CN103546221A - Method for estimating direction of arrival of bandwidth coherent signal - Google Patents

Abstract

The invention provides a method for estimating the direction of arrival of a bandwidth coherent signal. The method for estimating the direction of arrival of the bandwidth coherent signal comprises the following steps that covariance matrixes of different frequency points are transformed through a transformation matrix, and a guide covariance matrix is established; a propagator and noise subspace are estimated through the guide covariance matrix, and orthogonalization is conducted on the noise subspace; a spatial spectrum is calculated according to a formula shown in the specification according to the MUSIC algorithm, and the direction of arrival is estimated. Compared with an existing method for estimating the DOA of the bandwidth coherent signal based on the propagator method, the method for estimating the direction of arrival of the bandwidth coherent signal has the advantages of being suitable for various formations, free of direction estimation, good in statistics stability, capable of improving the performance of estimations of the DOA of the bandwidth coherent signal under the condition of low signal to noise ratio and a small number of snapshots, applicable to estimation of the DOA of the bandwidth coherent signal of a sonar system, a radar system and a wireless communication system.

Description

A kind of broadband coherent signal direction of arrival estimation method

Technical field

The present invention relates to a kind of broadband coherent signal ripple and reach angle (Direction of Arrival is called for short DOA) method of estimation.

Background technology

In various high resolution DOA estimation algorithms, the Subspace Decomposition class algorithm that the MUSIC algorithm of take is representative is because its resolving power is high, stable performance, be subject to extensive concern, yet this class algorithm needs pair array data covariance matrix to carry out Eigenvalues Decomposition in order to obtain noise subspace or signal subspace, operand is larger, and the software and hardware system that DOA is estimated is had relatively high expectations.Propagation operator method (PM) only can obtain the estimation of noise subspace by simple linear operation without Eigenvalues Decomposition, has reduced operand.Basic propagation operator method is for arrowband unrelated signal, and under middle high s/n ratio, performance approaches MUSIC algorithm, but under low signal-to-noise ratio hydraulic performance decline.Propagation operator method (PM) based on coherent signal subspace method (CSM) can realize the DOA estimation of broadband coherent signal, but need to estimate orientation, and estimate accuracy affects DOA estimated performance.

Summary of the invention

Technical problem to be solved by this invention is, overcome the deficiencies in the prior art, propose a kind of broadband DOA Estimation in Coherent Signal method based on propagation operator (PM) that does not need to estimate orientation, and can improve the DOA estimated performance of PM algorithm under low signal-to-noise ratio and fewer snapshots condition.

The invention provides a kind of broadband DOA Estimation in Coherent Signal method that can be used for sonar, radar and wireless telecommunication system, utilize the covariance matrix R of transformation matrix to different frequent points place _x(f _j) convert, structure guiding covariance matrix, utilizes guiding covariance matrix propagation operator and noise subspace, and computer memory is composed and is carried out DOA estimation.

Concrete grammar of the present invention is:

Step 1, utilize the covariance matrix R of transformation matrix to different frequent points place _x(f _j) convert structure guiding covariance matrix:

$Z\left(\mathrm{\θ}\right)=\underset{j=1}{\overset{J}{\mathrm{\Σ}}}{T}_j\left(\mathrm{\θ}\right)R_x\left(f_j\right)T_j^H\left(\mathrm{\θ}\right)$

Wherein, transformation matrix is:

ω wherein _j=2 π f _j, f _jit is the frequency of j frequency; τ _m(θ), 1≤m≤M, for inciding the signal of m array element with respect to the time delay of reference point from azimuth angle theta, M is array number; J is frequency number; H represents conjugate transpose;

Step 2, utilization guiding covariance matrix propagation operator and noise subspace, and carry out noise subspace orthogonalization; Step 3, adopt MUSIC algorithm to be composed by following formula computer memory and carry out ripple and reach angle and estimate:

$\hat{P}\left(\mathrm{\θ}\right)=\frac{1}{1^H{\hat{Q}}_0{\hat{Q}}_0^{H}1}$

1=[1 wherein, 1 ..., 1] ^tfor the guiding vector of M * 1 dimension, T represents transposition.

