Multi-Band/Multi-Frequency - Based Discrimination Given Minimal Apriori Information on Threat/Radar Geometry

The proposed effort will research the feasibility of S-Band and X-Band multi-frequency discrimination and potentially demonstrate how these features can be used to remove one of the biggest limitations of the current state of the art, i.e. reduction in classifier sensitivity to apriori information. The traditional approach to new discrimination research typically begins with a very narrow set of discrimination conditions (normally from a single or small test data set) and begins to formulate and test discrimination algorithms. This research will attempt to accelerate refinement of the multi-frequency-based discrimination algorithms by basing the algorithm design on tactically realistic conditions from the beginning (e.g. where minimal apriori information is available). The research will determine the feasibility of simultaneous collection of narrowband, frequency-based discrimination features and wideband discrimination features that enhance discrimination performance. Narrowband and wideband features will be collected versus lethal and non-lethal threat objects across a comprehensive set of threat/ship geometry conditions. Finally, the selected set of features will be used to evaluate discrimination performance versus a series of blind tests. If successful, this research will yield new, frequency-based discrimination features, as well as a highly efficient method for their collection.