A robust adaptive beamforming method under correlated interference sources

Abstract

This invention discloses a robust adaptive beamforming method under correlated interference sources, comprising: eigenvalue decomposition of the sampling covariance matrix and partitioning the eigenvectors into a signal subspace and a noise subspace; estimation of noise power based on eigenvalues; estimation of the initial angle between the desired signal and interference using subspace orthogonality, obtaining the nominal steering vector corresponding to the initial angle based on the known array shape, and projecting it onto the noise subspace to obtain the error vector of the nominal steering vector; updating and estimating the steering vector based on subspace orthogonality to obtain the steering vectors of the desired signal and interference; reconstructing the interference plus noise covariance matrix based on the estimated desired signal and interference steering vectors and the desired signal and interference portions of the sampling covariance matrix; and beamforming the array received data based on the estimated desired signal steering vector and the reconstructed interference plus noise covariance matrix.

Description

Technical Field

[0001] The present invention relates to the research field of beamforming in the field of array signal processing, and in particular to a robust adaptive beamforming method under correlated interference sources. Background Art

[0002] Although the traditional beamforming method can improve the robustness of the beamformer to a certain extent, when the desired signal appears in the covariance matrix, the beamformer will suppress the desired signal as interference, resulting in "self-cancellation" phenomenon, causing the performance of the beamformer to deteriorate. Although the current covariance matrix reconstruction-based algorithms can improve the performance of the beamformer to a certain extent, the estimation of the desired signal steering vector is often achieved by solving convex optimization problems, which has high complexity and many redundant components in the reconstructed covariance matrix. In addition, most of the current covariance matrix reconstruction-based algorithms use uncorrelated interference sources as assumptions, while in practice, more correlated interference sources appear. The performance of this type of algorithm is limited under the condition of correlated interference sources. In view of this, it is necessary to study new beamforming algorithms in the scenario of correlated interference sources to improve the performance of beamformers in practical applications. Summary of the invention

[0003] The purpose of the present invention is to provide a robust adaptive beamforming method under correlated interference sources, which further improves the performance of the beamformer under the conditions of actual array errors and the presence of correlated interference sources by estimating the steering vectors and interference power matrices of the desired signal and the correlated interference sources.

[0004] The objective of the present invention is achieved through the following technical solution: a robust adaptive beamforming method under correlated interference sources, comprising:

[0005] Step 1, perform eigendecomposition on the sampling covariance matrix of the array received data, call the eigenvectors corresponding to the eigenvalues ​​of the desired signal and the interference as the signal subspace, and call the eigenvectors corresponding to the remaining eigenvalues ​​as the noise subspace; obtain an estimated value of the noise power according to the eigenvalues ​​corresponding to the noise subspace;

[0006] Step 2: According to the orthogonality of the signal subspace and the noise subspace in step 1, the spectrum value of the whole space is obtained, the initial angle of the desired signal and the interference is obtained by searching the spectrum value for the peak, and the nominal steering vector corresponding to the initial angle is obtained according to the known array shape; the nominal steering vector of the desired signal and the interference is projected onto the noise subspace in step 1 to obtain the error vector of the desired signal and the interference steering vector;

[0007] Step 3: Using the nominal steering vectors of the desired signal and interference and the error vector in step 2, based on the subspace orthogonality principle, update and estimate the steering vectors of the desired signal and interference to obtain estimated values ​​of the steering vectors of the desired signal and interference;

[0008] Step 4, reconstructing the interference plus noise covariance matrix according to the noise power estimate in step 1, the estimated value of the desired signal and the interference steering vector in step 3, and the desired signal and interference part in the sampling covariance matrix of the array received data, to obtain a reconstructed interference plus noise covariance matrix;

[0009] Step 5: Based on the desired signal steering vector estimated in step 3 and the interference plus noise covariance matrix reconstructed in step 4, beamforming is performed on the array received data.

