imPROved Maritime awarENess by means of AI and BD mEthods

Informacje na temat projektu

PROMENADE

Identyfikator umowy o grant: 101021673

DOI

10.3030/101021673

Projekt został zamknięty

Data podpisania przez KE 26 Kwietnia 2021

Data rozpoczęcia 1 Października 2021

Data zakończenia 31 Marca 2023

Finansowanie w ramach

• Secure societies - Protecting freedom and security of Europe and its citizens

Koszt całkowity

€ 6 065 823,75

Wkład UE

€ 4 997 555,63

Koordynowany przez

YPOURGEIO NAFTILIAS KAI NISIOTIKIS POLITIKIS
Greece

Periodic Reporting for period 1 - PROMENADE (imPROved Maritime awarENess by means of AI and BD mEthods)

Okres sprawozdawczy: 2021-10-01 do 2023-03-31

Podsumowanie kontekstu i ogólnych celów projektu

Security risks and threats in the maritime domain have become more complex and intrinsically linked at the global level. European waters are navigated daily by some 12,000 vessels which share their positions to avoid collisions, generating a huge number of positional messages every month, besides a dramatic increase in irregular migration flows and human trafficking and smuggling, particularly through the Mediterranean Sea. There is an urgent need to increase Maritime Situational Awareness exploiting the overabundance of information already provided by sensors, platforms and legacy systems as well as the new and emerging data processing technologies to enhance the level of knowledge and allowing the operator to make timely and informed decisions. Then the challenge is twofold: on one side encourage the exchange of heterogeneous data among administration valorising the CISE network currently in place, on the other exploit at the best these datasets by means of automated processing in a way to enhance the Maritime Situation Awareness.

The objectives of PROMENADE project are:
❖ Improve solutions for the vessel tracking, behaviour analysis and automatic anomaly detection by means of the application Artificial Intelligence and Big Data technologies.
❖ Promote collaborative exchange of information between maritime surveillance authorities.
❖ Shorten the time to market and assure the compliance with legal and ethical regulations and norms
❖ Leverage and go beyond of the research undertaken in previous projects.

Prace wykonane od początku projektu do końca okresu sprawozdawczego oraz najważniejsze dotychczasowe rezultaty

The goal of PROMENADE project is to improve solutions for the vessel tracking, behaviour analysis and automatic anomaly detection by means of Artificial Intelligence (AI) and Big Data (BD) technologies application.
At this end PROMENADE developed a set of AI/BD based services for improved Maritime Situation Awareness (MSA) grouped in 5 main categories:
• Classification: PROMENADE's classification services provide automated data processing and classification for various sources, including cameras, satellite imagery, and vessel tracking stations applied in the maritime domain, related to vessel detection, classification, route classification, activity classification and oil spill detections.
• Pattern Detection; PROMENADE’s pattern detection services provide automatic pattern detection applied to multi-source data fusion, the extraction of Pattern of Life, behaviour analysis and anomaly detection, AIS/Satellite analysis
• Risk Assessment: PROMENADE’s Risk Assessment services provide automated data processing and risk assessment of vessels’ behavior and characteristics, though use of innovative data sources and algorithms.
• Future State Prediction: PROMENADE's future state prediction services allow to learn the motion of ships in particular areas of interest from historical data with the goal of predicting their future trajectories and anticipating their future behavior.
• Data Infrastructure which includes the data lake that provides for the ingestion, storage, processing, and distribution of data in a Big Data environment and the data exchange with the external CISE Network.

Innowacyjność oraz oczekiwany potencjalny wpływ (w tym dotychczasowe znaczenie społeczno-gospodarcze i szersze implikacje społeczne projektu)

Innovative Aspects
• Vessel detectors combined with classification, tracking and vessel visual identification algorithms, the inclusion of Artificial Intelligence models and the satellite missions which provide a dense image archive will improve the classical vessel detection.
• Cross-domain data fusion is enabled through Big Data, combining information from heterogeneous sources, including non-structured ones.
• The architecture of the Big Data System is innovative and specifically tailored for maritime data driven activities. Also, a microservices based paradigm will be the basis of applications development and design, according to current up to date best practices. This provides modularity and scalability of the innovative solution while making the system resilient to obsolescence and ready for technological upgrades.
• Vessel prediction improvements by putting together statistical signal processing, automated reasoning, and artificial intelligence techniques.
• Extend behaviour analysis algorithms to provide new functionalities and to utilize the state of the art in order to exploit additional information sources/databases and derive patterns based on expert knowledge.
• Machine learning can support the classification of maritime activities in real-time to deliver a reasoned understanding of shipping behaviour as it evolves and drive evidence-based decision-making.
• Innovation in risk assessment complementing the state of the art with an additional layer of anomalies, namely red flags in the characteristics of ship-owner companies and related individuals/entities.
• The use of a High-Performance Computer for training AI algorithms with Big Data, with an expected Peak power of more than 5 Petaflops and a state-of-the-art parallel file system to access more than 10 Petabytes of disk space.

Additionally, it is worth mentioning the use of different architectures deployed for each environment during the different phases of the project according to the specific purpose:
• Architecture for Training on the HPC: The PROMENADE toolkit will be compliant with this architecture when deployed on the HPC with the purpose of training algorithms which use artificial intelligence. The innovative use of advanced HPC allows Technical Partners to inject a huge amount of historical data in the Data Lake to make the results of these algorithms more accurate. This architecture will be used before the operational phase.
• Architecture for operational purposes: The PROMENADE toolkit will be compliant with this architecture in the operational phase when it is deployed in the real nodes of each country. In this case, the architecture will relate to real-time data coming from external sources and the AI services included in this architecture are already trained.
• The use of different architectures for each specific purpose or mission improves the obtained results and increases the performance of the toolkit.

PROMENADE architecture