Correlated Radar Clutter Map Creation for Multi-spectral Scene Generation

This SBIR affords the opportunity to design a high-speed, spatially correlated, RF clutter map generator that meets the Army’s requirements for testing and evaluating integrated flight simulations at AMRDEC. A successful design will input known terrain height and class maps, model the coupled, coherent interactions of natural and man-made scene scatterers (e.g. sand, gravel, asphalt, concrete, brick, grasses, trees and other foliage, and targets), track polarization components, and ultimately produce radar cross section maps which take into account natural statistical variation while remaining spatially correlated with the input height and class maps. Various techniques will be investigated, including fast Geometry Processing Unit (GPU)-based sampling, Fourier and fractal representations, and superposition of pre-generated bistatic scatter centers, the latter being a technique with which JRM has unique experience. In addition, for increased fidelity, the design will be able to incorporate known empirically-collected parameterizations for various terrain types, including wind-speed-dependence statistics. Metrics for assessing accuracy and performance against empirical databases and legacy non-real-time solutions will be developed and implemented. A clutter map generator of this type would address the above limitations to provide for more tractable stimulation of multispectral seeker models and data fusion algorithms in highly variable, complex background clutter environments.