Main Content

# peig

Pseudospectrum using eigenvector method

## Description

[S,wo] = peig(x,p) implements the eigenvector spectral estimation method and returns S, the pseudospectrum estimate of the input signal x, and a vector wo of normalized frequencies (in rad/sample) at which the pseudospectrum is evaluated. The pseudospectrum is calculated using estimates of the eigenvectors of a correlation matrix associated with the input data x. You can specify the signal subspace dimension using the input argument p.

[S,wo] = peig(x,p,wi) returns the pseudospectrum computed at the normalized frequencies specified in vector wi. The vector wi must have two or more elements, because otherwise the function interprets it as nfft.

[S,wo] = peig(___,nfft) specifies the integer length of the FFT, nfft, to use to estimate the pseudospectrum. This syntax can include any combination of input arguments from previous syntaxes.

[S,wo] = peig(___,'corr') forces the input argument x to be interpreted as a correlation matrix rather than a matrix of signal data. For this syntax, x must be a square matrix, and all of its eigenvalues must be nonnegative.

[S,fo] = peig(x,p,nfft,fs) returns the pseudospectrum computed at the frequencies specified in vector fo (in Hz). Supply the sample rate fs in Hz.

[S,fo] = peig(x,p,fi,fs) returns the pseudospectrum computed at the frequencies specified in the vector fi. The vector fi must have two or more elements, because otherwise the function interprets it as nfft.

[S,fo] = peig(x,p,nfft,fs,nwin,noverlap) returns the pseudospectrum S by segmenting the input data x using the window nwin and overlap length noverlap.

example

[___] = peig(___,freqrange) specifies the range of frequency values to include in fo or wo.

[___,v,e] = peig(___) returns the matrix v of noise eigenvectors, along with the associated eigenvalues in the vector e.

example

peig(___) with no output arguments plots the pseudospectrum in the current figure window.

## Examples

collapse all

Implement the eigenvector method to find the pseudospectrum of the sum of three sinusoids in noise. Use the default FFT length of 256. The inputs are complex sinusoids so you set p equal to the number of inputs. Use the modified covariance method for the correlation matrix estimate.

n = 0:99;
s = exp(1i*pi/2*n)+2*exp(1i*pi/4*n)+exp(1i*pi/3*n)+randn(1,100);
X = corrmtx(s,12,'mod');
peig(X,3,'whole')

Generate a real signal that consists of the sum of two sinusoids embedded in white Gaussian noise of unit variance. The signal is sampled at 100 Hz for 1 second. The sinusoids have frequencies of 25 Hz and 35 Hz. The lower-frequency sinusoid has twice the amplitude of the other.

fs = 100;
t = 0:1/fs:1-1/fs;

s = 2*sin(2*pi*25*t)+sin(2*pi*35*t)+randn(1,100);

Use the eigenvector method to compute the pseudospectrum of the signal between 0 and the Nyquist frequency. Specify a signal subspace dimension of 2 and a DFT length of 512.

peig(s,2,512,fs,'half')

It is not possible to resolve the two sinusoids because the signal is real. Repeat the computation using a signal subspace of dimension 4.

peig(s,4,512,fs,'half')

## Input Arguments

collapse all

Input signal, specified as a vector or matrix. If x is a vector, then it is treated as one observation of the signal. If x is a matrix, each row of x represents a separate observation of the signal. For example, each row is one output of an array of sensors, as in array processing, such that x'*x is an estimate of the correlation matrix.

Note

You can use the output of corrmtx to generate x.

Complex Number Support: Yes

Subspace dimension, specified as a real positive integer or a two-element vector. If p is a real positive integer, then it is treated as the subspace dimension. If p is a two-element vector, the second element of p represents a threshold that is multiplied by λmin, the smallest estimated eigenvalue of the signal's correlation matrix. Eigenvalues below the threshold λmin*p(2) are assigned to the noise subspace. In this case, p(1) specifies the maximum dimension of the signal subspace. The extra threshold parameter in the second entry in p provides you more flexibility and control in assigning the noise and signal subspaces.

Note

If the inputs to peig are real sinusoids, set the value of p to double the number of input signals. If the inputs are complex sinusoids, set p equal to the number of inputs.

Complex Number Support: Yes

Input normalized frequencies, specified as a vector.

Data Types: double

Number of DFT points, specified as a positive integer. If nfft is specified as empty, the default nfft is used.

Sample rate, specified as a positive scalar in Hz. in Hz. If you specify fs with the empty vector [], the sample rate defaults to 1 Hz.

Input frequencies, specified as a vector. The pseudospectrum is computed at the frequencies specified in the vector.

Length of rectangular window, specified as a nonnegative integer.

Number of overlapped samples, specified as a nonnegative integer smaller than the length of window.

Note

The arguments nwin and noverlap are ignored when you include 'corr' in the syntax.

Frequency range of pseudospectrum estimates,specified as one of 'half', whole, or 'centered'.

• 'half' — Returns half the spectrum for a real input signal x. If nfft is even, then S has length nfft/2 + 1 and is computed over the interval [0, π]. If nfft is odd, the length of S is (nfft + 1)/2 and the frequency interval is [0,π). When your specify fs, the intervals are [0, fs/2) and [0, fs/2] for even and odd nfft, respectively.

