Automated Association of Track Information from Sensor Sources with Non-Sensor Information in the Context of Maritime Surveillance

Maritime surveillance of coastal regions requires the processing of a large amount of data. Therefore, limited availability of human or material resources may prevent full use of information and knowledge. A software prototype that incorporates automated association of sensor information with non-sensor information has been developed. Sensor information, which provides contact reports, is modelled as tracks while nonsensor information, or user's knowledge that represents arbitrary ship behaviour, is modelled as templates. For the fusion process elements of fuzzy logic are introduced through membership functions to take into account uncertainties from track and template information. These membership functions are used to evaluate the possibility of track and template association. The knowledge-based system used for this level 2 data fusion incorporates a genetic algorithm for the track modelling and Boolean combination of parent templates for the modelling of user's knowledge. This paper also discusses database management and user interface