Well-Conditioned Multi-Level Fast Multipole Modeling of Military Communication Channels
Abstract
Duke University and Virginia Polytechnic Institute and State University (VPI&SU) propose to team in a research effort to develop computer codes for the analysis and prediction of electromagnetic wave (EM) propagation over terrain. Those features of the terrain to be included are geographic scales of surface elevation, surface roughness, variable electrical constitutive properties, and vegetation cover (e.g., single and multiple trees). The frequencies to be considered include those of greatest interest to the U.S. Army in their current and projected communications and remote sensing needs. The approach will focus on the solution of appropriate integral equations describing the interaction of the EM waves and the terrain. Both institutions will work on developing the most appropriate integral equations to describe generic propagation scenarios. In the numerical solution of these integral equations, VPI&SU will concentrate on preconditioning or renormalizing them so as to minimize the number of iterations necessary to solve them to a given level of accuracy. Duke will focus on developing improved fast solvers for the preconditioned equations based upon their previous work with the Multi-Level Fast Multipole Algorithm (MLFMA). Both organizations also will attempt to keep the developed methods and codes general to the extent that they also may be applied to propagation in an urban environment. The rigorous, well-conditioned MLFMA models will be used as benchmarks to calibrate the accuracy of traditional electromagnetic propagation models, including ray, beam, and parabolic-equation (PE) solvers. (10 refs.)
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 30, 2004
- Accession Number
- ADA426492
Entities
People
- Lawrence Carin
Organizations
- Duke University