Information Distances in Stochastic Resolution Analysis

07/11/2017
Auteurs : Radmila Pribić
Publication GSI2017
OAI : oai:www.see.asso.fr:17410:22564
contenu protégé  Document accessible sous conditions - vous devez vous connecter ou vous enregistrer pour accéder à ou acquérir ce document.
- Accès libre pour les ayants-droit
 

Résumé

A stochastic approach to resolution is explored that uses information distances computed from the geometry of data models characterized by the Fisher information in cases with spatial-temporal measurements for multiple parameters. Stochastic resolution includes probability of resolution at signal-to-noise ratio (SNR) and separation of targets. The probability of resolution is assessed by exploiting different information distances in likelihood ratios. Taking SNR into account is especially relevant in compressive sensing (CS) due to its fewer measurements. Our stochastic resolution is also compared with actual resolution from sparse-signal processing that is nowadays a major part of any CS sensor. Results demonstrate the suitability of the proposed analysis due to its ability to include crucial impacts on the performance guarantees: array configuration or sensor design, SNR, separation and probability of resolution.

Information Distances in Stochastic Resolution Analysis

Collection

application/pdf Information Distances in Stochastic Resolution Analysis Radmila Pribić
Détails de l'article
contenu protégé  Document accessible sous conditions - vous devez vous connecter ou vous enregistrer pour accéder à ou acquérir ce document.
- Accès libre pour les ayants-droit

A stochastic approach to resolution is explored that uses information distances computed from the geometry of data models characterized by the Fisher information in cases with spatial-temporal measurements for multiple parameters. Stochastic resolution includes probability of resolution at signal-to-noise ratio (SNR) and separation of targets. The probability of resolution is assessed by exploiting different information distances in likelihood ratios. Taking SNR into account is especially relevant in compressive sensing (CS) due to its fewer measurements. Our stochastic resolution is also compared with actual resolution from sparse-signal processing that is nowadays a major part of any CS sensor. Results demonstrate the suitability of the proposed analysis due to its ability to include crucial impacts on the performance guarantees: array configuration or sensor design, SNR, separation and probability of resolution.
Information Distances in Stochastic Resolution Analysis
application/pdf Information Distances in Stochastic Resolution Analysis (slides)

Média

Voir la vidéo

Métriques

0
0
958.22 Ko
 application/pdf
bitcache://b26fb1aaaefbba56b5cf82b45454de5f72b9b08d

Licence

Creative Commons Aucune (Tous droits réservés)

Sponsors

Sponsors Platine

alanturinginstitutelogo.png
logothales.jpg

Sponsors Bronze

logo_enac-bleuok.jpg
imag150x185_couleur_rvb.jpg

Sponsors scientifique

logo_smf_cmjn.gif

Sponsors

smai.png
gdrmia_logo.png
gdr_geosto_logo.png
gdr-isis.png
logo-minesparistech.jpg
logo_x.jpeg
springer-logo.png
logo-psl.png

Organisateurs

logo_see.gif
<resource  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
                xmlns="http://datacite.org/schema/kernel-4"
                xsi:schemaLocation="http://datacite.org/schema/kernel-4 http://schema.datacite.org/meta/kernel-4/metadata.xsd">
        <identifier identifierType="DOI">10.23723/17410/22564</identifier><creators><creator><creatorName>Radmila Pribić</creatorName></creator></creators><titles>
            <title>Information Distances in Stochastic Resolution Analysis</title></titles>
        <publisher>SEE</publisher>
        <publicationYear>2018</publicationYear>
        <resourceType resourceTypeGeneral="Text">Text</resourceType><subjects><subject>Information geometry</subject><subject>resolution</subject><subject>compressive sensing</subject><subject>radar</subject></subjects><dates>
	    <date dateType="Created">Wed 7 Mar 2018</date>
	    <date dateType="Updated">Wed 7 Mar 2018</date>
            <date dateType="Submitted">Mon 10 Dec 2018</date>
	</dates>
        <alternateIdentifiers>
	    <alternateIdentifier alternateIdentifierType="bitstream">b26fb1aaaefbba56b5cf82b45454de5f72b9b08d</alternateIdentifier>
	</alternateIdentifiers>
        <formats>
	    <format>application/pdf</format>
	</formats>
	<version>37288</version>
        <descriptions>
            <description descriptionType="Abstract">A stochastic approach to resolution is explored that uses information distances computed from the geometry of data models characterized by the Fisher information in cases with spatial-temporal measurements for multiple parameters. Stochastic resolution includes probability of resolution at signal-to-noise ratio (SNR) and separation of targets. The probability of resolution is assessed by exploiting different information distances in likelihood ratios. Taking SNR into account is especially relevant in compressive sensing (CS) due to its fewer measurements. Our stochastic resolution is also compared with actual resolution from sparse-signal processing that is nowadays a major part of any CS sensor. Results demonstrate the suitability of the proposed analysis due to its ability to include crucial impacts on the performance guarantees: array configuration or sensor design, SNR, separation and probability of resolution.
</description>
        </descriptions>
    </resource>
.