• 'whole' — Returns the whole spectrum for either real or complex input x. In this case, S has length nfft and is computed over the interval [0, 2π). When you specify fs, the frequency interval is [0, fs).

• 'centered' — Returns the centered whole spectrum for either real or complex input x. In this case, S has length nfft and is computed over the interval (–π, π] for even nfft and (–π, π) for odd nfft. When you specify fs, the frequency intervals are (–fs/2, fs/2] and (–fs/2, fs/2) for even and odd nfft, respectively.

Note

You can put the arguments freqrange or 'corr' anywhere in the input argument list after p.

## Output Arguments

collapse all

Pseudospectrum estimate, returned as a vector. The pseudospectrum is calculated using estimates of the eigenvectors of a correlation matrix associated with the input data x.

Output normalized frequencies, specified as a vector. S and wo have the same length. In general, the length of the FFT and the values of the input x determine the length of the computed S and the range of the corresponding normalized frequencies. The table indicates the length of S (and wo) and the range of the corresponding normalized frequencies for the first syntax.

S Characteristics for an FFT Length of 256 (Default)

Input Data TypeLength of S and w0Range of the Corresponding Normalized Frequencies

Real

129

[0, π]

Complex

256

[0, 2π)

If nfft is specified, the following table indicates the length of S and wo and the frequency range for wo.

S and Frequency Vector Characteristics

Input Data Typenfft Even or OddLength of S and wRange of w

Real

Even

(nfft/2 )+ 1

[0, π]

Real

Odd

(nfft + 1)/2

[0, π)

Complex

Even or odd

nfft

[0, 2π)

Output frequency, returned as a vector. The frequency range for fo depends on nfft, fs, and the values of the input x. The length of S (and fo) is the same as in the S and Frequency Vector Characteristics above. The following table indicates the frequency range for fo if nfft and fs are specified.

S and Frequency Vector Characteristics with fs Specified

Input Data Type

nfft Even/Odd

Range of f

Real

Even

[0, fs/2]

Real

Odd

[0, fs/2)

Complex

Even or odd

[0, fs)

Additionally, if nwin and noverlap are also specified, the input data x is segmented and windowed before the matrix used to estimate the correlation matrix eigenvalues is formulated. The segmentation of the data depends on nwin, noverlap, and the form of x. Comments on the resulting windowed segments are described in the following table.

Windowed Data Depending on x and nwin

form of x

Form of nwin

Windowed Data

Data vector

Scalar

Length is nwin.

Data vector

Vector of coefficients

Length is length(nwin).

Data matrix

Scalar

Data is not windowed.

Data matrix

Vector of coefficients

length(nwin) must be the same as the column length of x, and noverlap is not used.

See the Eigenvector Length Depending on Input Data and Syntax for related information on this syntax.

Noise eigenvectors, returned as a matrix. The columns of v span the noise subspace of dimension size(v,2). The dimension of the signal subspace is size(v,1)-size(v,2).

Estimated eigenvalues of the correlation matrix, returned as a vector.

## Algorithms

The eigenvector method estimates the pseudospectrum from a signal or a correlation matrix using a weighted version of the MUSIC algorithm derived from Schmidt's eigenspace analysis method [1] [2]. The algorithm performs eigenspace analysis of the signal's correlation matrix to estimate the signal's frequency content. If you do not supply the correlation matrix, the eigenvalues and eigenvectors of the signal's correlation matrix are estimated using svd. This algorithm is particularly suitable for signals that are the sum of sinusoids with additive white Gaussian noise.

The eigenvector method produces a pseudospectrum estimate given by

${P}_{\text{ev}}\left(f\right)=\frac{1}{\sum _{k=p+1}^{N}|{\text{v}}_{k}^{H}e\left(f\right){|}^{2}/{\lambda }_{k}}$

where N is the dimension of the eigenvectors and vkis the kth eigenvector of the correlation matrix of the input signal. The integer p is the dimension of the signal subspace, so the eigenvectors vk used in the sum correspond to the smallest eigenvalues λk of the correlation matrix. The eigenvectors used span the noise subspace. The vector e(f) consists of complex exponentials, so the inner product vkHe(f) amounts to a Fourier transform. This is used for computation of the pseudospectrum. The FFT is computed for each vk and then the squared magnitudes are summed and scaled.

## References

[1] Marple, S. Lawrence. Digital Spectral Analysis. Englewood Cliffs, NJ: Prentice-Hall, 1987, pp. 373–378.

[2] Schmidt, R. O. “Multiple Emitter Location and Signal Parameter Estimation.” IEEE® Transactions on Antennas and Propagation. Vol. AP-34, March, 1986, pp. 276–280.

[3] Stoica, Petre, and Randolph L. Moses. Spectral Analysis of Signals. Upper Saddle River, NJ: Prentice Hall, 2005.

## See Also

Introduced before R2006a

## Support

#### Deep Learning for Signal Processing with MATLAB

Download white paper