[0010] Furthermore, the sampling covariance matrix of the array received data is eigen-decomposed, and the eigenvectors corresponding to the eigenvalues ​​of the desired signal and the interference are called signal subspaces, and the eigenvectors corresponding to the remaining eigenvalues ​​are called noise subspaces; the estimated values ​​of the noise power obtained according to the eigenvalues ​​corresponding to the noise subspace include:

[0011] The method provided by the present invention is applicable to any array. For the convenience of description, an M-element uniform linear array is taken as an example, the array element spacing is d, and L far-field narrowband signal sources are received, including a desired signal and L-1 related interference sources. The desired signal and the related interference sources are not correlated with each other. Then, the received data of the array at time k can be expressed as:

[0012]

[0013] Where s0(k) represents the waveform of the desired signal at time k, a0 represents the steering vector of the desired signal, l=1,2,…,L-1 represents the waveform of the lth related interference source at time k, a l represents the steering vector of the lth related interference source, x n (k) represents the noise received by the array at time k. Assume that the mean values ​​of the desired signal, interference, and noise are all zero.

[0014] Furthermore, the relevant interference source can be expressed as:

[0015] s c =Γs uc

[0016] in, represents the relevant interference source vector received by the array, Represents the uncorrelated interference source vector, and Γ represents the coefficient matrix of the correlated interference source. When Γ is a diagonal matrix, the correlated interference source degenerates into an uncorrelated interference source. T represents the transpose operation of the matrix or vector. Ideally, the covariance matrix of the array receiving data can be expressed as:

[0017]

[0018] in, represents the power of the desired signal, A=[a0,A i ]=[a0,a1,…,a L-1 ] represents the steering vector matrix of the desired signal and interference, A i =[a1,a2,…,a L-1 ] represents the steering vector matrix of the interference source, represents the power matrix of uncorrelated interference sources, diag{} represents the diagonal matrix, l=1,2,…,L-1 represents the power of the lth uncorrelated interference source, ΓΣ uc Γ H represents the power matrix of the relevant interference sources, represents the noise power, I represents the identity matrix, and H represents the conjugate transpose operation of the matrix or vector.

[0019] In step 1, the sampling covariance matrix of the array received data is subjected to eigendecomposition, and the eigenvectors corresponding to the eigenvalues ​​of the desired signal and the interference are called signal subspaces, and the eigenvectors corresponding to the remaining eigenvalues ​​are called noise subspaces; and the estimated value of the noise power is obtained according to the eigenvalues ​​corresponding to the noise subspace, and the specific requirements include:

[0020] The sampling covariance matrix of the array received data is:

[0021]

[0022] Wherein, x(k) represents the data received by the array at time k, K represents the number of snapshots, and H represents the conjugate transpose operation of a matrix or vector.

[0023] Performing eigendecomposition on the sampling covariance matrix yields:

[0024]

[0025] Among them, M represents the number of array elements, represents the eigenvalues ​​in descending order, is the eigenvector corresponding to this eigenvalue. represents the eigenvector matrix, represents a diagonal matrix consisting of eigenvalues; represents the eigenvector matrix corresponding to the first L eigenvalues ​​and is called the signal subspace. represents the diagonal matrix composed of the first L eigenvalues, where L represents the number of desired signals and interferences; The eigenvector matrix corresponding to the remaining ML eigenvalues ​​is called the noise subspace. Represents the diagonal matrix composed of the remaining ML eigenvalues. According to the properties of eigendecomposition, the remaining ML eigenvalues ​​correspond to the power of the noise, and the estimated value of the noise power is expressed as:

[0026]

[0027] The step 2 comprises:

[0028] (21) According to the orthogonality of the signal subspace and the noise subspace in step 1, the spectrum value of the whole space is obtained as:

[0029]

[0030] in, represents the nominal steering vector corresponding to the angle θ, and Θ represents the entire space. By searching the spectrum value for the peak, the initial angles of the desired signal and the interference are obtained as follows: and The nominal steering vector corresponding to the initial angle obtained according to the known array shape is:

[0031] (22) The nominal steering vector of the desired signal and the interference is projected onto the noise subspace in step 1 to obtain the error vector of the desired signal and the interference steering vector:

[0032]

[0033] Among them, ||||2 represents the l2 norm of the matrix or vector.

