Periodic Reporting for period 1 - SMAUG (Smart Maritime and Underwater Guardian)
Período documentado: 2023-10-01 hasta 2025-03-31
The illegal trafficking of drugs and goods is a global problem, with an increasing proportion of it taking place in the maritime domain. The development and use of new transport methods, such as submersible and semi-submersible vessels, is becoming increasingly common. This poses new challenges to law enforcement as current detection methods are often ineffective. Furthermore, the ever-changing nature of cargo, coupled with advancements in smuggling and cartel transport methods, necessitates that law enforcement agencies continually update and enhance their technology to prevent illegal activities. This challenge is further compounded by the fact that the movement of illegal substances is a large-scale, heavily funded operation.
The mission of the SMAUG project is to improve and strengthen the security of ports and their gateways by harnessing the power of artificial intelligence (AI) through an integrated system capable of providing threat detection and analysis data between three main elements: port security infrastructure, advanced underwater detection systems, and surveillance vessels.
The four primary methods used for underwater detection and location are:
i) Acoustic detection using hydrophones to listen out for sounds emitted by small underwater vehicles.
ii) Rapid sonar hull scanning, which is used to scan ship's hulls and perform harbour floor scanning;
iii) High-resolution sonar inspection is used to inspect objects in water with poor visibility.
(iv) Collective autonomous location: a swarm of autonomous underwater vehicles providing information to artificial intelligence modules.
Combining these sources with port current information enables SMAUG to suggest solutions for detecting potential threats to infrastructure or vessels and identifying those carrying concealed goods.
The system will increase the response capabilities of ports providing: 1) a CISE Adaptor, 2) detection and identification of underwater anomalies, 3) autonomous inspection, 4) AI security improvement and 5) a command and control response capacity increment.
The mission of the SMAUG project is to improve and strengthen the security of ports and their gateways by harnessing the power of artificial intelligence (AI) through an integrated system capable of providing threat detection and analysis data between three main elements: port security infrastructure, advanced underwater detection systems, and surveillance vessels.
The four primary methods used for underwater detection and location are:
i) Acoustic detection using hydrophones to listen out for sounds emitted by small underwater vehicles.
ii) Rapid sonar hull scanning, which is used to scan ship's hulls and perform harbour floor scanning;
iii) High-resolution sonar inspection is used to inspect objects in water with poor visibility.
(iv) Collective autonomous location: a swarm of autonomous underwater vehicles providing information to artificial intelligence modules.
Combining these sources with port current information enables SMAUG to suggest solutions for detecting potential threats to infrastructure or vessels and identifying those carrying concealed goods.
The system will increase the response capabilities of ports providing: 1) a CISE Adaptor, 2) detection and identification of underwater anomalies, 3) autonomous inspection, 4) AI security improvement and 5) a command and control response capacity increment.
Current maritime management and seaport security systems have been studied in detail, including a state of the art analysis, the identification of potential users and stakeholders in the maritime domain and a description of their needs and requirements based on the responses received to a created questionnaire distributed among the project’s end users. Moreover, existing gaps and opportunities were mapped with the SMAUG objectives and potential capabilities to be added by SMAUG to existing systems were identified.
Regarding the detection systems, notable strides in integrating multiple advanced technologies into a cohesive monitoring system have been made. The integration of acoustic, sonar, aerial, and swarm-based detection methods have been progressively translated into functioning prototypes, field-tested hardware, and integrated software platforms.
From acoustic detection to robotic swarm intelligence, each subsystem has moved from conceptual design into functional prototypes tested in operational environments. With successful system integration and early field validation, the foundation is now firmly laid for the next phase of development, which will focus on increasing robustness, optimizing data processing, and conducting coordinated multi-system field trials.
Vessel features such as Docking Point Latch and Release, Surface and Underwater Vessel Operation, Energy Sources and Navigation have been built, evaluated and tested.
Software development has been done for navigation and control of the USVs (Unmanned Surface Vehicle), the UUV (Unmanned Underwater Vehicle) and the UAV (Unmanned Aerial Vehicle) of the project. Preparation to integrate the UAV on to a USV has been planned and prepared and is on track to be completed successfully.
The main architecture for the central command portal is ongoing, and planning for the demonstrator versions of this portal is also ongoing.
For the information exchange within the ecosystem, 2 AIS stations have been developed: The interface that collects all the data from the USVs /ROV (Remotely Operated Vehicle) and the dispatcher, and the interface between the dispatcher and the CISE node that will be used for the demonstration.
AIS and Sentinel satellites datasets have been created and prepared for integration into the SMAUG architecture supporting the integration of positioning and situational awareness capabilities relevant for maritime security.
An AIS coastal station has been developed to receive and analyse AIS data, alongside a dispatcher system that collects and displays data from all relevant sensors.
Data uplink functionality has been fully defined for 6 out of 9 detection devices. Partial testing has been completed for 4 devices, and full integration has been achieved for 1 device, which now serves as a reference model for partner implementation.
AI for security has progressed across several fronts, with significant developments in AI-based threat detection, model lifecycle management, and privacy-aware data handling:
• Anonymization strategies have been implemented and tested.
