Automatic Multi-sensor Data Fusion Processing Using Fuzzy Logic
BJ Jarvis, DJ Kewley · 1994
One aspect of system performance assessment of an Over-The-Horizon-Radar (OTHR) requires a meaningful comparison of radar tracking data (geographic, kinematic, etc) against external ground-truth data measuring these same quantities. These external data may originate from multiple sources each exhibiting unique characteristics even though they may describe a common target. To perform a comparison, the ground-truth data must be associated with the corresponding radar data. In general, the amount of data generated by these systems precludes a manual association of truth and radar data: an automated method is required. This paper successfully demonstrates the application of an artificial intelligence technique employing fuzzy logic to this problem, using real tracking data obtained from Australia's Jindalee OTHR. The method automatically selects radar data corresponding to available ground-truth data from an arbitrary number of independent sensors. The technique provides an aid to fast and efficient evaluation of detection, tracking, coordinate registration and operational performance of a radar. The principle can easily be extended to a real-time environment.