Anomaly Detection in Maritime AIS Tracks: A Review of Recent Approaches

Abstract

The automatic identification system (AIS) was introduced in the maritime domain to increase the safety of sea traffic, with its messages broadcast to nearby ships containing, among others, information about the identification, position, speed, and course of the sending vessels. With its growing use in worldwide surveillance and in support of vessel traffic services and collision risk assessment, there is an urgent need for the automated detection of anomalous AIS data that may indicate events relevant to safety and security. This paper surveys 44 research articles on anomaly detection of maritime AIS tracks, identifies the tackled anomaly types and their potential use cases, and closely examines the landscape of recent research and its limitations — most notably the focus on route-deviation anomalies, the regional confinement of models, the absence of a standard dataset with labeled anomalies, and emerging privacy concerns.

1. Introduction

Freight transport via sea is one of the major backbones of our highly connected global economy today, with global demand for freight transportation expected to more than double and over 70% of goods shipped by sea. In addition to commercial transport vessels, the world's oceans are traveled by a wide range of other ships, such as passenger ships, ferries, fishing vessels, or recreational crafts, making the safety and security of diverse maritime traffic necessary for the continued functioning of the globalized market economy and the well-being of passengers and marine ecosystems.

To augment safety and security at sea, the International Maritime Organization (IMO) designed the automatic identification system (AIS) in the 1990s as a complementary system to high-frequency radar, mandated it under the SOLAS agreement for ships above a certain size, and — with the advent of satellite-mounted AIS receivers — coverage became global and AIS data abundant, turning the system into a leading tool for collision risk assessment and surveillance such as anti-piracy operations and the prevention of illegal fishing.

With over 1,490,776 ships tracked worldwide, a manual unveiling of suspicious ship activities is infeasible, so in recent years many different approaches for automated anomaly detection of maritime AIS tracks have been proposed that exploit the fact that vessel traffic is to some extent predictable, especially within confined local regions, and that vessels cannot behave arbitrarily due to physical constraints or mandatory sea routes: such detectors train a model of normal vessel traffic patterns and mark outliers or violations as anomalies potentially indicating, e.g., accidents or criminal activities.

To provide a concise overview of this research field, this paper reviews current methodologies and their applications: it summarizes the anomaly types that suspicious AIS activity can reveal and maps them to common use cases, then presents a literature survey covering 44 publications on AIS track anomaly detection. It finds that most proposals focus on route-deviation anomaly detection and are limited to the confined geographical region their models were trained on, that authors oftentimes evaluate on inaccessible, self-recorded, or private samples, with notably no established dataset that includes labeled anomalies as ground truth. The rest of the paper covers AIS and anomaly detection background, then the survey, discussion, and conclusion.

2.1. AIS Background and Regulatory Requirements

AIS development began in the early 1990s to provide an additional ship-to-ship awareness system supplementing radar and visual observation, designed for three primary purposes specified by the IMO: first, collision avoidance enabled by exchanging AIS messages between all vessels in a certain vicinity, including the vessels' identities, precise locations, and further information to increase situational awareness and assist navigation; second, support for on-shore vessel traffic services (VTS) for guidance through heavily trafficked areas or particularly dangerous passages such as ports, using dedicated stationary Aids to navigation (AtoN) transceivers; and third, enabling littoral states to identify ships and their cargo in territorial waters, alongside greatly enhanced search and rescue (SAR) operations.

As an amendment to the 1974 SOLAS agreement, the IMO mandated that all ships on international voyages above 300 gross tonnage must use AIS, with many countries specifying stricter rules such as the United States or the European Union. AIS messages are periodically broadcast by transceivers onboard vessels over two VHF radio bands around 162 MHz based on time-division multiple access (TDMA), with an effective horizontal communication range up to 50 km limited by geodetic visibility — though in line-of-sight AIS signals can travel hundreds of kilometers, which is leveraged by satellite-mounted receivers (S-AIS) providing global monitoring and closing the gaps of terrestrial AIS in open-ocean observations.

The actual reporting interval of AIS messages lies in the range of 2–12 s (3 min for ships at anchor) and depends on the vessel's velocity as specified in IMO Resolution 74(69), whose requirements state that AIS electronics onboard vessels must automatically transmit information and automatically receive and process corresponding information, that high-priority and safety-related calls must be answered with minimum delay, with a separate display system for human operators. Positional data of ships is generally obtained via one of the available global navigation satellite systems (GNSS) such as GPS or Galileo.