adfa, p. 1, 2011. © Springer-Verlag Berlin Heidelberg 2011 Information Distances in Stochastic Resolution Analysis Radmila Pribić Sensors Advanced Developments, Thales Nederland Delft, The Netherlands Radmila.Pribic@nl.thalesgroup.com Abstract. A stochastic approach to resolution is explored that uses information distances computed from the geometry of data models characterized by the Fisher information in cases with spatial-temporal measurements for multiple parameters. Stochastic resolution includes probability of resolution at signal-to- noise ratio (SNR) and separation of targets. The probability of resolution is assessed by exploiting different information distances in likelihood ratios. Taking SNR into account is especially relevant in compressive sensing (CS) due to its fewer measurements. Our stochastic resolution is also compared with actual resolution from sparse-signal processing that is nowadays a major part of any CS sensor. Results demonstrate the suitability of the proposed analysis due to its ability to include crucial impacts on the performance guarantees: array configuration or sensor design, SNR, separation and probability of resolution. Keywords: resolution; information geometry; compressive sensing; radar; 1 Introduction Resolution is primarily described by the minimum distance between two objects that still can be resolved (e.g. [1]). Stochastic resolution has been introduced ([2]) by including the Cramér-Rao bound (CRB). The stochastic approach was extended with the probability of resolution at a given separation and signal-to-noise ratio (SNR) obtained via an asymptotic generalized likelihood ratio (GLR) test based on Euclidean distances ([3]). Information resolution have also been explored with an arbitrary test ([4]). For completeness of the stochastic approach, information geometry (IG, [5-9]) and compressive sensing (CS, [10-11]) are combined due to their focus on information content ([12-15]). In [13-14], the Fisher-Rao information distance is recognized in the asymptotic GLR. In [15], links to other information distances and tighter resolution bounds have been obtained that we expand here to multiple parameters. In the IG-based resolution analysis, the Fisher information metric ( FIM ) is employed for computing resolution bounds or information resolution. The stochastic analysis is crucial when using fewer measurements what is typical for compressive data acquisition in the front-end of a CS sensor (e.g. [10]). In the back-end, the analysis provides metrics for the high-resolution performance of sparse-signal processing (SSP). In radar, SSP can be seen as a model-based refinement of existing processing (e.g. [16]). Despite substantial CS research during last decade (e.g. [10- 11]), complete guarantees of CS-radar resolution performance have not been developed yet. Both IG and CS can improve existing sensors due to their focus on the information content in measurements rather than the sensing bandwidth only. The resolution ability is primarily given by the Rayleigh distance determined by the sensing bandwidth from sensor design. In array processing, this deterministic resolution relies on the sensor wavelength and the array size. The stochastic resolution analysis includes also SNR available from the data acquisition. Moreover, besides the sensor design, it also involves targets of interest with their SNR and separation. In [13- 14], we assess the probability of resolution at a given SNR and separation by applying the Fisher-Rao information distances in the asymptotic distribution of the GLR. In this paper and in [15], we extend the stochastic resolution analysis with tighter resolution bounds obtained via an LR test with different types of information distances. In Section 2, modeling of radar measurements and their SSP are summarized. In Section 3, our stochastic resolution analysis is explained and expanded to multiple parameters. In Section 4, numerical results from the stochastic resolution analysis and from SSP are compared. In the end, conclusions are drawn and future work indicated. 2 Data Modelling and SSP The data models in array processing and in SSP needed in Section 3 are given here. In a linear array (LA) of