DocumentCode :
3750626
Title :
Compressed sensing based on FE-BI for solving electromagnetic scattering problems
Author :
Hou Dayou;Chen Mingsheng;Kong Meng;Cao Xinyuan;Zhang Liang;Wu Xianliang
Author_Institution :
School of Electronic and Information Engineering, Hefei Normal University of China, Hefei 230601
Volume :
2
fYear :
2015
Firstpage :
1
Lastpage :
3
Abstract :
To calculate electromagnetic scattering problems over a wide incident angle by finite Element Boundary integral (FE-BI), the difficult usually lies in a great amount of repeated computation, since iterative operations are required at every angle with the incident direction changing. A new solution scheme based on compressed sensing (CS) is proposed combined with FE-BI to improve the computational efficiency, in which a new incident source including rich angle information is constructed. Under the incident source, measurements of field values can be obtained with FEM, and the field value can be reconstructed at each node and each angle of incidence with orthogonal matching pursuit (OMP). In contrast with traditional FE-BI, the calculated results by the improved method are proposed, it is shown that the new method can greatly reduce computation time and maintain high accuracy.
Keywords :
"Finite element analysis","Sparse matrices","Compressed sensing","Electromagnetic scattering","Matching pursuit algorithms","Transforms","Integral equations"
Publisher :
ieee
Conference_Titel :
Microwave Conference (APMC), 2015 Asia-Pacific
Print_ISBN :
978-1-4799-8765-8
Type :
conf
DOI :
10.1109/APMC.2015.7413113
Filename :
7413113
Link To Document :
بازگشت