[0034] The method of using the nominal steering vectors of the desired signal and the interference and the error vector in step 2 to update and estimate the steering vectors of the desired signal and the interference according to the subspace orthogonality principle to obtain estimated values ​​of the desired signal and the interference steering vectors includes:

[0035] According to the orthogonality of the signal subspace and the noise subspace, it can be seen that when the steering vector is closer to the true steering vector, the steering vector is more orthogonal to the noise subspace, and the corresponding spectrum value is larger, that is, the denominator in the spectrum value is smaller. Based on this principle, the result of steering vector estimation can be obtained as follows:

[0036]

[0037] in, bi represents the i-th value in the interval [-b, b] with an interval of η, b represents the upper bound of the search interval, Indicates the total number of searches.

[0038] The reconstructing of the interference plus noise covariance matrix according to the noise power estimation value in step 1, the estimation value of the desired signal and the interference steering vector in step 3, and the desired signal and interference part in the sampling covariance matrix of the array received data, to obtain the reconstructed interference plus noise covariance matrix includes:

[0039] For the covariance matrix of the ideal array received data, it can be expressed as:

[0040]

[0041] By performing feature decomposition, we can obtain:

[0042]

[0043] Among them, γ m represents the eigenvalues ​​in descending order, u m is the eigenvector corresponding to the eigenvalue. S U N ]=[u1,u2,…,u M ] represents the eigenvector matrix, Λ=diag{γ1,γ2,…,γ M} represents a diagonal matrix consisting of eigenvalues, U S =[u1,u2,…,u L ] represents the eigenvector matrix corresponding to the first L eigenvalues, and it is called the signal subspace, Λ S =diag{γ1,γ2,…,γ L} represents the diagonal matrix composed of the first L eigenvalues, U N =[u L+1 ,u L+2 ,…,u M ] is composed of the eigenvector matrix corresponding to the remaining ML eigenvalues, and it is called the noise subspace, Λ N =diag{γ L+1 ,γ L+2 ,…,γ M} represents the diagonal matrix composed of the remaining ML eigenvalues. Then we can get:

[0044]

[0045] Right now:

[0046]

[0047] in, represents the pseudoinverse of the matrix A.

[0048] The noise power estimated Steering vector matrix The sampling covariance matrix of the received data of the array can be used to obtain the power matrix of the desired signal and interference:

[0049]

[0050] Then, the reconstructed interference plus noise covariance matrix can be expressed as:

[0051]

[0052] The beamforming of the array received data based on the desired signal steering vector estimated in step 3 and the interference plus noise covariance matrix reconstructed in step 4 includes:

[0053] According to the reconstructed interference plus noise covariance matrix and the estimated steering vector of the desired signal, the weight vector of the beamformer is expressed as:

[0054]

[0055] in, represents the estimated desired signal steering vector.

[0056] The data received by the array is beamformed according to the weight vector:

[0057] y(k)=w H x(k)

[0058] Where y(k) represents the output signal of the beamformer, thereby enhancing the desired signal and suppressing interference and noise.

[0059] The advantages of the present invention compared with the prior art are:

[0060] (1) The present invention is based on the orthogonality of the signal subspace and the noise subspace, and obtains the initial angles of the desired signal and interference by searching the spectrum peaks. Compared with the Capon power spectrum in the traditional beam, the method provided by the present invention has better angle resolution, that is, it can obtain more accurate initial values ​​of the desired signal and interference angles.

[0061] (2) The present invention uses the projection of the nominal steering vector on the noise subspace as the error vector and the orthogonality of the subspace as the estimation criterion to estimate the steering vector, thereby avoiding the solution of the convex optimization problem for steering vector estimation in traditional beamforming, reducing the computational complexity, and being more conducive to implementation in practice.

[0062] (3) The present invention realizes the estimation of the interference power matrix by sampling the desired signal and interference parts of the covariance matrix and the estimated steering vector matrix, thereby improving the performance of the present invention under the condition of correlated interference sources. Since the uncorrelated interference source is a special case of the correlated interference source, the method of the present invention is also applicable to the adaptive beamforming of the uncorrelated interference source, that is, the present invention can simultaneously handle the beamforming problems of the correlated interference source and the uncorrelated interference source.

