Advanced Feature Extraction Algorithm (AFEA)

Defense against modern ballistic missiles requires discrimination between lethal and nonlethal objects within a large number of nearly identical threats. This MDA SBIR Topic "Advanced Feature Extraction Algorithm" (AFEA) has the objective of developing and demonstrating an algorithm that can contribute to this important function. It relies on matching multi-parameter simulations of the threat objects with measurements to infer the threat object geometry, motion, and thermal parameters. It relies on time resolved observations from passive and laser radar sensors. Phase I relies on a prior Aero Optics Phase I SBIR "Ladar Algorithms for Lethal Objects" (LALO) in which an inversion algorithm to extract object parameters was demonstrated. This earlier concept is augmented to account for multiple sensor platforms, improved convergence, shortened observation times and sensor noise. The algorithm is to be implemented in GPU software and hardware to speed computation and demonstrate applications to on-board, real-time processing. Phase I is a demonstration of the AFEA algorithm. Phase II will include further demonstrations, resolution of technical difficulties, and testing of the algorithm performance against large numbers and wide varieties of objects and engagements and selected flight test measurements.