Bergman had previously found false tracks on the AIS website. Last year, a virtual yacht from an online sailing racing game appeared on AISHub. But this was the first time he saw a real ship being impersonated, and the battleship was not inferior.
“At SkyTruth, we pay special attention to the ways in which fake data has affected the fishery,” Bergman said in a video telephone interview. “But we want to get an overview of how the data is falsified and what we can do to detect and correct it.”
Bergman identified nine battleships from the screenshots in the story, and then compared their fake AIS information with the real information broadcast by the same ships before and after the impostor. He immediately noticed that these were not amateur pranks or accidents. “These false information is very credible, but we have been confirmed by the Swedish Navy that these positions are false,” he said.
There are more than 20 types of AIS messages—some for supertankers and others for cruise ships—each containing multiple data fields, covering everything from navigation information to mysterious communication settings. By carefully comparing the areas that sailors usually don’t see, Bergman finally discovered the subtle difference between fake and real data. Then, he used the model to write queries for the global historical database of AIS messages, and the results shocked him.
His search found nearly a hundred sets of messages from multiple AIS data providers, which date back to September last year and span thousands of miles. What is more worrying is that the affected ships are almost entirely warships from European and NATO countries, including at least two US nuclear submarines.
“I was shocked when I realized that there are many other ships showing this unusual AIS profile,” Bergman said. But he needs to know that the suspicious AIS message is actually fake, not the result of a technical failure or a special military environment. In the following months, Bergman laboriously verified the actual location of the target ship. At first, he used open source data, including news reports, military press releases, and enthusiast sites such as Warshipcam.com. “A lot of people like to take photos of naval ships and post them on the Internet,” Bergman said. “I found examples of ships leaving or entering seemingly impossible areas.”
Bergman then superimposed the European Space Agency’s synthetic aperture radar and optical images. Sentry 1 and –2 Satellite to suspicious AIS ping. If they are real, the AIS data should be perfectly matched under the satellite image of the ship.
On the contrary, Bergman only saw the empty ocean time and time again. In fact, he said, “I haven’t found instances where tracks that were flagged as false by the query proved to be real.”
Stars are not always consistent with Bergman’s detective work. Some AIS trajectories are inconsistent with the overhead ESA satellites, or appear on cloudy days when optical images are useless. And some battleships do not have many Instagram fans. In the end, Bergman managed to confirm that about 15 sets of AIS data were absolutely false.
as well as Queen ElizabethIn h’s fictional fleet, Bergman found the false tracks of American, Dutch, Belgian, German, Lithuanian, Estonia, and other Swedish warships. A previously unreported suspicious trajectory shows the US guided missile destroyer USS Roosevelt In November last year, it sailed into the 4 kilometers of territorial waters around Kaliningrad, Russia. If this is true, it would be a reckless provocation. In June, there appeared to be five false invasions near Kaliningrad. One of them involved a Polish warship whose track, speed and heading were exactly the same as the Swedish frigate five days ago, and Bergman’s other sign indicated that the track was digitally generated.