Técnicas de estimación del orden basadas en subespacios en sistemas MIMO masivo

Order estimation, also known as source enumeration, is a classical problem in array signal processing which consists in estimating the number of signals received by an array of sensors. In the last decades, numerous approaches to this problem have been proposed. However, the need of working with lar...

Descripción completa

Detalles Bibliográficos
Autor: Garg, Vaibhav|||0000-0002-6639-3324
Tipo de recurso: tesis doctoral
Fecha de publicación:2021
País:España
Institución:Universidad de Cantabria (UC)
Repositorio:UCrea Repositorio Abierto de la Universidad de Cantabria
Idioma:inglés
OAI Identifier:oai:repositorio.unican.es:10902/24821
Acceso en línea:http://hdl.handle.net/10902/24821
Access Level:acceso abierto
Palabra clave:Order Estimation
Source enumeration
Subspace averaging
Matrix completion
Missing data
Array processing
Massive MIMO
Shift invariance
Sparse representation
Model order estimation
id ES_c96be2cdd15707e64c8e945e45fb2ae3
oai_identifier_str oai:repositorio.unican.es:10902/24821
network_acronym_str ES
network_name_str España
repository_id_str
dc.title.none.fl_str_mv Técnicas de estimación del orden basadas en subespacios en sistemas MIMO masivo
Subspace-based order estimation techniques in massive MIMO
title Técnicas de estimación del orden basadas en subespacios en sistemas MIMO masivo
spellingShingle Técnicas de estimación del orden basadas en subespacios en sistemas MIMO masivo
Garg, Vaibhav|||0000-0002-6639-3324
Order Estimation
Source enumeration
Subspace averaging
Matrix completion
Missing data
Array processing
Massive MIMO
Shift invariance
Sparse representation
Model order estimation
title_short Técnicas de estimación del orden basadas en subespacios en sistemas MIMO masivo
title_full Técnicas de estimación del orden basadas en subespacios en sistemas MIMO masivo
title_fullStr Técnicas de estimación del orden basadas en subespacios en sistemas MIMO masivo
title_full_unstemmed Técnicas de estimación del orden basadas en subespacios en sistemas MIMO masivo
title_sort Técnicas de estimación del orden basadas en subespacios en sistemas MIMO masivo
dc.creator.none.fl_str_mv Garg, Vaibhav|||0000-0002-6639-3324
author Garg, Vaibhav|||0000-0002-6639-3324
author_facet Garg, Vaibhav|||0000-0002-6639-3324
author_role author
dc.contributor.none.fl_str_mv Santamaría Caballero, Luis Ignacio
Universidad de Cantabria
dc.subject.none.fl_str_mv Order Estimation
Source enumeration
Subspace averaging
Matrix completion
Missing data
Array processing
Massive MIMO
Shift invariance
Sparse representation
Model order estimation
topic Order Estimation
Source enumeration
Subspace averaging
Matrix completion
Missing data
Array processing
Massive MIMO
Shift invariance
Sparse representation
Model order estimation
description Order estimation, also known as source enumeration, is a classical problem in array signal processing which consists in estimating the number of signals received by an array of sensors. In the last decades, numerous approaches to this problem have been proposed. However, the need of working with large-scale arrays (like in massive MIMO systems), low signal-to-noise- ratios, and poor sample regime scenarios, introduce new challenges to order estimation problems. For instance, most of the classical approaches are based on information theoretic criteria, which usually require a large sample size, typically several times larger than the number of sensors. Obtaining a number of samples several times larger than the number of sensors is not always possible with large-scale arrays. In addition, most of the methods found in literature assume that the noise is spatially white, which is very restrictive for many practical scenarios. This dissertation deals with the problem of source enumeration for large-scale arrays, and proposes solutions that work robustly in the small sample regime under various noise models. The first part of the dissertation solves the problem by applying the idea of subspace averaging. The input data are modelled as subspaces, and an average or central subspace is computed. The source enumeration problem can be seen as an estimation of the dimension of the central subspace. A key element of the proposed method is to construct a bootstrap procedure, based on a newly proposed discrete distribution on the manifold of projection matrices, for stochastically generating subspaces from a function of experimentally determined eigenvalues. In this way, the proposed subspace averaging (SA) technique determines the order based on the eigenvalues of an average projection matrix, rather than on the likelihood of a covariance model, penalized by functions of the model order. The proposed SA criterion is especially effective in high-dimensional scenarios with low sample support for uniform linear arrays in the presence of white noise. Further, the proposed SA method is extended for: i) non-white noises, and ii) non-uniform linear arrays. The SA criterion is sensitive with the chosen dimension of extracted subspaces. To solve this problem, we combine the SA technique with a majority vote approach. The number of sources is detected for increasing