While the AIS requirements state that cyber security mechanisms should be implemented, it is important to note that the AIS protocol itself is not secured and is publicly accessible: in particular, neither authentication, encryption, nor integrity protection is considered, so AIS is highly vulnerable to a wide range of cyber attacks such as false AIS message injection — not to mention that the privacy issues of ship crews, passengers, owners, and other associated persons are not addressed at all, with severe consequences discussed later.

2.2. Data Format, Surveillance, and Datasets

Information announced by AIS consists of static, dynamic, voyage-related, and safety-related data: static information most importantly includes the maritime mobile service identity (MMSI) number uniquely assigned to each vessel as its primary identifier, plus call sign and name, ship type, dimensions, and antenna location; dynamic data includes the vessel's position (latitude and longitude), accuracy and integrity, time in UTC, course over ground (COG), speed over ground (SOG), heading (HE), rate of turn (ROT), and navigational status; voyage-related data includes draught, hazardous cargo, destination (DST), and estimated time of arrival; safety-related data consists of text messages. Since the paper focuses on track anomaly detection, the dynamic part is of most interest, though destination and draught information may also be suitable for anomaly detection.

The use of AIS as a surveillance tool for authorities has steadily increased since its introduction, especially with satellite-based AIS enabling effective global vessel tracking, and with the increased desire by authorities to enhance maritime security after the September 11 attacks, AIS was recognized as a valuable tool for identifying vessels approaching coasts and ports and preventing possible terrorist attacks. Both civil and military ship-surveillance research has steadily increased, with surveillance approaches developed for a wide variety of domains — e.g., investigating criminal activities, monitoring fishing activities, or optimizing maritime logistics and supply chains.

By design, AIS is an entirely open protocol: communication is by radio broadcast that anyone in range can record and interpret with a suited receiver, and as AIS equipment and satellite-based receivers have become steadily more affordable and commercialized, a plurality of AIS datasets is available. A wide range of commercial providers readily supply AIS data to paying customers via public websites, alongside non-commercial providers offering data free of charge, with large differences between providers in pricing, availability, and quality — most providers generally do not fully provide all information carried via AIS, the available content of dynamic or voyage-related data depends on the individual provider, and tracks recordable only by satellite far from coasts may be excluded by non-commercial providers.

For scientific purposes, it is also possible to directly request AIS data from maritime authorities such as the European Maritime Safety Authority (EMSA), and Tu et al. explored different AIS data sources and assessed data quality among different providers.

3.1. Anomalous AIS Behaviors

The concept of anomalies in AIS tracks can be described as behavior that is not "normal" or not expected during regular operation — e.g., ship velocities not changing too rapidly, or ships usually traveling along common sea routes as observable in visualizations of historical AIS tracks. Following Lane et al., this paper uses five general anomalous behaviors derived from AIS tracks to classify recent research: (a) deviation from standard route, (b) unexpected AIS activity, (c) unexpected port arrival, (d) close approach, and (e) zone entry.

Deviation from standard route can be considered the most elementary anomaly: ships, especially cargo or passenger ships on repetitive routes, generally take the most direct path possible between origin and destination during travel (usually seaports), so in open waters a deviation from a straight route may indicate an anomaly, while in areas where landmasses or narrow sea lanes restrict traffic, ships can be expected to travel in relatively straight paths between common waypoints — particularly at seaport entrances or in canals when guided by VTS.

Unexpected AIS activity encompasses two types: first, AIS signal loss in areas where reception is usually adequate, which may indicate intentional on–off switching of AIS equipment; and second, unexpected route continuations after an AIS outage that might obfuscate malicious actions. Unexpected port arrival can be investigated using voyage-related AIS track data such as destination-port and waypoint information — e.g., when an illegal fishing catch is offloaded at different ports. Long-lasting close approaches between vessels should generally be rare except in emergencies, otherwise possibly indicating illegal activity such as contraband or drug exchange. Finally, zone entry anomalies are vessels entering an area for a significant amount of time where they are not expected or allowed to be, such as marine protected areas or other exclusion zones.

