While AIS data has many legitimate uses, it can be misused for illicit activities in many contexts, including human trafficking and sanctions evasion.
Since 2004, automatic identification systems (AIS) have been industry standard for vessels over 300 tonnes undertaking international voyages, allowing for tracking of vessel trajectories. These datasets are used not only for safety purposes, but also commercial purposes, such as trade analyses and analyses of ship and port performance.
With AIS data transmitted every 10 seconds, a single ship can generate 3 million records across a year. Such huge volumes of AIS data create challenges in data analysis. They also present an opportunity for open-source intelligence, given that the AIS data is publicly available and communicates more than just a vessels location, providing information such as physical dimensions, name, and appearance.
In partnership with the Royal United Services Institute (RUSI), the Institute for Security Science and Technology at Imperial College London hosted an event on the current state of AIS data and the implications of recent developments on security and defence.
Distributed Processing of Maritime Trajectory Data
Guang Yang is a PhD student in the Department of Computing at Imperial College London, focusing on databases, big data management, distributed systems, and the application of big data to traffic and maritime sectors.
For AIS data to provide meaningful information on the transit of vessels, the trajectories of individual vessels over time and space need to be constructed accurately over long distances. This is challenging given the high volume of AIS data produced for each vessel, the high number of vessels, and the high resolution of AIS data due to it being generated in near real-time.
Given these challenges, analysis of a single vessel’s movements often produces multiple trajectories that subsequently need to be joined into a single, continuous trajectory. The identification of individual trajectories as belonging to a single vessel and their joining is a demanding process.
An efficient management and processing system would aid processing of maritime trajectory data and increase the utility of AIS data for various analyses, including those relevant to open-source intelligence.
Guang has been building a distributed system optimised for trajectory datasets, with faster processing and features specific to trajectories considered. His current approach is based on the combing of AIS datasets for trajectories that are close in space and time. The conditions under which such trajectories can be considered similar enough in space and time to be joined are defined, and trajectories are joined. The resulting system has been shown to accelerate the trajectory join process by 43x compared to other state of the art methods.
Guang plans on further refining this infrastructure through incorporation of other data sources, such as satellite imagery. These additional sources would provide details about vessel trajectories that would allow joins based on novel variables, such as relative speed. He also plans on implementing methods by which special cases of interest can be identified through events such as sudden directional changes, which in the context of open-source intelligence would facilitate identification of suspicious vessels.
AIS Manipulation and Open-Source Intelligence
While AIS data has many legitimate uses, it can be misused for illicit activities in many contexts, including human trafficking, sanctions evasion, and smuggling.
Maritime Mobile Service Identity (MMSI) and International Maritime Organisation (IMO) numbers act as unique identifiers for a vessel and its accompanying AIS data. With AIS having been an industry standard for nearly two decades, illicit actors cannot rely on ‘going dark’; a vessel not transmitting AIS data would soon be flagged as suspicious.
Rather, AIS falsification allows an illegitimate vessel to use the MMSI or IMO number of a legitimate vessel as cover. This hinders detection of illegitimate vessels and has notably been used by the North Korean state to break sanctions.
Detection of illegitimate vessels utilising this method of AIS falsification is still possible through careful analysis of AIS data, a strength of open-source intelligence. If the legitimate and illegitimate vessels are separated over a great distance for instance, the AIS data associated with a MMSI number would show improbable distances travelled in short periods of time.
A stark example of the risks that AIS data manipulation pose for security and defence is HMS Defender, a Royal Navy destroyer. On June 19th, AIS data indicated that HMS Defender was sailing into the contested waters of Sevastopol port in Crimea, while webcam footage showed that it was actually located in the port of Odessa. With the responsible parties still unknown, their motivations can only be speculated upon.
An incident may have been avoided in this instance, but it is clear the consequences of future AIS data manipulation could be severe, and work must be done to keep abreast of recent developments.
Article text (excluding photos or graphics) © Imperial College London.
Photos and graphics subject to third party copyright used with permission or © Imperial College London.
Institute for Security Science & Technology