Fraunhofer Institute for High-Speed Dynamics, Ernst-Mach-Institut, EMIAutomatic hazard detection by analyzing large amounts of data.
The AI-ARC (Artificial Intelligence Based Virtual Control Room for the Arctic) project aimed to improve situational awareness in civil shipping and reduce safety risks – without increasing the workload of seafarers. Partners from 13 European countries were involved.
A key challenge was to efficiently process the AIS (Automatic Identification System) data from over 1500 ships as well as extensive weather information and satellite images on a daily basis. To achieve this, the partners developed innovative tools and services in which artificial intelligence (AI) plays a key role. In the AI-ARC project, AI automatically recognized suspicious activities and unusual occurrences.
These findings were incorporated into a new type of platform for situation visualization. Particular attention was paid to the reliable identification of icebergs for safe navigation in the Arctic. Fraunhofer EMI developed an advanced method for estimating forecast uncertainty that goes beyond conventional approaches. Unlike previous methods, which often require a change in the AI model architecture, Fraunhofer EMI uses an additional AI model to quantify the reliability of iceberg detection. This model-agnostic approach enables quick and easy implementation with comparable results, setting it apart from existing solutions. To detect illegal fishing and smuggling, the researchers analyzed ship routes for anomalies. Reliability assessments were carried out at Fraunhofer EMI to inform users about the background to an alarm and reduce false alarms. Thanks to these advanced approaches, the AI-ARC project is making a significant contribution to the safety of shipping in the Arctic.
