Signature Prediction and Uncertainty Analysis for Radar-based MDA Applications

A new, highly efficient physics-based approach is proposed for mapping the variation in RCS of MDA targets of interest over the full range of angle, frequency, geometry, and material properties relevant in missile defense scenarios. This approach implements recent developments in reduced-basis methodology (RBM) for Maxwell’s equations and combines the speed of RBM evaluations with the efficiency of stochastic collocation for uncertainty quantification to produce valid statistical measures of RCS variability, as well as RCS estimates of certified accuracy for each condition of observation. This RBM approach to RCS evaluation is built on generating accurate full-wave solutions for a small subset of the various parameters. In the proposed program, these solutions will be computed using the TEMPUS full-wave solver, which has demonstrated the ability to capture all the subtle effects of target structure on the radar return. For very large targets, the Cross-Flux technique will be used to combine TEMPUS and Xpatch solutions to produce an accurate hybrid solution. Speedup of the TEMPUS solver itself will be sought by implementing local time-stepping and variable-order local field representations. Porting the solver algorithms to a graphics card architecture will also be investigated, as that strategy offers speedups of 20 to 100.