dimensions of the SA technique and then a majority vote is applied to determine the final estimate. Further, to extend SA for arrays with arbitrary geometries, the SA is combined with a sparse reconstruction (SR) step. In the first step, each received snapshot is approximated by a sparse linear combination of the rest of snapshots. The SR problem is regularized by the logarithm-based surrogate of the l-0 norm and solved using a majorization-minimization approach. Based on the SR solution, a sampling mechanism is proposed in the second step to generate a collection of subspaces, all of which approximately span the same signal subspace. Finally, the dimension of the average of this collection of subspaces provides a robust estimate for the number of sources. The second half of the dissertation introduces a completely different approach to the order estimation for uniform linear arrays, which is based on matrix completion algorithms. This part first discusses the problem of order estimation in the presence of noise whose spatial covariance structure is a diagonal matrix with possibly different variances. The diagonal terms of the sample covariance matrix are removed and, after applying Toeplitz rectification as a denoising step, the signal covariance matrix is reconstructed by using a low-rank matrix completion method adapted to enforce the Toeplitz structure of the sought solution. The proposed source enumeration criterion is based on the Frobenius norm of the reconstructed signal covariance matrix obtained for increasing rank values. The proposed method performs robustly for both small and large-scale arrays with few snapshots. Finally, an approach to work with a reduced number of radio–frequency (RF) chains is proposed. The receiving array relies on antenna switching so that at every time instant only the signals received by a randomly selected subset of antennas are downconverted to baseband and sampled. Low-rank matrix completion (MC) techniques are then used to reconstruct the missing entries of the signal data matrix to keep the angular resolution of the original large-scale array. The proposed MC algorithm exploits not only the low- rank structure of the signal subspace, but also the shift-invariance property of uniform linear arrays, which results in a better estimation of the signal subspace. In addition, the effect of MC on DOA estimation is discussed under the perturbation theory framework. Further, this approach is extended to devise a novel order estimation criterion for missing data scenario. The proposed source enumeration criterion is based on the chordal subspace distance between two sub-matrices extracted from the reconstructed matrix after using MC for increasing rank values. We show that the proposed order estimation criterion performs consistently with a very few available entries in the data matrix.
publishDate 2021
dc.date.none.fl_str_mv 2021
2021-06-10
dc.type.none.fl_str_mv doctoral thesis
http://purl.org/coar/resource_type/c_db06
NA
http://purl.org/coar/version/c_be7fb7dd8ff6fe43
dc.type.openaire.fl_str_mv info:eu-repo/semantics/doctoralThesis
format doctoralThesis
dc.identifier.none.fl_str_mv http://hdl.handle.net/10902/24821
url http://hdl.handle.net/10902/24821
dc.language.none.fl_str_mv Inglés
eng
language_invalid_str_mv Inglés
language eng
dc.rights.none.fl_str_mv open access
http://purl.org/coar/access_right/c_abf2
Atribución 3.0 España
http://creativecommons.org/licenses/by/3.0/es/
dc.rights.openaire.fl_str_mv info:eu-repo/semantics/openAccess
rights_invalid_str_mv open access
http://purl.org/coar/access_right/c_abf2
Atribución 3.0 España
http://creativecommons.org/licenses/by/3.0/es/
eu_rights_str_mv openAccess
dc.source.none.fl_str_mv reponame:UCrea Repositorio Abierto de la Universidad de Cantabria
instname:Universidad de Cantabria (UC)
instname_str Universidad de Cantabria (UC)
reponame_str UCrea Repositorio Abierto de la Universidad de Cantabria
collection UCrea Repositorio Abierto de la Universidad de Cantabria
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869419369676668928
spelling Técnicas de estimación del orden basadas en subespacios en sistemas MIMO masivoSubspace-based order estimation techniques in massive MIMOGarg, Vaibhav|||0000-0002-6639-3324Order EstimationSource enumerationSubspace averagingMatrix completionMissing dataArray processingMassive MIMOShift invarianceSparse representationModel order estimationOrder estimation, also known as source enumeration, is a classical problem in array signal processing which consists in estimating the number of signals received by an array of sensors. In the last decades, numerous approaches to this problem have been proposed. However, the need of working with large-scale arrays (like in massive MIMO systems), low signal-to-noise- ratios, and poor sample regime scenarios, introduce new challenges to order estimation problems. For instance, most of the classical approaches are based on information theoretic criteria, which usually require a large sample size, typically several times larger than the number of sensors. Obtaining a number of samples several times larger than the number of sensors is not always possible with large-scale arrays. In