5
$\begingroup$

Setup

Suppose we have a complex $L\times 1$ signal $\mathbf{x}$ with two tones at (unknown) frequencies and phases defined as: $$ x_n = A_1 e^{j \omega_1n + \varphi_1} + A_2 e^{j \omega_2n + \varphi_2} $$ for $0 \leq n \leq L-1$. We observe $\mathbf{x}$ corrupted by additive white Gaussian noise: $$ \mathbf{y} = \mathbf{x} + \mathbf{w} $$ where $\mathbf{w} \sim \mathcal{CN}(0, \sigma^2 \mathbf{I})$.

Problem

I'm trying to show a connection between the signal amplitudes and eigenvalues from the eigendecomposition of the autocovariance matrix $\mathbf{R_y}$ of $\mathbf{y}$ (this is related to MUltiple SIgnal Classification i.e. MUSIC algorithm which splits the $L$ dimensional space into a 2 dimensional signal subspace and an $L-2$ dimensional noise subspace).

$$ \mathbf{R_y} := \mathbf{E}[\mathbf{y}\mathbf{y}^H] = \mathbf{x}\mathbf{x}^H + \sigma^2 \mathbf{I}. $$

Let $\mathbf{s}_1 = \begin{bmatrix} 1 & e^{j\omega_1} & e^{j2\omega_1} & \ldots & e^{j(L-1)\omega_1}\end{bmatrix}^T$ and similarly define $\mathbf{s}_2$. Then, $$ \mathbf{x} = \begin{bmatrix} \mathbf{s}_1 & \mathbf{s}_2 \end{bmatrix} \begin{bmatrix} A_1 e^{j\varphi_1} \\ A_2 e^{j\varphi_2} \end{bmatrix} $$ and the signal autocovariance matrix can be written as: \begin{eqnarray*} \mathbf{x}\mathbf{x}^H &=& \begin{bmatrix} \mathbf{s}_1 & \mathbf{s}_2 \end{bmatrix} \begin{bmatrix} A_1 e^{j\varphi_1} \\ A_2 e^{j\varphi_2} \end{bmatrix} \begin{bmatrix} A_1 e^{-j\varphi_1} & A_2 e^{-j\varphi_2} \end{bmatrix} \begin{bmatrix} \mathbf{s}_1^H \\ \mathbf{s}_2^H \end{bmatrix} \\ &=& \underbrace{\begin{bmatrix} \mathbf{s}_1 & \mathbf{s}_2 \end{bmatrix}}_{\mathbf{S}} \underbrace{\begin{bmatrix} A_1^2 & A_1 A_2 e^{j(\varphi_1-\varphi_2)} \\ A_1 A_2 e^{-j(\varphi_1-\varphi_2)} & A_2^2 \end{bmatrix}}_{\mathbf{M}} \underbrace{\begin{bmatrix} \mathbf{s}_1^H \\ \mathbf{s}_2^H \end{bmatrix}}_{\mathbf{S}^H} \end{eqnarray*} where we note that $\mathbf{M}$ is Hermitian symmetric and therefore has a decomposition $\mathbf{M} = \mathbf{Q\Lambda Q}^H$ where $\mathbf{\Lambda}$ is a 2x2 diagonal matrix and $\mathbf{Q}^H\mathbf{Q} = \mathbf{Q}\mathbf{Q}^H = \mathbf{I}$.

Therefore, \begin{eqnarray*} \mathbf{R_y} &=& \mathbf{SQ \Lambda Q^HS^H} + \sigma^2 \mathbf{I}. \end{eqnarray*}

Now I am stuck. I suspect I should be able to eventually express $\mathbf{R_y} = \mathbf{PDP^H}$ where the diagonal matrix $\mathbf{D}$ should have two large eigenvalues related to the signal amplitudes $A_1$ and $A_2$, but since $\mathbf{SS^H}\neq \mathbf{I}$ I have no way to absorb the $\sigma^2$'s into the diagonal matrix $\mathbf{\Lambda}$.

