Multiple sensor fusion for effective abnormal behaviour detection in counter-piracy operations

We propose a two-stage method based on multiple distributed sensors for detection of piracy operations at sea. The proposed method is based on fusion of evidence from radar and optical sensors as well as AIS signals. The sensors are land based, air based and space based. In the first stage, a number of detectors perform recognition of low-level events, such as vessel speed, direction, size and identification. Their outputs act as input to the second stage, which performs fusion and disambiguation of the first-stage detections taking into account multiple sensor fusion, multi temporal fusion and distributed fusion. The result should be effective abnormal behaviour detection, by which we mean a method that is able to detect piracy operation at an early stage, i.e. close to the time, or possibly before, the attack has occurred. A numerical example with a pirate attack is presented and discussed.