Abstract: A system for analyzing a plurality of channels of data received from a sensor array. The system includes a data acquisition system that receives and independently processes each channel. A low-level processing section receives each channel of processed data and identifies signals of interest in one channel. Signals of interest are stored in an event database. A high-level processing section analyzes data occurring over a preset duration of time and across multiple channels of data and communicates with an operator machine interface. The operator machine interface provides analysis to an operator. Further aspects of the system characterize the data in order to indicate the data source and alert the operator to signals having certain predefined characteristics.

Abstract: A system to extend transducer arrays from the hull of a vessel while it is underway provides for increased array gain, volumetric aperture (very large aperture arrays), and rejection of hull-bound noise for submarines, surface ships, weapons, and unmanned underwater and surface vehicles. A plurality of arms extend radially outwardly from the hull of a submarine, and passive and/or active elements of the arrays are mounted around the hull on the arms, trail in the water from the arms, or extend between different ones of the arms or between arms and the hull. The system of arms and the array are folded into elongate recesses in the hull to lower noise during higher speed transit by the submarine. The arrays can be made up of sensors and/or projectors of radar, sonar signals, optical energy, vibrational energy, magnetic influence, temperature, etc., and combinations of these sensors.

Abstract: A system and method for ranking features by exploiting their relationship the Fisher projection space. The system ranks n features in a feature set using a design set comprising exemplars from each of M possible event classes of an associated feature-based classification system. A training set is created by randomly selecting exemplars from each of the M classes in the design set. A "smoothed" Fisher projection space for the training set is created by replacing the sample means and the within-class sample covariance matrix normally used in deriving a Fisher projection space with expressions for the mean vectors and covariance matrices derived from event class probability density function estimates. The angle between a given feature and the smoothed Fisher projection space is calculated for each feature in the feature set, and the features are then ordered by increasing numerical size of this angle.