Title :
MUSIC Algorithm for Two-Dimensional Inverse Problems With Special Characteristics of Cylinders
Author :
Chen, Xudong ; Agarwal, Krishna
Author_Institution :
Dept. of Electr. & Comput. Eng., Nat. Univ. of Singapore, Singapore
fDate :
6/1/2008 12:00:00 AM
Abstract :
A multiple signal classification (MUSIC) algorithm is proposed to determine the positions of small cylinders, which have some special characteristics that may lend traditional MUSIC unsuccessful. The TM case is discussed in detail. Up to three independent currents are induced on each cylinder in the TM case. However, for a cylinder, if the material property is the same as the background medium in one or two of the principal axes, or if the geometry makes one or two of the induced currents very small with respect to the others (so as to be negligible in noisy scenario), the number of significant sources may be less than three and the traditional MUSIC may fail in locating the cylinder. In such event, knowledge of the direction of the induced source becomes mandatory to locate the cylinder. The proposed method determines the direction of the induced source and uses the background Green´s function corresponding to it as MUSIC´s test function. The method uses minimization of the analytical expressions and gives good results in noise-free as well as noisy scenarios.
Keywords :
Green´s function methods; inverse problems; signal classification; Green function; MUSIC Algorithm; induced source direction; multiple signal classification algorithm; two-dimensional inverse problems; Acoustic scattering; Classification algorithms; Electromagnetic scattering; Engine cylinders; Geometry; Inverse problems; Material properties; Multiple signal classification; Tensile stress; Testing; Multiple signal classification (MUSIC); two-dimensional electromagnetic inverse problems;
Journal_Title :
Antennas and Propagation, IEEE Transactions on
DOI :
10.1109/TAP.2008.923333