One of the main drawbacks of AIS is its inherent unreliability: AIS tracks often show large blank spots that may be due to unintentional technical problems, radio interference, attenuation, or actual tampering such as intentionally turning off AIS transceivers — and moreover AIS signals can be easily spoofed and manipulated by attackers willing to obscure their locations, as reported in practice. Even with complete tracks, manual detection of vessel movements indicating emergencies or dangerous or illegal behaviors is often infeasible simply because the total number of ships is too large, so much research effort has gone into automated anomaly detection that autonomously identifies anomalous behaviors, possibly acting on knowledge of data representing the norm.

3.2. Applications of AIS-Based Anomaly Detection

Initially developed for safety purposes, the focus of AIS has shifted towards a surveillance and security tool for maritime authorities, where anomaly detection can efficiently identify conspicuous, dangerous, or even illegal behavior at sea across a wide range of applications broadly categorized into monitoring criminal activities, supporting safety activities, and environmental monitoring. Much research focuses on criminal activities such as drug smuggling or piracy: AIS transceivers of Iranian-flagged tankers were tampered with to disguise vessel identities to evade oil-export sanctions, some detection methods have focused on concrete activities such as ship rendezvous (close approach) on the open ocean for smuggling or drug trade, while others have focused on more general anomalies such as on–off AIS switching or route deviations used to detect pirate attacks, for example.

The majority of the anomaly detection research considered in this review confirms the use of the proposed methods for safety activities: anomaly detection may identify zone entry where travel is considered dangerous, can effectively evaluate collision risks (close approach) and thus support traffic surveillance and collision risk prediction, may even decrease coast guards' rescue times by identifying when and where SAR missions occur from the anomalous AIS tracks of involved vessels, and can further increase marine traffic management and planning efficiency — reducing environmental impact and overall shipping costs through lowered accident risk or decreased fuel consumption — while identifying sea routes with many close-approach anomalies could give engineers insights to improve canal and traffic-lane design.

Finally, AIS anomaly detection can benefit environmental monitoring to prevent or reduce marine environmental degradation, since ocean fish depletion due to overfishing poses an immense risk: data analyses of AIS tracks have already helped map European fishing effort, and anomaly detection could supplement existing methods by identifying illegal fishing to properly manage resources and prevent overfishing — with innovative approaches proposed such as detecting intentional AIS switch-off on fishing boats — plus reducing other risks and locating potential environmental incidents. In summary, AIS anomaly detection enables diverse applications across the maritime domain, with likely more future AIS applications increasing the effectiveness of AIS data utilization.

4.1. Related Work and 4.2. Survey Methodology

Interest in maritime anomaly detection has received steady attention over the years, yielding several scientific publications: related to this work are surveys with similar intentions — a 2008 publication deriving a taxonomy for the term anomaly in the maritime domain, followed by a summary of selected approaches in 2011, then the first systematic review by Sidibé and Shu covering 2011–2016 literature focused on detection methodology, approach type, AIS attributes, and dataset used, an extensive survey by Riveiro et al. whose considered approaches are largely generic or focus on non-AIS technologies, Tu et al.'s survey carefully differentiating geographically confined (map-dependent) from universally applicable (map-independent) approaches, the most recent study covering only 18 publications, and finally Zissis et al.'s extensive literature review in the context of modeling maritime routes.

However, while extensive, most studies lack a clear differentiation between methods developed primarily for detecting AIS track anomalies and those originally addressing related problems: some approaches predict vessel trajectories or their time of arrival, and although deviations from these predictions might indicate anomalies, the authors evaluated their approaches only in terms of predictive quality rather than anomaly detection capability — with similar exclusion reasons for related methods investigating AIS signal strengths or optimizing human–machine interface visualizations, which contribute to anomaly detection only as a secondary use case. This survey therefore focuses on publications specifically designed for AIS track anomaly detection, allowing qualitative assessment and comparison of their evaluation methodologies.

To identify publications explicitly concerned with AIS track anomaly detection, a literature review was conducted using the keywords "AIS", "anomaly detection", "maritime", and "tracks", plus publications covered by previous reviews — the main exclusion reason being off-topic papers or those not focusing exclusively on AIS track anomalies — yielding a selection of 44 papers spanning 2007–2021. For each publication, the following detector properties were extracted: the underlying detection method grouping similar methods, the addressed anomaly types, the AIS features used or external non-AIS features, each approach's scope regarding fixed region versus global applicability, vessel scope (specific versus independent), time scope (whether temporal relationships are incorporated), and finally the dataset type and whether evaluation compares against a ground truth of known anomalies.