[0063] The differences between the present invention and the prior art are: the processing objects are different, the present invention processes the situation of correlated interference sources, while the prior art processes uncorrelated interference sources; the processing methods are different: the present invention constructs the error vector of the steering vector based on the orthogonality of the subspace, and estimates the steering vector based on the maximum orthogonality of the noise subspace, thereby reducing the computational complexity of the method of the present invention; the estimation of the power matrix of the correlated interference source is realized by using the expected signal and interference components in the sampled covariance matrix, and the interference plus noise covariance matrix is ​​reconstructed, which is more suitable for processing practical problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0065] Figure 1 A flowchart of a robust adaptive beamforming method under a correlated interference source of the present invention;

[0066] Figure 2 A schematic diagram of a linear array signal receiving model provided by an embodiment of the present invention;

[0067] Figure 3 A performance curve diagram provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0068] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] like Figure 1 As shown, the embodiment of the present invention provides a robust adaptive beamforming method under a correlated interference source, which mainly includes the following steps:

[0070] Step 1, perform eigendecomposition on the sampling covariance matrix of the array received data, call the eigenvectors corresponding to the eigenvalues ​​of the desired signal and the interference as the signal subspace, and call the eigenvectors corresponding to the remaining eigenvalues ​​as the noise subspace; obtain an estimated value of the noise power according to the eigenvalues ​​corresponding to the noise subspace;

[0071] Step 2: According to the orthogonality of the signal subspace and the noise subspace in step 1, the spectrum value of the whole space is obtained, the initial angle of the desired signal and the interference is obtained by searching the spectrum value for the peak, and the nominal steering vector corresponding to the initial angle is obtained according to the known array shape; the nominal steering vector of the desired signal and the interference is projected onto the noise subspace in step 1 to obtain the error vector of the desired signal and the interference steering vector;

[0072] Step 3: Using the nominal steering vectors of the desired signal and interference and the error vector in step 2, based on the subspace orthogonality principle, update and estimate the steering vectors of the desired signal and interference to obtain estimated values ​​of the steering vectors of the desired signal and interference;

[0073] Step 4, reconstructing the interference plus noise covariance matrix according to the noise power estimate in step 1, the estimated value of the desired signal and the interference steering vector in step 3, and the desired signal and interference part in the sampling covariance matrix of the array received data, to obtain a reconstructed interference plus noise covariance matrix;

[0074] Step 5: Based on the desired signal steering vector estimated in step 3 and the interference plus noise covariance matrix reconstructed in step 4, beamforming is performed on the array received data.

[0075] like Figure 2 The M-element uniform linear array shown in the figure has an element spacing of d and receives L far-field narrowband signal sources, including a desired signal and L-1 related interference sources. The desired signal and the related interference sources are uncorrelated with each other. The received data of the array at time k can be expressed as:

[0076]

[0077] Where s0(k) represents the waveform of the desired signal at time k, a0 represents the steering vector of the desired signal, l=1,2,…,L-1 represents the waveform of the lth related interference source at time k, a l represents the steering vector of the lth related interference source, x n (k) represents the noise received by the array at time k. Assume that the mean values ​​of the desired signal, interference, and noise are all zero.

[0078] Furthermore, the relevant interference source can be expressed as:

[0079] s c =Γsuc

[0080] in, represents the relevant interference source vector received by the array, Represents the uncorrelated interference source vector, and Γ represents the coefficient matrix of the correlated interference source. When Γ is a diagonal matrix, the correlated interference source degenerates into an uncorrelated interference source. T represents the transpose operation of the matrix or vector. Ideally, the covariance matrix of the array receiving data can be expressed as:

[0081]

[0082] in, represents the power of the desired signal, A=[a0,A i ]=[a0,a1,…,a L-1 ] represents the steering vector matrix of the desired signal and interference, A i =[a1,a2,…,a L-1 ] represents the steering vector matrix of the interference source, represents the power matrix of uncorrelated interference sources, diag{} represents the diagonal matrix, l=1,2,…,L-1 represents the power of the lth uncorrelated interference source, ΓΣ uc Γ H represents the power matrix of the relevant interference sources, represents the noise power, I represents the identity matrix, and H represents the conjugate transpose operation of the matrix or vector.

