Robust and GPU-accelerated IPO to improve high frequency EM scattering from large scale radar scene simulation

被引:0
作者
Pajot, Anthony [1 ]
Douchin, Nicolas [1 ]
Mametsa, Henri-Jose [2 ]
机构
[1] Oktal SE, Toulouse, France
[2] Off Natl Etud & Rech Aerosp, Toulouse, France
来源
2014 INTERNATIONAL RADAR CONFERENCE (RADAR) | 2014年
关键词
GMRES; ALGORITHM; RCS;
D O I
暂无
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Tools for electromagnetic simulation make EM computations down to millimeter wavelength possible from large complex targets and their environments. In this paper, we present how we add localized support for N-OP, most useful for cavities and other complex objects subject to multiple reflections, to our shooting-and-bouncing-ray SE-RAY-EM software. This extension is made with the Iterative Physical Optics (IPO) method, which we improve to make it suitable for use in intensive EM simulation of large scale scenes with localized N-OP bounces accuracy. As we present, this requires a generalization in materials, numerical robustness, computational efficiency and a more precise hybridization method than existing work.
引用
收藏
页数:6
相关论文
共 14 条
[11]  
SAAD Y, 1986, SIAM J SCI STAT COMP, V7, P856, DOI 10.1137/0907058
[12]   A FLEXIBLE INNER-OUTER PRECONDITIONED GMRES ALGORITHM [J].
SAAD, Y .
SIAM JOURNAL ON SCIENTIFIC COMPUTING, 1993, 14 (02) :461-469
[13]  
Telles J., ENG ANAL BOUNDARY EL
[14]   GMRESR: a Family of Nested GMRES Methods [J].
van der Vorst, H. A. ;
Vuik, C. .
NUMERICAL LINEAR ALGEBRA WITH APPLICATIONS, 1994, 1 (04) :369-386