Maritime Domain Awareness: AI’s Growing Role

Introduction

Artificial intelligence is transforming maritime domain awareness by fusing satellite AIS data, synthetic aperture radar imagery, and RF signal detection to monitor vessel activity across the world’s oceans – revealing that up to 76 percent of industrial fishing vessels operate outside public tracking systems, according to a 2024 Nature study analyzing coastal waters from 2017 to 2021.

The world’s oceans carry 90 percent of global trade by volume. Monitoring this vast maritime domain presents an extraordinary challenge: millions of square miles of ocean, hundreds of thousands of vessels, and threats ranging from piracy and illegal fishing to strategic competition between great powers. The International Maritime Organization defines maritime domain awareness as “the effective understanding of anything associated with the maritime domain that could impact the security, safety, economy, or environment.” AI systems now process data from satellite constellations, coastal radar networks, and RF detection satellites to build that understanding at a scale no human workforce could match.

Satellite Constellations and AIS Monitoring

At any given moment, approximately 200,000 vessels broadcast their positions via the Automatic Identification System, with the IMO requiring AIS transponders on all international-voyage ships above 300 gross tonnage since December 2004 – creating the foundational dataset for global maritime surveillance that AI systems now analyze at scale.

The foundation of modern maritime surveillance is the Automatic Identification System. AIS transponders broadcast vessel identity, position, course, speed, and destination. The International Maritime Organization mandated AIS carriage for all ships of 300 gross tonnage and above on international voyages, all cargo ships of 500 gross tonnage and above regardless of route, and all passenger ships regardless of size, with requirements becoming effective for all covered vessels by 31 December 2004. This data, collected by a global network of terrestrial receivers and satellite constellations, provides unprecedented visibility into maritime traffic.

Global Fishing Watch uses machine learning to process AIS data from these broadcasts, with approximately 200,000 vessels publicizing their locations via AIS at any given time. Their algorithms distinguish between different fishing techniques – purse seine, trawl, longline, and pole-and-line fishing – based on distinctive AIS track patterns. Panama and Chile have released their Vessel Monitoring System data to the platform, expanding coverage of fishing fleets that may not carry AIS.

Multiple commercial providers now offer space-based AIS services. SpaceX’s Starlink constellation and dedicated maritime AIS satellite operators have expanded satellite AIS coverage beyond the line-of-sight range of coastal receivers, enabling tracking of vessel movements in open ocean areas previously invisible to terrestrial stations.

Synthetic Aperture Radar and Satellite Imagery

SAR satellites provide all-weather, day-night vessel detection that deep learning algorithms process with high accuracy, while the European Space Agency’s Copernicus programme delivers free SAR and optical imagery through its Sentinel satellites – enabling researchers worldwide to develop and test maritime detection algorithms that complement AIS tracking.

Beyond AIS, synthetic aperture radar satellites provide all-weather, day-night surveillance capability. SAR satellites detect vessels regardless of cloud cover or darkness, filling critical gaps in optical imaging coverage. Convolutional neural networks trained on SAR imagery have achieved vessel detection accuracy exceeding 90 percent on standard benchmark datasets, according to multiple studies published in the IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

Capella Space and ICEYE have deployed commercial SAR constellations offering sub-meter resolution imaging. These systems can distinguish between vessel types, estimate vessel dimensions, and in some cases identify specific vessels based on radar signature characteristics. The European Maritime Safety Agency operates the Copernicus Maritime Surveillance service, combining satellite imagery with vessel identification and behavioral analysis to support EU member states’ maritime monitoring operations.

The European Space Agency’s Copernicus programme provides both SAR and optical imaging through its Sentinel-1 and Sentinel-2 satellites. Free access to Copernicus data has accelerated AI development for maritime applications, creating a global commons of training data for vessel detection algorithms.

Dark Vessel Detection

A January 2024 study published in Nature found that 72 to 76 percent of the world’s industrial fishing vessels are not publicly tracked, with much of that unmonitored fishing concentrated around South Asia, Southeast Asia, and Africa – exposing a massive surveillance gap that AI-driven sensor fusion between AIS, SAR, and RF detection now works to close.

A subset of vessels deliberately disable their AIS transponders to avoid tracking. These “dark vessels” may be engaged in sanctions evasion, illegal fishing, or other activities that operators wish to conceal. Detecting them represents one of maritime domain awareness’s most challenging problems.

The landmark study by Fernando Paolo, David Kroodsma, and colleagues at Global Fishing Watch, published in Nature in January 2024 (DOI: 10.1038/s41586-023-06825-8), combined satellite imagery, vessel GPS data, and deep-learning models to map industrial vessel activities across the world’s coastal waters from 2017 to 2021. The researchers found that 72 to 76 percent of the world’s industrial fishing vessels are not publicly tracked, with much of that fishing taking place around South Asia, Southeast Asia, and Africa. They also found that 21 to 30 percent of transport and energy vessel activity is missing from public tracking systems.