$\endgroup$
1
$\begingroup$

The piece I was missing was the distribution of the initial phase values $\varphi_1$ and $\varphi_2$. It is standard to assume that these are uniformly distributed [^]. This leads to: $$ \mathbf{R_x} = \mathbf{E[xx^H]} = \mathbf{S \Lambda S^H} $$ where $\mathbf{\Lambda} = \begin{bmatrix} A_1^2 & 0 \\ 0 & A_2^2 \end{bmatrix}$ and $\mathbf{S} = \begin{bmatrix}\mathbf s_1 & \mathbf s_2 \end{bmatrix}$ so that $$ \mathbf{R_y} = \mathbf{S \Lambda S^H} + \sigma^2 \mathbf{I}. $$

Now note that $\mathbf{R_y}$ is Hermitian symmetric and must possess an eigendecomposition with eigenvalues $\lambda_1 \geq \lambda_2 \geq \lambda_3 \geq \cdots \geq \lambda_L$. However, also note that $\mathbf{R_x}$ has two nonzero eigenvalues, say, $\tilde\lambda_1 \geq \tilde\lambda_2$. Since $\mathbf{R_y} = \mathbf{R_x} + \sigma^2 I$ it is easy to verify that $\lambda_1 = \tilde \lambda_1 + \sigma^2$, $\lambda_2 = \tilde \lambda_2 + \sigma^2$ and $\lambda_i = \sigma^2$ for $3\leq i \leq L$.

In the eigendecomposition of $\mathbf{R_y}$, let $\mathbf e_1$ and $\mathbf e_2$ be the eigenvectors corresponding to $\lambda_1$ and $\lambda_2$ and $\mathbf g_1, \ldots, \mathbf g_{L-2}$ be the eigenvectors corresponding to the eigenvalues $\lambda_3, \lambda_4,\ldots, \lambda_L$. One can now show that $\mbox{span}\{\mathbf e_1, \mathbf e_2\} = \mbox{span}\{\mathbf s_1, \mathbf s_2\}$. This is often called the "signal subspace" and the orthogonal subspace $\mbox{span}\{ \mathbf g_1,\ldots \mathbf g_{L-2}\}$ is called the "noise subspace."

Here's the punchline: To estimate $\omega_1$ and $\omega_2$ we need to solve for $\omega$ in $\mathbf s(\omega)^H \mathbf G \mathbf G^H \mathbf s(\omega) = 0$, where $\mathbf G = \begin{bmatrix} \mathbf g_1 & \ldots & \mathbf g_{L-2}\end{bmatrix}$, and $\mathbf s(\omega) = \begin{bmatrix} 1 & e^{j\omega} & e^{j2\omega} & \ldots & e^{j(L-1)\omega} \end{bmatrix}$.

Reference

[^] P. Stoica, R. Moses, Spectral Analysis of Signals 1st Ed. Ch. 4.

$\endgroup$
0
$\begingroup$

I can't fully agree with what you imply for $\mathrm E\left[\mathbf{xx^H}\right]$.

From what I roughly recall upon skimming old literature [1, p. 20f] If your input signal is harmonic, the autocorrelation (or the autocovariance) should have diagonal shape – that personally feels intuitively right (... if you consider the the complex sinusoids as a base of the space of power-limited functions, which is exactly what the Fourier transform does to a signal represented with "time" as base, then it becomes obvious that different sinusoids signals are orthogonal, for they have zero coefficients in all but one, different, entry).


[1] S. Ehrhard, M. Fischer, M. Fuhr, M. Mouazzen, M. Müller, and M. L. Schulz, “Spektralschätzung mit MUSIC und ESPRIT,” Communications Engineering Lab (Institut für Nachrichtentechnik), Tech. Rep., Mar. 2011. Available via OpenAccess.

$\endgroup$
  • $\begingroup$ I think the situation is different here because we are taking the outer product $\mathbf{x}\mathbf{x}^H$, so I cannot use the orthogonality of the Fourier basis (which would require taking inner products of the columns of $\mathbf{S}$). $\endgroup$ – Atul Ingle Aug 15 '16 at 12:52
  • $\begingroup$ I don't agree. The autocorrelation matrix of $\mathbf x$ should be diagonal. $\endgroup$ – Marcus Müller Aug 15 '16 at 13:00
  • $\begingroup$ I still don't see it mathematically though - I double checked my derivation above and still can't rid of the off-diagonal terms of $\mathbf{M}$. $\endgroup$ – Atul Ingle Aug 15 '16 at 13:03
  • $\begingroup$ I agree, but that means that $\mathbf {xx^H}$ can't be $\mathbf{E[yy^H]}$ in case of zero noise, can it :) ? $\endgroup$ – Marcus Müller Aug 15 '16 at 13:13
  • $\begingroup$ In case of zero noise, $\sigma^2\mathbf{I}=0$ so that $\mathbf{R_y} = \mathbf{xx^H}$. $\endgroup$ – Atul Ingle Aug 15 '16 at 13:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.