Aspects of systems engineering for a matched-field processor

Matched-field processing (MFP) offers improvements over more conventional plane wave processing techniques for source detection and localization. However, MFP requires an adequate knowledge of the geoacoustic environment and comes at the expense of greatly increased algorithm complexity and computational requirements. A realistic MFP system will require billions of arithmetic operations per second and billions of bytes of data storage to provide the sonar operator with accurate and current information about ocean traffic. The authors have developed a parallel processor architecture which is capable of solving a baseline MFP real-time processing scenario. An MFP system also requires three dimensional visualization of system data objects. They show that effective presentation of this data can enhance the interpretive capabilities of the sonar operator. Data fusion techniques are combined with advanced visualizations to facilitate the detection, localization and tracking of an acoustic target in three dimensions