4.3. Methods for AIS Anomaly Detection

Regardless of the specific method, what most AIS anomaly detectors have in common is requiring a learned underlying model using normal and sometimes anomalous AIS training data, on which basis detectors decide whether new data classifies as normal or anomalous behavior — with only a few model-free approaches, e.g., d'Afflisio et al., who deterministically verify whether trajectories could have occurred during an (intentional) AIS outage considering pre-outage data. The survey identified ten groups of detection methods: one major group for machine learning with either neural networks or clustering — notably DBSCAN for clustering, widely used because it appropriately clusters similar regions such as maritime vessel tracks and can incorporate course (COG) or speed (SOG) as additional features besides position, so tracks outside clustered regions indicate anomalies — and another large group for stochastic methods such as Gaussian mixture models, which assume AIS track properties underlie probabilistic variations.

Complementing these fundamental methodologies, also applied in other intrusion detection domains, further specialized approaches utilizing geometric properties exist — e.g., computing the convex hull of tracks in a certain region or modeling each AIS message as a potential electrical charge decaying over distance — and finally there are approaches that incorporate multiple methods into a comprehensive framework rather than relying on a single detection method, hence assigned to a separate (framework) category in the survey, alongside miscellaneous approaches such as Gaussian processes and Bayesian networks.

The sources reveal that most works group into one of the major groups — machine learning, stochastic, or geometry approaches — besides frameworks or miscellaneous approaches, as the classification of underlying anomaly detection methods illustrates.

4.4. Anomaly Types Addressed by Research

Research is classified according to the five categories proposed by Lane et al.: current AIS track anomaly detection research considers all five anomaly types, yet proposed methods often identify only a single anomaly type. Deviation from a standard route is the most prominent type, addressed in 37 out of 44 publications: many approaches extract frequently traveled sea routes from historical AIS data — e.g., via clustering — then compare unknown AIS tracks to check sufficient similarity to extracted routes or membership in an identified route cluster, and these approaches work very well where many ships take similar routes, as demonstrated for bay areas and port entrances — while route deviations can also be used without a priori knowledge of common routes.

Seven approaches in the survey particularly consider unexpected AIS activity: d'Afflisio et al. aim to identify velocity changes during AIS outages assuming ships travel at similar velocity most of the time, arguing intentional on–off switching generally hides changes in nominal vessel velocity occurring, e.g., when vessels change course — while Singh and Heymann's methodology relies on known AIS tracks from other vessels and also addresses intentional AIS on–off switching detection. Note that many other methods in the unexpected-AIS-activity context aim to detect anomalies by monitoring radio signal strengths or area-specific outages — most notably Mazzarella et al. — which are excluded from this paper due to its clear AIS-track focus, yet represent promising complementary approaches usable alongside pure AIS-based methods to improve detection capability.

Three publications explicitly consider anomalous port arrival: Zor and Kittler, for instance, look at ferries running regular routes on a fixed schedule with only a single ship considered at a time, so departure and expected-arrival ports are known a priori — and since the destination port is AIS-transmitted, one can simply check whether the vessel's track violates the port order. Similarly, close approach anomalies appear in only a few publications studying the number of close ship interactions possibly indicating cargo handover, or forecasting vessels' closest point of arrival, which could benefit VTS navigational guidance. Zone entry is considered only marginally, with restricted zones learned implicitly as part of general shipping routes and trajectories — standalone zone entry perhaps deemed too trivial to implement — though more elaborate methods such as predicting likely imminent zone entry are conceivable.

4.5. Detection Features and 4.6. Scope of Detection Methods

As outlined in Section 2.2, AIS messages carry a wide range of different data, yet all anomaly detection methods generally use only a limited subset of available features — implicitly requiring the MMSI number to recover ship tracks from individual AIS data points. Because AIS-track-based approaches are considered, unsurprisingly all methods use vessel positions (PO), while speed over ground (SOG), heading (HE), or course over ground (COG) are often additionally analyzed; some approaches utilize a minimum of features only, whereas others have seemingly adopted a maximalist approach including most dynamic AIS information plus a vast range of external non-AIS features such as weather or daytime — while only a few approaches use port destinations (DST), vessel types, or status (STA) features, sometimes remaining unclear whether destination information comes from actual AIS data or other sources.

