findings indicate that LNG consistently yields the highest NPV, particularly for autonomous vessels, suggesting economic advantages and reduced emissions. These insights are crucial for optimizing fuel selection and operational strategies in marine transportation and offer valuable guidance for decision-making and investment in the marine sector, ensuring regulatory compliance and environmental sustainability.

Analyzing AIS and wave hindcast data for global wave scatter diagrams with seasonality

Global scatter diagrams are used to estimate possible ship motions in a specified route and are required for comprehensive cargo-loading considerations of merchant ships. This study aims to update global wave scatter diagrams using the latest wave hindcast and Automatic Identification System (AIS) data. The AIS data for over 7500 merchant ships were collected between 2015 and 2022. The wave hindcast data ECMWF Reanalysis v.5 (ERA5), preferred for its recent coverage, were adjusted based on CCI satellite data. The duration of the AIS data is much shorter than the design lifetime of merchant ships, which is generally 25 years. This study found almost linear relationships between the maximum significant wave heights in storms and the generalized extreme value (GEV) parameters representing the distribution of encountered significant wave heights, with adjustments made for seasonality and tropical cyclones. By simulating the parameters based on these relationships and the wave hindcast data over 25 years, the encountered significant wave height distributions over 25 years were obtained. The results indicated a reduction in extremely significant wave heights compared with the BMT-GWS, in agreement with previous studies. This study also found that the difference in the wave period distribution between natural and encountered sea states was small in most sea areas for all significant wave height levels. Taking advantage of the abundant data volume of ERA5, this study adopted an empirical distribution of the wave period conditioned by the significant wave height. The novelty of this study is that we obtained wave scatter diagrams globally by month to consider the seasonal effect in detail and evaluated the confidence interval of the encountered wave height in its probability tail. In addition, this study quantifies the uncertainty of the wave height distribution by Monte Carlo simulation, which is not within the scope of previous studies on wave scatter diagrams based on the significant wave heights encountered by merchant ships around the world.

TPTrans: Vessel Trajectory Prediction Model Based on Transformer Using AIS Data

The analysis of large amounts of vessel trajectory data can facilitate more complex traffic management and route planning, thereby reducing the risk of accidents. The application of deep learning methods in vessel trajectory prediction is becoming more and more widespread; however, due to the complexity of the marine environment, including the influence of geographical environmental factors, weather factors, and real-time traffic conditions, predicting trajectories in less constrained maritime areas is more challenging than in path network conditions. Ship trajectory prediction methods based on kinematic formulas work well in ideal conditions but struggle with real-world complexities. Machine learning methods avoid kinematic formulas but fail to fully leverage complex data due to their simple structure. Deep learning methods, which do not require preset formulas, still face challenges in achieving high-precision and long-term predictions, particularly with complex ship movements and heterogeneous data. This study introduces an innovative model based on the transformer structure to predict the trajectory of a vessel. First, by processing the raw AIS (Automatic Identification System) data, we provide the model with a more efficient input format and data that are both more representative and concise. Secondly, we combine convolutional layers with the transformer structure, using convolutional neural networks to extract local spatiotemporal features in sequences. The encoder and decoder structure of the traditional transformer structure is retained by us. The attention mechanism is used to extract the global spatiotemporal features of sequences. Finally, the model is trained and tested using publicly available AIS data. The prediction results on the field data show that the model can predict trajectories including straight lines and turns under the field data of complex terrain, and in terms of prediction accuracy, our model can reduce the mean squared error by at least 6×10−4 compared with the baseline model.