addition, most of the methods found in literature assume that the noise is spatially white, which is very restrictive for many practical scenarios. This dissertation deals with the problem of source enumeration for large-scale arrays, and proposes solutions that work robustly in the small sample regime under various noise models. The first part of the dissertation solves the problem by applying the idea of subspace averaging. The input data are modelled as subspaces, and an average or central subspace is computed. The source enumeration problem can be seen as an estimation of the dimension of the central subspace. A key element of the proposed method is to construct a bootstrap procedure, based on a newly proposed discrete distribution on the manifold of projection matrices, for stochastically generating subspaces from a function of experimentally determined eigenvalues. In this way, the proposed subspace averaging (SA) technique determines the order based on the eigenvalues of an average projection matrix, rather than on the likelihood of a covariance model, penalized by functions of the model order. The proposed SA criterion is especially effective in high-dimensional scenarios with low sample support for uniform linear arrays in the presence of white noise. Further, the proposed SA method is extended for: i) non-white noises, and ii) non-uniform linear arrays. The SA criterion is sensitive with the chosen dimension of extracted subspaces. To solve this problem, we combine the SA technique with a majority vote approach. The number of sources is detected for increasing dimensions of the SA technique and then a majority vote is applied to determine the final estimate. Further, to extend SA for arrays with arbitrary geometries, the SA is combined with a sparse reconstruction (SR) step. In the first step, each received snapshot is approximated by a sparse linear combination of the rest of snapshots. The SR problem is regularized by the logarithm-based surrogate of the l-0 norm and solved using a majorization-minimization approach. Based on the SR solution, a sampling mechanism is proposed in the second step to generate a collection of subspaces, all of which approximately span the same signal subspace. Finally, the dimension of the average of this collection of subspaces provides a robust estimate for the number of sources. The second half of the dissertation introduces a completely different approach to the order estimation for uniform linear arrays, which is based on matrix completion algorithms. This part first discusses the problem of order estimation in the presence of noise whose spatial covariance structure is a diagonal matrix with possibly different variances. The diagonal terms of the sample covariance matrix are removed and, after applying Toeplitz rectification as a denoising step, the signal covariance matrix is reconstructed by using a low-rank matrix completion method adapted to enforce the Toeplitz structure of the sought solution. The proposed source enumeration criterion is based on the Frobenius norm of the reconstructed signal covariance matrix obtained for increasing rank values. The proposed method performs robustly for both small and large-scale arrays with few snapshots. Finally, an approach to work with a reduced number of radio–frequency (RF) chains is proposed. The receiving array relies on antenna switching so that at every time instant only the signals received by a randomly selected subset of antennas are downconverted to baseband and sampled. Low-rank matrix completion (MC) techniques are then used to reconstruct the missing entries of the signal data matrix to keep the angular resolution of the original large-scale array. The proposed MC algorithm exploits not only the low- rank structure of the signal subspace, but also the shift-invariance property of uniform linear arrays, which results in a better estimation of the signal subspace. In addition, the effect of MC on DOA estimation is discussed under the perturbation theory framework. Further, this approach is extended to devise a novel order estimation criterion for missing data scenario. The proposed source enumeration criterion is based on the chordal subspace distance between two sub-matrices extracted from the reconstructed matrix after using MC for increasing rank values. We show that the proposed order estimation criterion performs consistently with a very few available entries in the data matrix.This work was supported by the Ministerio de Ciencia e Innovación (MICINN) of Spain, under grants TEC2016-75067-C4-4-R (CARMEN) and BES-2017-080542.Santamaría Caballero, Luis IgnacioUniversidad de Cantabria20212021-06-10doctoral thesishttp://purl.org/coar/resource_type/c_db06NAhttp://purl.org/coar/version/c_be7fb7dd8ff6fe43info:eu-repo/semantics/doctoralThesishttp://hdl.handle.net/10902/24821reponame:UCrea Repositorio Abierto de la Universidad de Cantabriainstname:Universidad de Cantabria (UC)Inglésengopen accesshttp://purl.org/coar/access_right/c_abf2Atribución 3.0 Españahttp://creativecommons.org/licenses/by/3.0/es/info:eu-repo/semantics/openAccessoai:repositorio.unican.es:10902/248212026-06-02T12:39:31Z
score 15,300724

Similares en LA Referencia