[0083] The sampling covariance matrix of the array received data is subjected to eigendecomposition, and the eigenvectors corresponding to the eigenvalues ​​of the desired signal and the interference are called signal subspaces, and the eigenvectors corresponding to the remaining eigenvalues ​​are called noise subspaces; and the estimated value of the noise power is obtained according to the eigenvalues ​​corresponding to the noise subspace, including:

[0084] The sampling covariance matrix of the array received data is:

[0085]

[0086] Wherein, x(k) represents the data received by the array at time k, K represents the number of snapshots, and H represents the conjugate transpose operation of a matrix or vector.

[0087] Performing eigendecomposition on the sampling covariance matrix yields:

[0088]

[0089] Among them, M represents the number of array elements, represents the eigenvalues ​​in descending order, is the eigenvector corresponding to this eigenvalue. represents the eigenvector matrix, represents a diagonal matrix consisting of eigenvalues; represents the eigenvector matrix corresponding to the first L eigenvalues ​​and is called the signal subspace. represents the diagonal matrix composed of the first L eigenvalues, where L represents the number of desired signals and interferences; The eigenvector matrix corresponding to the remaining ML eigenvalues ​​is called the noise subspace. Represents the diagonal matrix composed of the remaining ML eigenvalues. According to the properties of eigendecomposition, the remaining ML eigenvalues ​​correspond to the power of the noise, and the estimated value of the noise power is expressed as:

[0090]

[0091] The step 2 comprises:

[0092] (21) According to the orthogonality of the signal subspace and the noise subspace in step 1, the spectrum value of the whole space is obtained as:

[0093]

[0094] in, represents the nominal steering vector corresponding to the angle θ, and Θ represents the entire space. By searching the spectrum value for the peak, the initial angles of the desired signal and the interference are obtained as follows: and The nominal steering vector corresponding to the initial angle obtained according to the known array shape is:

[0095] (22) The nominal steering vector of the desired signal and the interference is projected onto the noise subspace in step 1 to obtain the error vector of the desired signal and the interference steering vector:

[0096]

[0097] Among them, || ||2 represents the l2 norm of the matrix or vector.

[0098] The method of using the nominal steering vectors of the desired signal and the interference and the error vector in step 2 to update and estimate the steering vectors of the desired signal and the interference according to the subspace orthogonality principle to obtain estimated values ​​of the desired signal and the interference steering vectors includes:

[0099] According to the orthogonality of the signal subspace and the noise subspace, it can be seen that when the steering vector is closer to the true steering vector, the steering vector is more orthogonal to the noise subspace, and the corresponding spectrum value is larger, that is, the denominator in the spectrum value is smaller. Based on this principle, the result of steering vector estimation can be obtained as follows:

[0100]

[0101] in, b i represents the i-th value in the interval [-b, b] with an interval of η, b represents the upper bound of the search interval, Indicates the total number of searches.

[0102] The reconstructing of the interference plus noise covariance matrix according to the noise power estimation value in step 1, the estimation value of the desired signal and the interference steering vector in step 3, and the desired signal and interference part in the sampling covariance matrix of the array received data, to obtain the reconstructed interference plus noise covariance matrix includes:

[0103] For the covariance matrix of the ideal array received data, it can be expressed as:

[0104]

[0105] By performing feature decomposition, we can obtain:

[0106]

[0107] Among them, γ m represents the eigenvalues ​​in descending order, u m is the eigenvector corresponding to the eigenvalue. S U N ]=[u1,u2,…,u M ] represents the eigenvector matrix, Λ=diag{γ1,γ2,…,γ M} represents a diagonal matrix consisting of eigenvalues, U S =[u1,u2,…,u L ] represents the eigenvector matrix corresponding to the first L eigenvalues, and it is called the signal subspace, Λ S =diag{γ1,γ2,…,γ L} represents the diagonal matrix composed of the first L eigenvalues, U N =[u L+1 ,u L+2 ,…,u M ] is composed of the eigenvector matrix corresponding to the remaining ML eigenvalues, and it is called the noise subspace, Λ N =diag{γ L+1 ,γ L+2 ,…,γ M} represents the diagonal matrix composed of the remaining ML eigenvalues. Then we can get:

[0108]

[0109] Right now:

[0110]

[0111] in, represents the pseudoinverse of the matrix A.

