To be able to see if anything deviates from the normal in this flood of data, to find patterns indicating that something unwanted is about to happen. That was the aim of the FMV project concerning sea surveillance.
Terrorism and other situations of crisis can show unusual patterns of behaviour. The ability to recognise these signs of potential threats is of great importance. FMV has worked on these questions within the framework of a research and technology development program.
Ingemar Kjölander, Project Manager, tells that the purpose of the program was to develop and tune an experimental system to be used for sea surveillance, SDS, (Situation Detection Service).
The basis of this system for compilation of naval traffic was developed in earlier research programmes. Now Ingemar Kjölanderhe task was to further improve the algorithms.
The SDS-system gets its data from radar, cameras and transponders in the vessels. The system is learning what can be considered a “normal state” by recording all ship movements during a certain period. Vessels moving outside of this pattern are detected.
The recorded normal states are one side of the system, the other is a so called Rule-editor. In this it can be decided what is to be considered a deviation. It can be too high or low speed or that a vessel is inside or outside of specific areas such as established naval routes.
Is there a need to have the ability to find these deviations? Yes, both the Swedish armed forces and the civilian rescue services are asking for a tool of this kind that can distinguish deviations from the normal. The aim is, of course, to identify possible threats as early as possible in order to eliminate them.
Unlawful fishing and pirate activities can be a cause of unusual behaviour on the water. It can also be some problem with the ship or the crew. These are things that the sea surveillance authorities want to know in order to take the right actions.
By applying more advanced rules of deviation there could for example be an alarm when two ships from different harbours meet at sea and then return to their respective harbours. This could be a sign of smuggling activities.
After display and demonstrations at the armed forces development centre in Enköping the programme manager Ingemar Kjölander concluded that the experimental system, SDS, delivered according to expectations.
– In the first version there were too many alarms. When we introduced adjustable parameters for properties and behaviour of the ships the number of alarms went down at the same time the operators found it easier to see the cause of the alarm.
The conclusion from the research and technology programme, SDS, was that the algorithms worked and that it is possible to build systems to detect deviations. This is also confirmed by that the delivering company SAAB already has implemented the algorithms in new products.
Apart from the armed forces and the rescue service the coast guard, police and custom authorities are potential users of this system.