• Classification of underwater acoustic signals to identify vessels in real-time validated in realistic and noisy port conditions. Integration with AIS data under development to further strengthen the decision-making process.
• Development of models to identify vessels, people, and relevant objects in drone-captured images of maritime environments has been made and tested to Search and Rescue scenarios.
• Sensor-based Alert Generation Using Machine Learning, converting raw data streams into actionable alerts. Two data sources—AIS and sonar—were selected and are currently being fine-tuned and prepared for integration into the main alerting infrastructure.
• Advanced architecture for orchestrating responses to detected maritime threats has been developed. Module in a proof-of-concept stage. A virtual port environment has been developed to simulate realistic conditions.
• Creation of a model monitoring module that enables continuous tracking of model behavior for early detection of anomalies.
Use case scenarios have been defined, and testing has been planned and started across the participating ports.
Regarding the detection systems, notable strides in integrating multiple advanced technologies into a cohesive monitoring system have been made. The integration of acoustic, sonar, aerial, and swarm-based detection methods have been progressively translated into functioning prototypes, field-tested hardware, and integrated software platforms.
From acoustic detection to robotic swarm intelligence, each subsystem has moved from conceptual design into functional prototypes tested in operational environments. With successful system integration and early field validation, the foundation is now firmly laid for the next phase of development, which will focus on increasing robustness, optimizing data processing, and conducting coordinated multi-system field trials.
Vessel features such as Docking Point Latch and Release, Surface and Underwater Vessel Operation, Energy Sources and Navigation have been built, evaluated and tested.
Software development has been done for navigation and control of the USVs (Unmanned Surface Vehicle), the UUV (Unmanned Underwater Vehicle) and the UAV (Unmanned Aerial Vehicle) of the project. Preparation to integrate the UAV on to a USV has been planned and prepared and is on track to be completed successfully.
The main architecture for the central command portal is ongoing, and planning for the demonstrator versions of this portal is also ongoing.
For the information exchange within the ecosystem, 2 AIS stations have been developed: The interface that collects all the data from the USVs /ROV (Remotely Operated Vehicle) and the dispatcher, and the interface between the dispatcher and the CISE node that will be used for the demonstration.
AIS and Sentinel satellites datasets have been created and prepared for integration into the SMAUG architecture supporting the integration of positioning and situational awareness capabilities relevant for maritime security.
An AIS coastal station has been developed to receive and analyse AIS data, alongside a dispatcher system that collects and displays data from all relevant sensors.
Data uplink functionality has been fully defined for 6 out of 9 detection devices. Partial testing has been completed for 4 devices, and full integration has been achieved for 1 device, which now serves as a reference model for partner implementation.
AI for security has progressed across several fronts, with significant developments in AI-based threat detection, model lifecycle management, and privacy-aware data handling:
• Anonymization strategies have been implemented and tested.
• Classification of underwater acoustic signals to identify vessels in real-time validated in realistic and noisy port conditions. Integration with AIS data under development to further strengthen the decision-making process.
• Development of models to identify vessels, people, and relevant objects in drone-captured images of maritime environments has been made and tested to Search and Rescue scenarios.
• Sensor-based Alert Generation Using Machine Learning, converting raw data streams into actionable alerts. Two data sources—AIS and sonar—were selected and are currently being fine-tuned and prepared for integration into the main alerting infrastructure.
• Advanced architecture for orchestrating responses to detected maritime threats has been developed. Module in a proof-of-concept stage. A virtual port environment has been developed to simulate realistic conditions.
• Creation of a model monitoring module that enables continuous tracking of model behavior for early detection of anomalies.
Use case scenarios have been defined, and testing has been planned and started across the participating ports.
Two main Expected Outcomes are foreseen at the end of the project:
• Improved security of maritime infrastructures and maritime transport, including sea harbours and their entrance route.
SMAUG will integrate information from different sources to create a system that will help operators to take action in the event of an unexpected event. SMAUG will add a decision support system in the port systems, keeping the operator always in the loop.
• Improved detection of illicit and dangerous goods and/or of threats hidden below the water surface, either threatening infrastructures or vessels, or moving alone or connected to vessels.
By integrating the scans and analysis from different types of sensors SMAUG will be able to detect an ever increasing quantity of illegal goods or threats entering the port, specifically underwater. Once any of the sensors detect something suspicious, SMAUG will provide the ability to send other types of UUV with other capabilities to
investigate.
• Improved security of maritime infrastructures and maritime transport, including sea harbours and their entrance route.
SMAUG will integrate information from different sources to create a system that will help operators to take action in the event of an unexpected event. SMAUG will add a decision support system in the port systems, keeping the operator always in the loop.
• Improved detection of illicit and dangerous goods and/or of threats hidden below the water surface, either threatening infrastructures or vessels, or moving alone or connected to vessels.
By integrating the scans and analysis from different types of sensors SMAUG will be able to detect an ever increasing quantity of illegal goods or threats entering the port, specifically underwater. Once any of the sensors detect something suspicious, SMAUG will provide the ability to send other types of UUV with other capabilities to
investigate.