ARRIVAL-X is a cloud-based, AI-powered solution designed to enhance operational efficiency, reduce idle waiting times, and optimize resource allocation for ports, shipping companies and logistics providers.
Developed under the European Space Agency’s (ESA) NAVISP program, the new system uses machine learning algorithms integrating automatic identification system (AIS) data, GPS/Galileo positioning data, and weather data from the Copernicus Marine Service, to deliver precise and real-time ETA (estimated time of arrival) predictions.
Readers will know, maritime transport has been and remains the backbone of global trade, enabling of over 80% of the world’s long-distance movement of goods. However, inefficiencies in port operations, unreliable (ETA) predictions, and congestion contribute to delays, cost overruns and higher carbon emissions.
Miftah Zeya, CEO of Munich-based StellarShip Labs UG, also known as SPACE-SHIP, delivered the final presentation of the ARRIVAL-X project at a recent event hosted by ESA. He described the new system as a groundbreaking, real-time, maritime situational awareness and decision support tool that enhances the operational efficiency and global competitiveness of ports. Inaccurate ETAs, he said, continue to be one of the most intractable pain points in global shipping logistics, affecting two-thirds of ships globally.
Testing in a real port scenario
ARRIVAL-X is an AI-driven platform requiring zero integration with port infrastructure. The system was field-tested in December 2024, at the Port of Algeciras, Spain, Europe’s fourth busiest maritime hub.
Two modeling approaches were assessed. The first of these is a global approach that works for all the major seaports across the world. It can be deployed in one hour and is suitable for long-term predictions. The second approach is a local approach, which encompasses a higher level of customization, requiring retraining with local datasets and suitable for last-mile predictions.
Overall, error distribution data revealed very good levels of accuracy within the last six hours prior to arrival, beating conventional prediction methods and far surpassing the project team’s expectations. ARRIVAL-X also includes a system dashboard, comprising a standalone web interface, rendering complex prediction outputs into actionable information, easy to understand and deploy, even for non-expert users.
SPACE-SHIP is already looking towards further refinements to the system, including infrastructure scaling to support multiple users at the same time, user interface and user experience (UI/UX) improvements, and an application programming interface (API) service for broader market reach. A second pilot test campaign is slated for June 2025. This project was co-funded by NAVISP Element 2, which aims to increase competitiveness in the European PNT sector.