"The world's population increasingly relies on the ocean for food, energy production and global trade, yet human activities at sea are not well quantified. We combine satellite imagery, vessel GPS data and deep-learning models to map industrial vessel activities and offshore energy infrastructure across the world's coastal waters from 2017 to 2021."
-- Paolo et al., "Satellite mapping reveals extensive industrial activity at sea," Nature 625, 85-91 (2024)

HawkEye 360 operates a constellation of 36 RF-detection satellites that collect and analyze radio frequencies used by ships for navigation, identifying vessels whose true locations diverge from AIS data or whose transponders have been disabled entirely. In May 2025, the U.S. Defense Security Cooperation Agency approved a $131 million foreign military sale of HawkEye 360 technology to India, including SeaVision software, to enhance maritime domain awareness in the Indo-Pacific region.

Illegal Fishing Detection

The UN Food and Agriculture Organization estimates illegal, unreported, and unregulated fishing causes global losses of approximately $23 billion per year, with IUU fishing accounting for up to 30 percent of total catches in some important fisheries and up to 37 percent of fish caught in West Africa.

Illegal, unreported, and unregulated fishing threatens marine ecosystems, undermines legitimate fishing economies, and in some regions enables forced labor. AI-enabled maritime domain awareness provides new tools for detecting and deterring IUU fishing at scale.

According to the UN Food and Agriculture Organization, IUU fishing causes estimated global losses of approximately $23 billion per year. NOAA Fisheries acknowledges that while precise quantification is difficult, “there is little disagreement that it is in the billions, or even tens of billions, of dollars each year.” IUU fishing accounts for up to 30 percent of total catches in some important fisheries, with West Africa losing up to 37 percent of its catch to unreported or illegal fishing.

The Paolo et al. Nature study’s finding that three-quarters of industrial fishing vessels operate outside public tracking systems underscores the scale of the detection challenge. Global Fishing Watch’s machine learning models analyze AIS track patterns to identify likely fishing behavior even when vessels do not self-report as fishing vessels. When AIS signals disappear, SAR satellite imagery provides an independent observation layer – detecting vessel radar signatures in ocean areas where AIS-tracked vessels should be but are not.

The CSIS Asia Maritime Transparency Initiative published the most comprehensive study to date of China’s maritime militia, identifying over 120 militia vessels in the South China Sea using remote sensing data and open-source Chinese-language research. These vessels, ostensibly commercial fishing boats, operate alongside Chinese law enforcement and military to assert control over disputed waters – a gray-zone challenge that traditional AIS monitoring alone cannot address.

Naval Operations and Autonomous Systems

The U.S. Navy established Task Force 59 in September 2021 as its first unmanned and AI task force, operating autonomous surface vessels alongside manned platforms in the Middle East before transitioning its capabilities to the Fifth Fleet in 2024, while DARPA’s completed Ocean of Things program demonstrated persistent maritime awareness through distributed sensor networks of intelligent floats.

AI-enhanced maritime domain awareness supports operational naval forces through real-time operational pictures and autonomous sensing platforms. The U.S. Navy established Task Force 59 in September 2021 under U.S. Naval Forces Central Command as its first task force dedicated to integrating unmanned systems and artificial intelligence into maritime operations. Operating in the Middle East, TF-59 deployed multiple unmanned surface vessel types – including Saildrone Explorers and MARTAC T-38 Devil Ray vehicles – to conduct surveillance patrols across the Fifth Fleet’s area of responsibility. The task force transitioned its unmanned capabilities to the broader Fifth Fleet structure in 2024 after demonstrating that autonomous platforms could extend maritime awareness coverage.

DARPA’s Ocean of Things program, now completed, explored persistent maritime situational awareness through thousands of small, intelligent sensor floats. These battery-powered floats collected sea surface temperature, sea state, and vessel activity data using edge processing, then transmitted observations via satellite to cloud-based analytics platforms. The program demonstrated that distributed, expendable sensor networks could maintain awareness over large ocean areas – each float operating for approximately one year before safely self-scuttling in deep ocean.

NATO’s Allied Command Transformation has explored shared maritime awareness among coalition forces. AI tools assist multinational coordination by automatically correlating vessel observations across national sensor networks, reducing the time required to build a common operational picture from hours of manual cross-referencing to near-real-time fusion.

Future Directions

Future maritime AI development focuses on closing the gap between what satellites can observe and what navies need to understand in contested environments, with research priorities spanning edge processing on unmanned vessels, multi-national sensor fusion, and resilient architectures that maintain effectiveness when adversaries deliberately degrade GPS and satellite communications.

Maritime AI capabilities continue to advance across multiple dimensions. The gap identified by the Paolo et al. Nature study – that up to three-quarters of industrial fishing vessels and nearly a third of transport activity remain invisible to public tracking – defines the central research challenge.