Besides detected anomaly types and utilized AIS features, detector scope may differ largely regarding regions, vessel types, and time: 31 publications incorporate temporal relationships, while the remaining 13 operate time-independently, merely testing, e.g., whether new AIS messages' speed vectors resemble adjacent historic ones. The survey assesses regional scope — whether the learned model is valid only in the confined training-data region or generalizes to unknown untrained areas: the majority of approaches (38) are region-centric, with only six globally applicable, and the geographical size of regional methods varies significantly from bays and port entrances to continental-scale areas — with the major disadvantage of heavy reliance on the region's historical AIS data, making them inapplicable where no or sparse information exists.

Approaches also differ in how vessel types are considered for detection: the distinction is whether methods detect anomalies from one specific vessel type or for any other (unknown) vessel types — apparently many developed methodologies (28 papers) generalize to all other vessels, yet not insignificantly few (16 papers) focus on unique vessels and anomalies, such as Zor and Kittler considering route anomalies of ferries oscillating between two fixed ports, exploiting ferry-traffic characteristics of always following very similar known paths, with models built from individual routes of specific ferries and present behavior analyzed for anomalies — though it remains questionable whether their approach applies to other ships traveling a similar route.

A limitation of single-vessel-specific approaches is that vessels involved in anomalous behavior must be known in advance, and sufficient regular non-anomalous AIS data must be recorded before anomaly detection can occur: while suitable for ferry anomalies, such approaches seem too specific for most scenarios beyond that context — and in security- and safety-related scenarios such as terrorist attacks or emergencies, meaningfully selecting ships for advance training of a vessel-specific detection model is challenging because it is unknown which ships will likely be involved in an upcoming incident or illicit activities.

4.7. Evaluation Datasets and Ground Truth

Besides proposing new anomaly detectors, evaluating their effectiveness and comparing them to the state of the art is essential in scientific research, with the usual procedure of splitting a given dataset into training and test data containing (known) anomalies expected to be found. Regarding dataset types, the survey differentiates publicly available, self-recorded, private/closed, and synthetic AIS data: data sources vary widely between research approaches — only 7 publications revealed their data sources, while 32 approaches are evaluated on closed datasets (22 private, 8 self-recorded, 2 synthetic) and 5 with unknown sources. Many authors obtained data from (military) authorities or defense contractors such as the Swedish navy or BAE Systems, while others relied on commercial platforms such as MarineTraffic, some groups self-recorded needed AIS data given inexpensive commercial off-the-shelf AIS equipment, whereas others artificially generated short AIS tracks claimed consistent with real-world traffic and terrorist behavior.

Moreover, the lack of ground truth is observed as a common challenge in this research area: with real-world data, many researchers struggle to find valid ground truth and resort to artificial AIS anomaly generation — besides 8 publications with decent ground truth derived, e.g., from real reported European incidents, suspected illegal-fishing rendezvous, domain-expert-labeled tracks, or severe weather situations, there are 11 publications with makeshift ground truth including hand-drawn anomalies, random data injection, or added synthetic anomalies, among others. Validating methods exclusively with self-generated anomalies may limit the developed methods' real-world reliability, the approaches' credibility, and research-community comparison.

Overall, this heterogeneous AIS data-source landscape constrains science: a dedicated scientific database for regular AIS data might add little value given already numerous available sources, but such a database including known and confirmed anomalous AIS tracks would be of great value — with three general scientific requirements: first, representativeness in relevance and coverage; second, crucial accessibility with data made available and easily accessible; and most importantly in the AIS anomaly detection context, anomalous traces labeled appropriately. Note that Mao et al. already laid the foundation for a scientific AIS database — but for evaluating trajectory prediction rather than anomaly detection methods.

5. Discussion: Limitations and Privacy Considerations

The literature review reveals that AIS track anomaly detection research follows a relatively homogeneous path: firstly, the bulk of research targets route anomaly detection and thus lacks diversity — zone-entry and unexpected-port-arrival detection methods could be effortlessly implemented and successfully complement route-deviation detection, while close approach anomalies may be more challenging to detect reliably, especially in highly trafficked areas, because they simultaneously require position, course, and speed information of multiple ships — yet surprisingly, particularly given AIS's original safety goal, there is not more research focused on close approach detection, though simple collision avoidance mechanisms are already built into existing AIS equipment and likely uncovered by the survey by design.

