Recovering inhomogeneities in a waveguide using eigensystem decomposition

Abstract

We present an eigensystem decomposition method to recover weak inhomogeneities in a waveguide from knowledge of the far-field scattered acoustic fields. Due to the particular geometry of the waveguide, which supports only a finite number of propagating modes, the problem of recovering inhomogeneities in a waveguide has a different set of challenges than the corresponding problem in free space. Our method takes advantage of the spectral properties of the far-field matrix, and by using its eigenvalues and its eigenvectors we obtain a representation of the linearized solution to the inverse problem in terms of products of fields which are linear combinations of the propagating modes of the waveguide with weights given by the eigenvectors of the far-field matrix. The problem of finding the unknown inhomogeneity reduces to a problem of determining some coefficients in a finite system of linear equations, whose coefficients depend on the background medium and the eigenvalues and the eigenvectors of the far-field matrix. By numerically implementing our inverse algorithm we reconstructed the inhomogeneities present in the waveguide. We show that (1) even with as few as seven propagating modes we obtain a good recovery of the size and shape of the inhomogeneity; (2) multiple inhomogeneities can be well recovered and, as expected, (3) the recovery improves when the number of propagating modes increases.

Sojourn times, manifolds with infinite cylindrical ends, and an inverse problem for planar waveguides

“…Most of these seem to be related to the fixed energy recovery problem, where one needs more entries in the scattering matrix. See [2] for a recent result and further references.…”

Imaging Extended Reflectors in a Terminating Waveguide

“…defined for each point y s in the search domain S. Note that by replacing (16) into (17) and using the expression of G n given in (12), it is easy to show that…”

The Reconstruction of Obstacles in a Waveguide Using Finite Elements

“…For planar waveguides, there exist eigenvalues and eigenfunctions of Sturm-Liouville type (see [8]) given by…”

“…In [18], Xu et al applied a method using generalized dual space indicator for an obstacle in a shallow water waveguide. Dediu and McLaughlin [8] proposed an eigensystem decomposition to recover weak inhomogeneities in a planar waveguide from far-field data. In [4,5], Bourgeois and Lunéville employed the linear sampling method to reconstruct sound soft obstacles as well as cracks in a planar waveguide from near-field data.…”