Described estimation propagation operator and noise subspace are adopted with the following method:

1) to guiding covariance matrix Z (θ) piecemeal;

$Z\left(\mathrm{\θ}\right)=\left[\begin{array}{cc}\hat{G}& \hat{H}\end{array}\right]$

Wherein for the matrix of M * N, matrix for M * (M-N);

2) calculate propagation operator;

$\hat{P}={\left({\hat{G}}^H\hat{G}\right)}^{-1}{\hat{G}}^H\hat{H}$

3) structure noise subspace;

${\hat{Q}}^H=[{\hat{P}}^H,-I]$

Wherein I is the unit matrix of M-N dimension;

4) to noise subspace orthogonalization:

Compare with the existing broadband DOA Estimation in Coherent Signal method based on propagation operator method, the present invention is suitable for any formation, do not need to estimate orientation, there is good statistics stability, be especially applicable to the broadband DOA Estimation in Coherent Signal of fast umber of beats less and under Low SNR.

Accompanying drawing explanation

Fig. 1 is the broadband DOA Estimation in Coherent Signal method flow diagram of the embodiment of the present invention.

Embodiment

Basic ideas of the present invention are to replace data covariance matrix with the guiding covariance matrix of array signal, utilize guiding covariance matrix propagation operator, structure noise subspace, and carry out DOA estimation.

Propagation operator method (PM) can obtain the estimation of noise subspace by simple linear operation without Eigenvalues Decomposition, has reduced the operand of Estimation of Spatial Spectrum.

Below in conjunction with Fig. 1 and embodiment, describe the present invention.

Fig. 1 is the broadband DOA Estimation in Coherent Signal method flow diagram of the embodiment of the present invention.As shown in Figure 1, the method comprises the steps:

1) reception data are divided into K section, every section of J point, does FFT conversion and obtains frequency domain reception data matrix;

$\left[\begin{array}{c}{x}_1\left(f_j\right)\\ x_2\left(f_j\right)\\ .\\ .\\ .\\ x_M\left(f_j\right)\end{array}\right]=\left[\begin{array}{cccc}{e}^{-j{\mathrm{\ω}}_j{\mathrm{\τ}}_{11}}& e^{-j{\mathrm{\ω}}_j{\mathrm{\τ}}_{12}}& ...& e^{-j{\mathrm{\ω}}_j{\mathrm{\τ}}_{1N}}\\ e^{-j{\mathrm{\ω}}_j{\mathrm{\τ}}_{21}}& e^{-j{\mathrm{\ω}}_j{\mathrm{\τ}}_{22}}& ...& e^{-j{\mathrm{\ω}}_j{\mathrm{\τ}}_{2N}}\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ e^{-j{\mathrm{\ω}}_j{\mathrm{\τ}}_{M1}}& e^{-j{\mathrm{\ω}}_j{\mathrm{\τ}}_{M2}}& ...& e^{-j{\mathrm{\ω}}_j{\mathrm{\τ}}_{\mathrm{MN}}}\end{array}\right]\left[\begin{array}{c}{s}_1\left(f_j\right)\\ s_2\left(f_j\right)\\ .\\ .\\ .\\ s_N\left(f_j\right)\end{array}\right]+\left[\begin{array}{c}{n}_1\left(f_j\right)\\ n_2\left(f_j\right)\\ .\\ .\\ .\\ n_M\left(f_j\right)\end{array}\right],1\≤j\≤J$

Wherein M is array number, and N is information source number, and information source and noise are separate, ω _j=2 π f _j, f _jbe the frequency of j frequency, τ _mnit is the time delay that n signal arrives m array element relative reference point.Above formula can be write as the form of matrix:

X(f _j)＝A(f _j)S(f _j)+N(f _j) 1≤j≤J

A (f wherein _j) be array manifold matrix.