[0112] The noise power estimated Steering vector matrix The sampling covariance matrix of the received data of the array can be used to obtain the power matrix of the desired signal and interference:

[0113]

[0114] Then, the reconstructed interference plus noise covariance matrix can be expressed as:

[0115]

[0116] The beamforming of the array received data based on the desired signal steering vector estimated in step 3 and the interference plus noise covariance matrix reconstructed in step 4 includes:

[0117] According to the reconstructed interference plus noise covariance matrix and the estimated steering vector of the desired signal, the weight vector of the beamformer is expressed as:

[0118]

[0119] in, represents the estimated desired signal steering vector.

[0120] The data received by the array is beamformed according to the weight vector:

[0121] y(k)=w H x(k)

[0122] Where y(k) represents the output signal of the beamformer, thereby enhancing the desired signal and suppressing interference and noise.

[0123] Figure 3 Given in the figure, under the condition of interference-to-noise ratio of 20 dB, the direction of arrival error obeys a uniform distribution on [-4°, 4°], and the performance curves of different methods with signal-to-noise ratio (SNR) in 200 Monte Carlo experiments. It can be seen from the figure that under the condition of correlated interference sources, the method proposed in the present invention can achieve significant performance improvement and obtain performance close to the theoretical optimal beamformer, which illustrates the effectiveness of the method proposed in the present invention. In addition, the method of the present invention does not involve the solution of any convex optimization problem, so the computational complexity is relatively low, which is more conducive to application in practice. The method of the present invention is also applicable to adaptive beamforming of uncorrelated interference sources.

[0124] like Figure 2 As shown, it is a schematic diagram of a linear array signal receiving model provided by an embodiment of the present invention. Other arrays can be used. The array is composed of M omnidirectional microphones, of which the rightmost is a reference array element, the mth array element, m = 1, 2, ..., M-1, and the distance to the reference array element is d m When a far-field narrowband signal from an angle θ is incident on the array, the distance to the mth array element is d m sinθ will produce a certain time difference, which is reflected as a phase difference, so that the nominal steering vector can be obtained based on the estimated angle and array shape.

[0125] like Figure 3 As shown, a uniform linear array with M=10 elements provided by an embodiment of the present invention, the element spacing is half a wavelength, the desired signal comes from θ0=-5°, the related interference sources come from θ1=-40° and θ2=30°, and the correlation coefficient matrix is ​​Γ=[g1,g2,g3], where the correlation coefficient vectors are g1=[e jπ / 3 ,0.5e jπ ,0.5e jπ / 9 ] T ,g2=[0.8e jπ / 6 ,0.9e j6π / 5 ,0.8e jπ / 2 ] T and g3=[0.6e jπ / 4 ,0.4e jπ / 5 ,0.1e j6π / 5 ] T , and there is an angle estimation error of 3°. In the method of the present invention, b=0.01, η=10 -3 . The method of the present invention is proposed, and the comparison methods are: theoretical optimal beamformer (Optimal), covariance matrix reconstruction method based on maximum entropy spectrum (INCM-MEPS), covariance matrix reconstruction algorithm based on spatial power spectrum sampling (INCM-SPSS), robust adaptive beamforming method based on subspace (INCM-subspace), robust adaptive beamforming method based on projection (INCM-projection), covariance matrix reconstruction method based on simple power estimation (INCM-SIPE), diagonal loading algorithm (LSMI) and worst case performance optimal algorithm (WCB). The error in the direction of arrival obeys a uniform distribution on [-4°, 4°], and the performance curves of different methods with signal-to-noise ratio (SNR) under 200 Monte Carlo experiments are shown.