Edge processing represents a key frontier. Current architectures typically relay raw sensor data to centralized cloud processing. Future systems will push AI inference to the sensor platform itself – enabling autonomous USVs and sensor floats to classify vessel contacts locally and transmit only actionable intelligence, reducing bandwidth requirements and latency in communications-denied environments.

Explainable AI addresses operator trust challenges. Black-box detection systems that flag vessels without explaining their reasoning create bottlenecks when operators must decide whether to investigate. Research programs at multiple naval laboratories are developing interpretable models that provide analysts with the reasoning chain behind each alert.

Quantum sensing represents a longer-term research direction. Quantum magnetometers and gravity gradiometers may eventually enable detection of submerged vessels from aircraft or space. While practical maritime deployment remains years away, the Office of Naval Research and DARPA continue investing in quantum sensing research for potential undersea awareness applications.

Conclusion

Maritime domain awareness has become one of the clearest demonstrations of AI’s operational value in defense. The 2024 Nature finding that up to 76 percent of industrial fishing vessels and 30 percent of transport activity escape public tracking systems reveals both the scale of the problem and the opportunity for AI-driven sensor fusion. Satellite constellations, SAR imaging, RF detection, and behavioral analysis now combine to make persistent ocean monitoring technically feasible – but the gap between detection capability and global coverage remains the defining challenge for the next generation of maritime AI systems.

Comparison: Maritime Surveillance Technologies

Technology	Coverage	All-Weather	Dark-Vessel Detection	Primary Limitation
Terrestrial AIS	Coastal only	Yes	No	Line-of-sight range
Satellite AIS	Global	Yes	No	Dependent on transponder
Optical Imagery	Point coverage	No	Limited	Cloud cover
SAR Imagery	Point coverage	Yes	Yes	Revisit rate
RF Detection (HawkEye 360)	Global	Yes	Yes	Signal required
Acoustic Networks	Regional	N/A	Yes	Range limitations
Multi-modal AI Fusion	Integrated	Yes	Yes	Integration complexity

Frequently Asked Questions

What is maritime domain awareness? Maritime domain awareness is the effective understanding of all activities and entities in the maritime environment. The International Maritime Organization defines it as “the effective understanding of anything associated with the maritime domain that could impact the security, safety, economy, or environment.” This includes monitoring vessel traffic, tracking potential threats, identifying illegal activities such as piracy or smuggling, and maintaining situational awareness for operational and strategic decision-making.

How is AI used in maritime domain awareness? AI systems process data from multiple sources including satellite AIS receivers, synthetic aperture radar imagery, optical satellites, and RF detection satellites to identify vessels, predict behavior patterns, detect anomalies, and flag potential threats. Machine learning algorithms can distinguish between fishing techniques based on vessel track patterns and detect when vessels disable their AIS transponders by correlating gaps in AIS data with independent satellite observations.

What is AIS and why does it matter for MDA? The Automatic Identification System is a shipboard broadcast transponder operating in the VHF maritime band. The International Maritime Organization requires AIS on all international-voyage ships above 300 gross tonnage, all cargo ships above 500 gross tonnage, and all passenger ships. AIS broadcasts vessel identity, position, course, and speed, creating the foundational dataset for global maritime surveillance – though the system depends on voluntary compliance, and vessels engaged in illicit activities frequently disable their transponders.

What are the main MDA challenges AI addresses? Key challenges include the sheer volume of maritime traffic to monitor, distinguishing legitimate vessels from those engaged in illegal activities, maintaining awareness across vast ocean areas beyond coastal radar range, and integrating data from multiple heterogeneous sensors into a coherent operational picture. A 2024 Nature study found 72 to 76 percent of industrial fishing vessels are not publicly tracked, underscoring the scale of the surveillance gap.

How does AI detect dark vessels that disable AIS? AI systems combine SAR satellite imagery with AIS data and RF signal detection to identify vessels that have turned off their transponders. When a vessel’s AIS signal disappears, algorithms query SAR and RF satellites to image the expected location. HawkEye 360 operates 36 RF-detection satellites that locate vessels by their radio emissions even without AIS. Deep-learning models trained on vessel radar signatures can match SAR detections to known vessel profiles, sometimes identifying specific ships without transponder data.

What nations lead in maritime AI capabilities? The United States, China, and European nations including the UK and France lead in maritime AI development. The U.S. Navy’s Task Force 59 pioneered unmanned AI-integrated maritime patrols in the Middle East. China has deployed extensive coastal surveillance networks integrating radar, AIS, and satellite data. The European Maritime Safety Agency operates the Copernicus Maritime Surveillance service using AI-enhanced satellite analysis. India approved a $131 million purchase of HawkEye 360 RF detection technology to enhance Indo-Pacific maritime awareness.

Related reading on Defense AI Weekly: Is Singapore Stronger Than Malaysia’s Military?