Discrimination Planning Technique Development, Assessment, and Proof Of Principle Demonstration

Period of Performance: 02/28/2006 - 08/29/2006

$98.9K

Phase 1 SBIR

Recipient Firm

Magnacom, Inc.
615 Discovery DriveSuite B
Huntsville, AL 35806
Principal Investigator

Abstract

The purpose of this task is to demonstrate proof of principle software capable of selecting a sensor from a suite of sensor realizations that is best suited to provide critical discrimination information for a single pre-defined threat engagement. A sensor realization may be a specific sensor site for operational area planning or a specific sensor type for C2BMC inline discrimination sensor selection. Threat signature information, threat trajectory information, and discrimination timeline data are all inputs to the optimization process. The offline and inline planning software samples the signature data as presented to each sensor and sensor site by the threat object and collects both medium band and/or wideband feature measurements. A separation analysis will then performed on the computed (and stored) feature measurements to estimate what sensors or sensor sites provide promising discrimination performance for the proposed discrimination timeline. The available sensors or sensor sites will then be ranked according to their efficacy for discrimination support using several specifically defined measures of effectiveness (MOEs) that operate on the feature separation analysis results. The suitability of the MOEs will also be addressed relative to execution speed and their ability to predict actual classifier behavior.