The majority of research models "normal" behavior specific to a geographically confined region: while these approaches seem to work very well, they may suffer disadvantages — in areas with insufficient or sparse data they can hardly be used; when entering AIS data or configuring AIS equipment, mistakes are often introduced that can be unintentionally and wrongly incorporated as normal into the learned detection model; and finally, with upcoming autonomous vessels and ships, embedding AIS track anomaly detectors directly into moving vessels may benefit safe automated navigation — yet region-specific approaches may not be fully applicable there since they require re-training for each region the vessel moves to — so these region-specific approaches are deduced as mostly suitable only for static use cases and stationary AIS receivers.

Regarding evaluation datasets, it turns out that in many cases AIS anomaly detection is funded or performed by researchers associated with maritime authorities or defense contractors, so unsurprisingly these entities often supply their datasets too — yet some researchers (8 papers) also recorded their own datasets, while only a few (seven papers) use publicly available sources, and unfortunately known anomalous AIS situations representing reliable ground truth are lacking, so many researchers resort to simulating their own anomalies as substitute ground truth, with existing approaches ranging from manually drawing AIS tracks to simulating data with rigid-hulled inflatable vehicles. Overall, the lack of a common dataset heavily reduces transparency, hinders result replication, and makes sound scientific evaluation and comparison of different approaches nearly impossible — so developing an established, freely available AIS database, as suggested by Mao et al., is urgently necessary to advance the domain's science.

Still, as recently found by Serra-Sogas et al., only a minority of total vessels contribute to AIS, with about 70% remaining invisible — severely limiting detection-method applicability, especially to small recreational vessels accounting for 53% of non-AIS-equipped vessels — so supplementary methods tracking vessels regardless of AIS equipment, such as radar-based ones, may prove beneficial alongside AIS-based anomaly detection. Regarding information security, the Section 2.1 problem is the AIS protocol's general openness, accessible by anyone at any time: the trade-off between security/safety gains through AIS and privacy loss for vessel travelers may skew toward the former, with known crews' and passengers' locations revealable in real time to anyone willing to pay, and globally placed AIS receivers sometimes Internet-connected without any security mechanisms. The IMO noted as early as 2004 that freely available AIS data may undermine the initial safety and security goals, the European Data Protection Supervisor acknowledged AIS may have infringed EU data protection regulation since inception, with similar US freedom-of-information-act considerations. To the authors' knowledge, no anomaly detection research considers privacy issues: existing work may identify and track individuals involved in (suspected) illicit activities by anyone obtaining AIS data — including rights-violating authorities and governments that might actively prohibit identified SAR missions and persecute those involved — partially remediable by extending confidentiality and authentication protection to AIS as suggested by Goudossis and Katsikas, Aziz et al., and Kessler.

6. Conclusions

To increase the safety and security of globally growing maritime traffic, AIS was introduced by fitting transceivers to each vessel, primarily to suit as a digitized ship-to-ship collision avoidance system. With the ability to overhear AIS position broadcasts from ships worldwide — e.g., via satellite-based receivers — AIS became the leading tool enabling surveillance of the entire maritime domain, and with plenty of available AIS datasets and vessels following predictable routes and maneuvers, automated anomaly detection promises to detect unintended or even malicious behavior, e.g., by learning frequent shipping routes.

In this paper, a literature survey classified academic anomaly detection approaches utilizing AIS tracks, and based on a detailed assessment of 44 publications, it reveals that most research tackles deviations from expected or well-known shipping routes by focusing on confined regions rather than training universally applicable models — suitable for stationary VTSs or repetitive-route vessels but hindering adoption of new use cases such as future autonomous vessels on worldwide cargo routes — with objective, sound comparison between approaches hardly possible due to closed evaluation datasets and missing anomaly ground truth. With AIS as a powerful worldwide surveillance tool contradicting individuals' privacy rights, as recognized by the IMO and at state level, awareness should be raised — especially for future AIS tracking and automatic anomaly detection — and AIS extensions providing both pseudonymity and confidentiality to vessels are needed to protect the privacy and data sovereignty of individuals at sea.

Abbreviations

AIS: automatic identification system; AtoN: aids to navigation; COG: course over ground; COTS: commercial off-the-shelf; DST: destination; EMSA: European Maritime Safety Authority; GNSS: global navigation satellite system; GPS: global positioning system; HE: heading; IMO: International Maritime Organization; MMSI: maritime mobile service identity; PO: position; ROT: rate of turn; S-AIS: satellite-based AIS; SAR: search and rescue; SOG: speed over ground; SOLAS: safety of life at sea; STA: status; TDMA: time-division multiple access; VHF: very high frequency; VTS: vessel traffic service.