Scribd
Upload
81 views20 pages

DOA Estimation Using MUSIC and Root MUSIC Methods: Chhavipreet Singh (Y5156) Siddharth Sahoo (Y5827447)

This document discusses direction of arrival (DOA) estimation methods MUSIC and Root-MUSIC. It begins with an introduction to the DOA problem and the Cramer-Rao lower bound for DOA estimation. It then describes how MUSIC estimates DOAs by exploiting the orthogonality between the signal and noise subspaces. Root-MUSIC is introduced as a model-based method that directly estimates DOAs by finding the roots of a polynomial derived from the received data and steering vectors. The document provides mathematical details on how both methods calculate DOA estimates from the received signals and antenna array model.

Uploaded by

ali
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
81 views20 pages

DOA Estimation Using MUSIC and Root MUSIC Methods: Chhavipreet Singh (Y5156) Siddharth Sahoo (Y5827447)

This document discusses direction of arrival (DOA) estimation methods MUSIC and Root-MUSIC. It begins with an introduction to the DOA problem and the Cramer-Rao lower bound for DOA estimation. It then describes how MUSIC estimates DOAs by exploiting the orthogonality between the signal and noise subspaces. Root-MUSIC is introduced as a model-based method that directly estimates DOAs by finding the roots of a polynomial derived from the received data and steering vectors. The document provides mathematical details on how both methods calculate DOA estimates from the received signals and antenna array model.

Uploaded by

ali
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 20

DOAEstimationUsingMUSIC

andRootMUSICMethods
Chhavipreet Singh(Y5156)
SiddharthSahoo(Y5827447)
EE602 StatisticalSignalProcessing
Instructor:DrR.Hegde

Overview

IntroductiontotheDOAproblem
CRBforDOAEstimation
DOAEstimationusingMUSIC
Root MUSIC:ModelBasedParameter
Estimation

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

DirectionofArrivalEstimation
Theproblemathand:
Estimatethedirectionofasignalfromthereceivedsignals

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

DirectionofArrivalEstimation
Problemdescription:
Asetofincomingsignals
Incomingsignaldirection,i
Alinear,equispacedantennaarraywithNelements

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

DirectionofArrivalEstimation
Difficultiesfaced
Noofincomingsignalsunknown
Unknowndirectionandamplitude
Thesignalsarecorruptedbynoise

Assumption:
NumberofunknownsignalsisM

M<N
WhiteGaussianNoise
EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

CRBforDOAEstimation
x islengthNvectorofreceivedsignals
isthesetofparameters
v isaknown functionofparameters
weknowthat,
where

J istheFischerInformationMatrix
EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

CRBforDOAEstimation
Assumingasinglesignalcorruptedbynoise,
wheres isthesteeringvectorofthesignal
niszeromeanGaussianwithcovariance
Assuming
Inourcase

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

CRBforDOAEstimation

TheCRBfortheDOAEstimationproblemistherefore,

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

DOAEstimationusingMUSIC
MUSIC:MUltiple SIgnal Classification
ThematrixS isaN xM matrixoftheM steeringvectors

Thecorrelationmatrixofx canbewrittenas

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

DOAEstimationusingMUSIC
Rs isN xM matrixwithrankM
IthasN M eigenvectorscorrespondingtozeroeigenvalue
Letqm besuchaneigenvector

ThisimpliesthatallN M eigenvectors(qm)ofRs correspondingto


thezeroeigenvalues areorthogonal toallM signalsteeringvectors
EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

DOAEstimationusingMUSIC
Letqm betheNx(N M)matrixofeigenvectorsofRs
Thepseudo spectrumisgivenby

TheeigenvectorsmakingupQn areorthogonaltothesignal
steeringvectors
Thedenominatorbecomeszerowhen isasignaldirection
TheestimatedsignaldirectionsaretheM largestpeaksinthe
pseudo spectrum

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

DOAEstimationusingMUSIC
EstimatingtheeigenvectorsinQn fromtheeigenvectorsofR
Foranyeigenvector qm Q,

ThereforeanyeigenvectorofisalsoaneigenvectorofR
Thecorrespondingeigenvalue +2
If

then

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

DOAEstimationusingMUSIC
WecanpartitiontheeigenvectormatrixQ intoasignalmatrix
Qs andnoisesubspaceQn
TheM columnsofQs correspondtotheM signaleigenvalues
TheN M columnsofQn correspondtothenoiseeigenvalues
Themth signaleigenvalue isgivenby

Byorthogonality ofQ, Qs isorthogonaltoQn

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

DOAEstimationusingMUSIC
Wesawthatallnoiseeigenvectorsareorthogonaltothe
signalsteeringvectors
ThisisthebasisforMUSIC

whereqm isoneofthe(N M)noiseeigenvectors


If isequaltoDOAofoneofthesignals,thedenominatoriszero
MUSICidentifiesthepeaksofthefunctionPMUSIC()asthe
directionsofarrival
EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

Root MUSIC:ModelBased
ParameterEstimation
InMUSICaccuracyislimitedbythediscretizationatwhich
PMUSIC()isevaluated
Itrequireshumaninteractionoracomprehensivesearch
algorithmtodeterminethepeaks
TheRoot MUSICmethodresultsdirectlyinnumericvalues
fortheestimateddirections
Itusesamodel ofthereceivedsignalasafunctionoftheDOA
TheDOA,,isaparameter inthismodel.
Basedonthismodelandthereceiveddatatheparameteris
estimated
EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

Root MUSIC:ModelBased
ParameterEstimation
Weusesteeringvectorasthemodel
Assumingthereceivingantennaisthelineararrayof
equispaced,isotropicelements

Theinnerproductofeigenvectorqm andthesteeringvector
s()isequivalenttoapolynomialinz
Form=m=(M+1),...,N,wearelookingfortheroots
ofapolynomial
EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

Root MUSIC:ModelBased
ParameterEstimation
Thepolynomialweuseisgivenby

where

C=Qn QnH

Takingl=n m,

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

Root MUSIC:ModelBased
ParameterEstimation
Thepolynomialobtainedisofdegree(2N2)
Ifz isazeroofthepolynomialthensois1/z*
zand1/z*havethesamephaseandreciprocalmagnitude
hencebothcarrythesameinformation
Oneofthesetwoliewithintheunitcircle
Ofthe(N 1) rootswithintheunitcirclechosetheMclosest
totheunitcircle(zm ,m=1,,M)

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

Root MUSIC:ModelBased
ParameterEstimation
Weobtainedthedirectionsofarrivalusing

AsRoot MUSIConlyworriesaboutthephaseoftheroots,
errorsandthemagnitudeareirrelevant

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

THANKYOU

EE602 StatisticalSignalProcessing
Instructor:Dr.R.Hegde

You might also like