Minimum-time optimization of in-situ antenna performance

Period of Performance: 04/10/2013 - 06/15/2014

$437K

Phase 2 SBIR

Recipient Firm

Mathematical Systems & Solutions, Inc.
685 Busch Garden Dr. Array
Pasadena, CA 91105
Principal Investigator

Abstract

We propose development of a software capability which, based on use of accurate and efficient exact-physics computational electromagnetics (CEM) solvers together with a combination of automated and interactive optimization tools, will enable optimization of the properties of on-platform mounted-antenna systems. For accuracy and modeling flexibility the proposed codes are based on the fast, high-order frequency-domain algorithms put forth recently by our team-which can accurately and efficiently resolve both low- and high-frequency behavior while accounting for scattering, transmission and absorption by metallic, dielectric and magnetic components. These solvers have provided some of the most efficient PDE solution methods in existence: they can successfully tackle electrically large problems in complex engineering geometries, including problems which are significantly beyond the capabilities of the most competitive alternative solvers. Our novel parallel CPU implementations of these algorithms, in turn, demonstrate a high degree of parallel efficiency. Our overall approach to in-situ antenna optimization, finally, relies on a combination of classical gradient-based optimization methods together with a new technique based on use of "low-rank electromagnetic bases", which can yield a large number of scattering solutions (and thus, enable efficient optimization algorithms) from a limited number of actual solution runs.