2) estimate the data covariance matrix R of each subband _x(f _j);

$R_x\left(f_j\right)=\frac{1}{K}\underset{k=1}{\overset{K}{\mathrm{\Σ}}}X\left(f_j\right)X^H\left(f_j\right)$

Wherein K is data hop count, X (f _j) be frequency f _jcorresponding reception data matrix, H represents conjugate transpose.

3) calculate guiding covariance matrix Z (θ);

$Z\left(\mathrm{\θ}\right)=\underset{j=1}{\overset{J}{\mathrm{\Σ}}}{T}_j\left(\mathrm{\θ}\right)R_x\left(f_j\right)T_j^H\left(\mathrm{\θ}\right)$

T wherein _j(θ) be transformation matrix,

4) estimate propagation operator;

To guiding covariance matrix Z (θ) piecemeal,

$Z\left(\mathrm{\θ}\right)=\left[\begin{array}{cc}\hat{G}& \hat{H}\end{array}\right]$

Wherein

for the matrix of M * N,

matrix for M * (M-N).

Estimate propagation operator

$\hat{P}={\left({\hat{G}}^H\hat{G}\right)}^{-1}{\hat{G}}^H\hat{H}$

5) structure noise subspace, and orthogonalization;

${\hat{Q}}^H=[{\hat{P}}^H,-I]$

${\hat{Q}}_0=\hat{Q}{\left({\hat{Q}}^H\hat{Q}\right)}^{-1/}$

Wherein I is the unit matrix of M-N dimension.

6) adopt MUSIC method estimation space spectrum, search spectrum peak obtains DOA and estimates;

$\hat{P}\left(\mathrm{\θ}\right)=\frac{1}{1^H{\hat{Q}}_0{\hat{Q}}_0^{H}1}$

1=[1 wherein, 1 ..., 1] ^tguiding vector for M * 1 dimension.

Compare with the existing broadband DOA Estimation in Coherent Signal method based on propagation operator method, the present invention is suitable for any formation, do not need to estimate orientation, there is good statistics stability, can effectively improve the DOA estimated performance of broadband coherent signal under low signal-to-noise ratio and fewer snapshots condition.

In addition, DOA method of estimation amount of calculation based on propagation operator is proportional to the ratio of information source number and array number, extremely be conducive to the many and real-time direction finding of the Underwater Acoustic Environment that information source number is less of array element, also can be used for the broadband DOA Estimation in Coherent Signal of radar and wireless telecommunication system simultaneously.

Claims (2)

1. a broadband coherent signal direction of arrival estimation method, is characterized in that, step is as follows:

Step 1, utilize the covariance matrix R of transformation matrix to different frequent points place _x(f _j) convert structure guiding covariance matrix:

Wherein, transformation matrix is:

ω wherein _j=2 π f _j, f _jit is the frequency of j frequency; τ _m(θ), 1≤m≤M, for inciding the signal of m array element with respect to the time delay of reference point from azimuth angle theta, M is array number; J is frequency number; H represents conjugate transpose;

Step 2, utilization guiding covariance matrix propagation operator and noise subspace, and carry out noise subspace orthogonalization;

Step 3, adopt MUSIC algorithm to be composed by following formula computer memory and carry out ripple and reach angle and estimate:

1=[1 wherein, 1 ..., 1] ^tfor the guiding vector of M * 1 dimension, T represents transposition.

2. according to right, want the broadband coherent signal direction of arrival estimation method described in 1, it is characterized in that, described estimation propagation operator and noise subspace are adopted with the following method:

1) to guiding covariance matrix Z (θ) piecemeal;

Wherein

for the matrix of M * N,

matrix for M * (M-N);

2) calculate propagation operator;

3) structure noise subspace;

Wherein I is the unit matrix of M-N dimension;

4) to noise subspace orthogonalization;

。

Images