Iteratively reweighted tensor SVD for robust multi-dimensional harmonic retrieval
Weize Sun, Xin Lin, Hing-Cheung So, Lei Huang, Qiang Li
Abstract
In this paper, parameter estimation for multi-dimensional sinusoids in additive impulsive noise is addressed. Our underlying idea is to minimize the ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</sub> -norm of the residual error tensor, where 1 <; p <; 2, and transform this problem to an iterative ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -norm minimization. In doing so, we can utilize the tensorial structure of the received data and then apply iteratively reweighted tensor singular value decomposition, referred to as IR-t-SVD, to recover the subspace or the signal tensor. After the recovery step, standard subspace techniques can be applied for parameter estimation. Based on the numerical results, IR-t-SVD outperforms several state-of-the-art methods in terms of mean square frequency error under α-stable noise.
BibTeX
@inproceedings{icassp2016_iterativelyrewei,
title = {Iteratively reweighted tensor SVD for robust multi-dimensional harmonic retrieval},
author = {Weize Sun and Xin Lin and Hing-Cheung So and Lei Huang and Qiang Li},
booktitle = {ICASSP 2016},
year = {2016}
}