Synthetic Signature Prediction and Feature Analysis for Recognition Applications
Small Business Information
31255 Cedar Valley Drive, Suite 327, Westlake Village, CA, 91362
AbstractThe ability to accurately predict the scattering and radiation behavior for broad band (up to 100 GHz) electromagnetic illumination over complex targets with geometrical details including gaps, cracks, thin edges, cavities and embedded antennae coupled with realistic material treatments is a critical requirement for many DoD low observable and automatic target modeling requirements. For applications such as SAR imagery that require hundreds of frequencies and hundreds of illumination angles for complex targets over a wide range of frequency spectrum, among the various approaches available (high frequency methods, fast frequency domain techniques, and full wave time-domain methods) the time-domain offer the best accurate approach. At HyPerComp, we have undertaken the development of a state-of-the-art time-domain CEM technology known as TEMPUS (Time-Domain EM Parallel Unstructured Simulator). TEMPUS is a complete industrial grade CEM environment that includes all aspects of a CEM simulation such as CAD geometry modeling/repair, unstructured gridding for full-scale targets with general materials, parallel run set up (for PC- and workstation clusters) and higher order accurate solvers for Maxwell's equations, and postprocessing utilities for solution visualization and extraction of final results like bistatic/monostatic RCS, SAR images, and range profiles. The primary goal of this Phase II effort is to significantly advance the state of the art of TEMPUS in the following areas: 1) maturing a hybrid TEMPUS/Xpatch (a high frequency) tool set for cost-effective simulation of electrically large problems (thousands of ƒÜƒnsize), 2) improving the choice of higher order basis functions to minimize any wave distortion (phase and amplitude errors) for long time wave propagation while reducing the computational effort for general unstructured grids, 3) implementing a high order absorbing outer boundary conditions that will allow placing the computational boundary very close (a few ƒÜƒndistance away) to the target, and 4) characterization of general material response.
* information listed above is at the time of submission.