Artificial Intelligence based Virtual Control Room for the Arctic

Modern maritime navigational, surveillance, and communications systems, combined with today’s network technologies and data storage and sharing capabilities have created enormous amounts of available data collected from vessels operating worldwide. Currently, however, all of this available data does not necessarily reflect reality or help maritime operations. In fact, it presents two core problems.

The first problem is that due to these vast amounts of data, threats and other significant maritime events of interest are often lost and hidden in regular traffic patterns. This presents a need to develop an intelligent data processing functionality, which highlights unusual and inconsistent behavior (also called an anomaly) within this large set of data. By creating these functionalities and increasing automation at higher fusion levels, this would support maritime operators who have limited resources by giving them access to information such as early warnings about approaching threats.

The second problem is that due to these advancements in network technology and storage capacities an enor-mous amount of data is now available to the various maritime user communities. All of this data has the potential to create a comprehensive knowledge base from which to operate. However, without intelligent fil-tering of data, there is a risk of information overload, the operational picture getting over complicated and decision-making efforts delayed or impaired. The consequences are potentially dramatic in terms of accidents, pollutions, border infringements and criminal activities and are possibly further aggravated in areas so remote as the High North Atlantic, Arctic and the upper High Norths which are not monitored as closely as other European waters. This presents an urgent need to create an intelligent filtering and retrieval mechanism to navigate this dataset.

In AI-ARC, RISE developed intention recognition methods that help maritime surveillance operators find vessels that perform illegal actions. A demonstration was performed in Karlskrona in September 2023 in cooperation with the Swedish Coast Guard as well as other project partners, in which several use cases were presented and evaluated

• A method for early warning of groundings
• Methods for detecting possible environmental violations, including finding indications of possible perpetrators
• Methods for detecting illegal fishing
• Methods for detecting "dark vessels", ships that do not emit AIS signals
• Methods for detecting suspicious movement patterns that, combined with other information, could indicate attempts at sabotaging underwater infrastructure or smuggling.