Detecting illegal shipping activity with AI, machine learning
Marie Donlon | April 08, 2021Researchers at the Technical University of Denmark (DTU) and from the data and analysis company Gatehouse are jointly developing technology for identifying illegal shipping traffic in Arctic waters.
The project, dubbed Dark Ships, seeks to combine data obtained from ships’ Automatic Identification System (AIS) transponders and satellites to determine if a vessel in the enormous Arctic region is there for nefarious reasons — potentially engaged in illegal fishing, illegal weapons trading or other illegal activities.
Source: DTU
According to DTU and Gatehouse, the technology will use artificial intelligence (AI) algorithms and machine learning to cull through AIS transponder data and satellite data to assess the behavior of ships in the Arctic. Alone, satellites monitoring ships in the Arctic region cannot distinguish ships from icebergs. Yet, in combination with AIS transponder data, which reveals ship data such as a ship’s cargo, next port of call or route, the researchers believe it will be possible to better detect suspicious behaviors associated with illegal shipping activity.
“Sometimes ships forget to update the information, so their AIS data does not show the correct route or time of arrival at a port. At the more serious end of the spectrum, there are the ships that turn off the transponder when they enter our waters, to hide their position and route. This can mean that they are engaged in illegal activities, and that the ship should be found and stopped,” explained Henning Heiselberg, head of security DTU.
As such, being able to distinguish ships from other objects like icebergs in the water and not being able to locate that ship's AIS data would likely serve as identification of a “dark,” or illegal, ship.
According to the project's participants, the AI and machine learning algorithms will sift through the vast amounts of data much faster than humans and could even use that data to make predictions about which direction these dark ships are destined.
“Once a dark ship has been identified, it will already have moved. Based on our knowledge of which routes ships usually follow and where they typically sail with a given cargo, we can start predicting where they are going to be if an authority needs to make contact, for example by flying out,” explained Poul Bondo deputy director of Gatehouse Maritime.