Planned intervention: On Wednesday June 26th 05:30 UTC Zenodo will be unavailable for 10-20 minutes to perform a storage cluster upgrade.
Published June 29, 2020 | Version v1
Project deliverable Open

D3.2 Maritime Use Case: Initial Evaluation and Validation Report

Description

This report summarises the results of the preliminary evaluation of the components developed for the Maritime Use

Case that will be integrated and tested in a final experimentation at sea toward the end of the Project INFORE.

The Maritime Use Case consists in developing and deploying a hybrid sensor network for ship target detection,

localization and classification. Sensor data for the global scale and the local scale surveillance scenario, resulting in

an extreme scale data stream, will be integrated in a maritime situation awareness platform to perform high level

fusion tasks and target behavioural analysis in order to detect behavioural anomalies of vessels with respect to the

normal traffic in the area of interest. The developed platform will be integrated/using the INFORE architecture,

utilizing its capabilities to analyse extreme scale streamed data.

The evaluation is performed using a set of key performance indices (KPIs) defined for the single components of the

system and for the system as a whole, including human factors. The data used in the evaluation process are historical

data available at the beginning of the Project INFORE or generated by simulating surveillance scenarios.

The tested components include i) the CMRE coordinated sea surface robotic sensor network for target detection and

localization by acoustic data, the target detection and classification systems for satellite data and for thermal and RGB

video camera data, and the MarineTraffic situation awareness platform algorithms for complex event detection.

The results of the evaluation are in general in line with the expectations. The coordinated robotic sensor network

outperforms the base network without coordination, the classification from satellite data and camera data performs

with classification accuracy around or greater than 90% and around 80%, respectively. The complex event classifier

system has analogous performance with accuracy achieving, in some cases, values greater than 90%.

The report provides a complete description of each component, including theoretical details, the evaluation of the

components and finally a preliminary plan of the final experiment at sea to evaluate the Maritime Use Case as a whole.

Future work before the final experiment includes implementation of the control and cooperation software on board

the surface robot vehicles and further test of the classification algorithms for satellite and camera data and for complex

events. Furthermore, the detailed plan, including logistic aspects, of the final experiment at sea will be started soon

this year and finalized next year.

Files

D3.2 Maritime Use Case Requirements, Scenario Definitions and Initial Evaluation Report.pdf

Additional details

Funding

INFORE – Interactive Extreme-Scale Analytics and Forecasting 825070
European Commission