[0126] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above embodiments can be implemented by software, or by means of software plus necessary general hardware platforms. Based on such understanding, the technical solutions of the above embodiments can be embodied in the form of software products, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0127] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A robust adaptive beamforming method under correlated interference sources, characterized in that: include: Step 1, perform eigendecomposition on the sampling covariance matrix of the array received data, call the eigenvectors corresponding to the eigenvalues ​​of the desired signal and the interference as the signal subspace, and call the eigenvectors corresponding to the remaining eigenvalues ​​as the noise subspace; obtain an estimated value of the noise power according to the eigenvalues ​​corresponding to the noise subspace; Step 2: According to the orthogonality of the signal subspace and the noise subspace in step 1, the spectrum value of the whole space is obtained, the initial angle of the desired signal and the interference is obtained by searching the spectrum value for the peak, and the nominal steering vector corresponding to the initial angle is obtained according to the known array shape; the nominal steering vector of the desired signal and the interference is projected onto the noise subspace in step 1 to obtain the error vector of the desired signal and the interference steering vector; Step 3: Using the nominal steering vectors of the desired signal and interference and the error vector in step 2, based on the subspace orthogonality principle, update and estimate the steering vectors of the desired signal and interference to obtain estimated values ​​of the steering vectors of the desired signal and interference; Step 4, reconstructing the interference plus noise covariance matrix according to the noise power estimation value in step 1, the estimation value of the desired signal and the interference steering vector in step 3, and the desired signal and interference part in the sampling covariance matrix of the array received data, to obtain a reconstructed interference plus noise covariance matrix; Step 5: Based on the desired signal steering vector estimated in step 3 and the interference plus noise covariance matrix reconstructed in step 4, beamforming is performed on the array received data.

2. The robust adaptive beamforming method under correlated interference sources according to claim 1, characterized in that: The step 1 is specifically implemented as follows: The sampling covariance matrix of the array received data is: Where x(k) represents the data received by the array at time k, K represents the number of snapshots, and H represents the conjugate transpose operation of the matrix or vector; Performing eigendecomposition on the sampling covariance matrix yields: Among them, M represents the number of array elements, represents the eigenvalues ​​in descending order, is the eigenvector corresponding to the eigenvalue; represents the eigenvector matrix, represents a diagonal matrix consisting of eigenvalues; represents the eigenvector matrix corresponding to the first L eigenvalues ​​and is called the signal subspace. represents the diagonal matrix composed of the first L eigenvalues, where L represents the number of desired signals and interferences; The eigenvector matrix corresponding to the remaining ML eigenvalues ​​is called the noise subspace. Represents the diagonal matrix composed of the remaining ML eigenvalues. According to the properties of eigendecomposition, the remaining ML eigenvalues ​​correspond to the power of the noise. The estimated value of the noise power is expressed as:

3. The robust adaptive beamforming method under correlated interference sources according to claim 1, characterized in that: The step 2 includes: (21) According to the orthogonality of the signal subspace and the noise subspace in step 1, the spectrum value of the whole space is obtained as: in, represents the nominal steering vector corresponding to the angle θ, and Θ represents the entire space. By searching the spectrum value for the peak, the initial angles of the desired signal and the interference are obtained as follows: and The nominal steering vector corresponding to the initial angle obtained according to the known array shape is: (22) The nominal steering vector of the desired signal and the interference is projected onto the noise subspace in step 1 to obtain the error vector of the desired signal and the interference steering vector: Among them, ||||2 represents the matrix or vector Norm.

4. The robust adaptive beamforming method under correlated interference sources according to claim 1, characterized in that: The estimated values ​​of the desired signal and the interference steering vector in step 3 are: in, b i represents the i-th value in the search interval [-b, b] with an interval of η, b represents the upper bound of the search interval, Indicates the total number of searches.

5. The robust adaptive beamforming method under correlated interference sources according to claim 1, characterized in that: The interference plus noise covariance matrix reconstructed in step 4 is: in, represents the estimated interference steering vector matrix, represents the estimated interference power matrix, and I represents the unit diagonal matrix.

6. The robust adaptive beamforming method under correlated interference sources according to claim 1, characterized in that: In step 5, beamforming is performed on the array received data based on the desired signal steering vector estimated in step 3 and the interference plus noise covariance matrix reconstructed in step 4, including: According to the reconstructed interference plus noise covariance matrix and the estimated steering vector of the desired signal, the weight vector of the beamformer is expressed as: in, represents the estimated desired signal steering vector; The data received by the array is beamformed according to the weight vector: y(k)=w H x(k) Wherein, y(k) represents the output signal of the beamformer, which enhances the desired signal and suppresses interference and noise.

Images