AIR ASW ACOUSTIC CLARIFICATION

The ASW threat in the current, post-cold-war environment is the slow-moving, low-target strength, quiet diesel submarine. The shallow water environments they typically operate in make the task of detection even more difficult. These conditions will require the use of an active sonar with a classifier architecture that can operate robustly in a variety of harsh environments. ORINCON proposes to develop both classification feature algorithms and feature vector classifier architectures. Features that are useful in both monostatic and multistatic scenarios will be developed. Classical feature vector classifiers as well as neural network classifiers will be examined. We have specific experience using neural networks to classify active returns recorded with the BSY-1 system and the Low Frequency Active (LFA) system. Automatic classification systems of this type provide 360-degree classification coverage, assistance to inexperienced operators, and the ability to "dig" targets out of clutter and reverberation. ORINCON proposes to identify techniques for active classification and define promising active classification algorithms. We will identify existing recorded data collected during Fleet operations for algorithm testing. A general method to prepare simulated recordings of active returns in a realistic background is also presented and a plan to solidify it is proposed.