Improved echo classification in shallow water

This paper describes a classification approach designed to automate operations and reduce false alarms for active sonars in shallow water environments. There are two main innovations that characterize the approach. The first innovation is to break down the classifier into four sub-classifiers, each targeting a different target characteristic. These characteristics are: (a) echo spatial and temporal shape, (b) echo amplitude, (c) echo tracking consistency, and (d) Doppler. The classifier forms a likelihood statistic for each characteristic which is represented in the display symbols. The second innovation is the use of parameters for conditioning the distributions which sets the stage for environmental adaptability. The paper stresses system design and not performance.