0% found this document useful (0 votes)

149 views7 pages

Music Algorithm (Multiple User Signal Classification Algorithm)

The document discusses the MUSIC algorithm for direction of arrival (DOA) estimation using antenna arrays. It introduces MUSIC as an improvement over previous eigenstructure methods that correctly exploits the measurement model. The key steps of MUSIC are described: 1) estimating the signal and noise subspaces through eigendecomposition of the covariance matrix, 2) computing the MUSIC spatial spectrum to identify peaks corresponding to DOAs. Simulation results show that increasing the number of antenna elements, using an element spacing of λ/2, and higher frequencies improve the resolution of DOA estimation.

Uploaded by

Fakhar

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as DOCX, PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

149 views7 pages

Music Algorithm (Multiple User Signal Classification Algorithm)

Uploaded by

Fakhar

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as DOCX, PDF, TXT or read online on Scribd

You are on page 1/ 7

CHAPTER 2

MUSIC ALGORITHM (MULTIPLE

USER SIGNAL CLASSIFICATION
ALGORITHM)
2.1 Introduction:
There have been several solution to the problem including maximum likelihood (ML) of Capon
(1969) and Burgs maximum entropy method (ME) although successful and implemented these
methods have fundamental limitations especially bias and sensitivity in parameter estimates
largely because they use incorrect method ( i.e AR 1 instead of the specialized ARMA 2) of the
measurements. Pisarenko (1973) was one of the first in exploiting the structure of data model
doing so in context of estimation of parameters of cissoids(a curve generated from two given
curves) in additive noise using a covariance approach. Schmidt (1977) and Bienvenu (1979) were
the first to correctly exploit the measurement model in the case of sensor arrays of arbitrary
form ,Schmidt in particular ,accomplished this by first deriving a complete geometric solution in
the absence of noise, then cleverly extending the geometric concepts to obtain a reasonable
approximated solution in the presence of noise. The resulting algorithm was called
MUSIC(Multiple Signal Classification)[7 ].
Other eigen structure methods include ESPIRIT (estimation of signal parameters via rotational
invariance techniques)method, minimum norm methods and the weighted subspace fitting
method.

In statistics and signal processing, an autoregressive (AR) model is a type of random process which is
often used to model and predict various types of natural and social phenomena.
2

In statistics and signal processing, autoregressive moving average (ARMA) models, sometimes
called Box-Jenkins models after the iterative Box-Jenkins methodology usually used to estimate them, are
typically applied to auto correlated time series data. The ARMA model is a tool for understanding and,
perhaps, predicting future values in this series. The model consists of two parts, an autoregressive (AR) part
and a moving average (MA) part. The model is usually then referred to as the ARMA(p, q) model
where p is the order of the autoregressive part and q is the order of the moving average part .

Real Time Implementation of Direction of Arrival Estimation

Figure: 2.1 antenna array system

Antenna resolution properties have been nvestigated by a large number of researchers to compare
the DOA estimation and beam forming methods to different array geometries. In detailed
performance evaluation based upon hundred of evaluation it has been declared by MITs Lincoln
laboratory , has declare MUSIC as the most promising and leading algorithm for detailed study
and hardware implementation amongst high resolution algorithms then available.
The popularity of MUSIC is due to its generality. For example it is applicable to arrays of any
dimension and known configuration and response and can be used to estimate various parameters
per source. MUSIC requires priori knowledge of second order spatial statistics of the background
noise and interference field[4].

2.2 Implementation of MUSIC Algorithm:

The MUSIC algorithm was developed by Schmidt by noting that the desired signal array response
is orthogonal to the noise subspace. The signal and noise subspace are first identified using eigen
decomposition of the received signal covariance matrix. Following , the MUSIC spatial spectrum
is computed from which the DOAs are estimated. Inside the algorithm,we first define the general
array to be the set:
A={a(i): i

(2.1)

For some region of interest in the DOA space .The array manifold 3 is assumed unambiguous
and known for all values of angle either analytically or through some calibration procedure. The
objective is to apply appropriate methods to the received signals so as to extract region of out
of range of [4].
If noise was absent in equation n(t)=a( ) the observation could be confined entirely to Mdimensional subspace defined by the span of A().For the no noise case ,it is simply a matter of
finding unique elements of A that intersects with this subspace. A different approach is necessary
in the case when noise is present since observation becomes full rank .The approach of
MUSIC and other subspace based algorithms is to estimate dominant subspace of the
observations and then find the elements of A() that are in some sense closest to this
subspace[4].
12

Real Time Implementation of Direction of Arrival Estimation

The subspace estimation step is typically achieved by eigen decomposition of the auto covariance
matrix of the received data R. For MUSIC to be applicable, the emitter covariance is required to
be full-rank, i.e that M=M. The eigen vectors can assumed to be forming an orthonormal basis 4
i.e QQH=QHQ. Once the subspaces are determined the DOA of the desired signals can be
calculated over the spectrum region of interest using

PM()= (a() a())/( a()EN EN* a()) [7]

(2.2)

Number of signals that an array can detect depends upon the number of elements . It has been
verified that N elements array can detect upto N-1 uncorrelated signals. [4].

2.3 Simulation Results:

To demonstrate the efficiency of the algorithm we choose a ULA(uniform linear array) and
different parameter variations.

2.3.1-Variation of Number of Antenna Elements:

There are four sources located in far field of the array with 1= -5,2= 0,3= 5 and 4=10. The
spacing between two elements, d=/2. The number of snap shots (number of samples) is 1000.
The result of the simulation is given

Figure 2.1 The number of elements is 8.

Figure 2.2 The number of elements is 16

array manifold is a set array response vectors corresponding to all possible DOA

In linear algebra, a basis is a set of linearly independent vectors that, in a linear combination, can represent
every vector in a given vector space or free module.

MUSIC (Multiple User Signal Classification) Algorithm

Figure 2.3 The number of elements is 32.

Figure 2.4 The number of elements is 64.

2.3.2-Changing the Spacing between Elements:

There are four sources located in far field of the array with 1= -5,2= 0,3= 5 and 4=10. The
antenna array has 8 elements. The number of snap shots (number of samples) is
1000.

Figure 2.5 d= /4.

Figure 2.6 d= /5.

Real Time Implementation of Direction of Arrival Estimation

Figure 2.7 d= d= /6.

Figure 2.9 d= /8.

Figure 2.11 d= /2.

Figure 2.8 d= /7.

Figure 2.10 d= /9

Figure 2.12 d= 0.8 .

MUSIC (Multiple User Signal Classification) Algorithm

Figure 2.13 d= 1 .

Figure 2.14 d= 2 .

2.3.3-Changing Frequency:
There are four sources located in far field of the array with 1= -5,2= 0,3= 5 and 4=10.The
antenna array has 8 elements and the spacing between two elements ,d= 0.5 . The number of
snap shots is 1000.

Figure 2.15 f= 1 GHz.

Figure 2.16 f= 1MHz.

=v/f;

=v/f

=(340)/(1000);

=340/(1x10)

=0.340m

=3.4x10 -4m

Real Time Implementation of Direction of Arrival Estimation

2.4 CONCLUSION:
From the simulation results discussed:

Greater the number of elements better the response of an antenna .

Best performance at an antenna spacing of /2.
If the spacing is increased beyond /2 alaising occurs.
If the spacing is less than /2 the performance is not good.
Greater frequency smaller the value of and hence better response.