A near-field strong interference suppression method based on spherical wave deconvolution beamforming positioning

Abstract

The application discloses a near-field strong interference suppression method based on spherical wave deconvolution beam forming positioning, which comprises the following steps: performing fine grid division on a near-field scanning area of a linear array; performing spherical wave focusing beam forming on the near-field scanning area and two-dimensional deconvolution beam forming about distance-angle under the condition of unknown sound source position; calculating the distribution entropy characteristics of the deconvolution beam forming result in the angle dimension, and using a preset threshold to correlate the interference positioning of a target; constructing an array flow pattern matrix A of each near-field target l and generating a covariance matrix A; calculating the orthogonal complement space projection matrix of the covariance matrix A, and applying the orthogonal complement space projection matrix to the array element domain covariance matrix C to perform near-field interference suppression; using a far-field plane wave model to perform deconvolution beam forming calculation on the covariance matrix for suppressing near-field interference, so as to obtain passive detection results of the near-field strong interference suppression. The application can realize near-field interference positioning and interference suppression under the condition that the interference position is unknown.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of towed line array passive target detection, and particularly relates to a near-field strong interference suppression method based on spherical wave deconvolution beam forming positioning. BACKGROUND

[0002] With the development of vibration and noise reduction technology, the target noise level, especially the high-frequency noise level, decreases sharply, but the low-frequency noise is difficult to reduce. The towed line array is suitable for underwater acoustic target detection due to its large aperture, low detection frequency, and small interference noise from the ship. The near-field strong interference has a great influence on the target detection performance of the large-aperture towed line array, especially when the azimuths of the target of interest and the near-field strong interference are relatively close.

[0003] The common near-field interference suppression method at present is generally to use the inverse beam forming combined with adaptive filtering to realize interference cancellation. However, the position of the interference target is unknown or inaccurate, and it is difficult to have good interference suppression performance. SUMMARY

[0004] The purpose of the present application is to provide a large-aperture towed line array near-field strong interference suppression method based on spherical wave deconvolution beam forming positioning, to realize accurate positioning of the near-field interference and suppress the near-field interference, and improve the target detection performance.

[0005] To achieve the above purpose, the technical solutions of the present application are as follows:

[0006] A near-field strong interference suppression method based on spherical wave deconvolution beam forming positioning, the method comprising:

[0007] performing fine grid division on the near-field scanning area of the towed line array;

[0008] performing spherical wave focusing beam forming and two-dimensional deconvolution beam forming with respect to distance-angle under the condition of unknown sound source position in the near-field scanning area;

[0009] calculating the distribution entropy feature of the deconvolution beam forming result in the angle dimension, and using a preset threshold to associate the interference positioning of the target;

[0010] constructing an array flow pattern matrix A of each near-field target l and generating a covariance matrix A of the array flow pattern matrix;

[0011] calculating the orthogonal complement space projection matrix of the covariance matrix A, and applying the projection matrix to the element domain covariance matrix C to perform near-field interference suppression;

[0012] using a far-field plane wave model to perform deconvolution beam forming calculation on the covariance matrix for suppressing near-field interference, to obtain the passive detection result of the near-field strong interference suppression.

[0013] Preferably, the fine mesh division of the near-field scanning region of the towed linear array includes:

[0014] The far field and near field of the towed linear array are coarsely divided to obtain the near field scanning region of the towed linear array.

[0015] Based on the required near-field estimation resolution, a fine grid is divided into the near-field scanning region of the towed linear array.

[0016] Preferably, the coarse division of the far field and near field of the towed array adopts the towed array focusing distance division method, including:

[0017] Assuming the target is a far-field plane wave, the scanning region R = [r1, r2, ..., r... n Based on the signal magnitude SL(r) obtained at different focusing distances after array gain obtained by spherical wave beamforming, the signal loss Loss(r) caused by focusing is calculated as Loss(r) = SL - SL(r), where SL is the signal magnitude obtained by plane wave beamforming.

[0018] Compare the signal loss Loss(r), r = 1, 2, ..., n, with the allowable signal loss σ. When Loss(r) ≥ σ and min(Loss(r) - σ), the corresponding r is the focusing distance when the signal loss Loss is closest to σ. Let R be the critical distance between the near and far fields of the linear array. m =r.

[0019] Preferably, the spherical wave focusing beamforming algorithm is expressed as follows:

[0020] ,

[0021] Where P′(r, θ) is the spatial spectrum at a focusing distance of r and a focusing azimuth of θ, c is the wave velocity, d is the element spacing of the towed array, M is the number of elements in the towed array, r is the focusing distance, θ is the focusing azimuth, x0 is the distance difference between the sound signal to each sound channel and to the reference sound channel, j is an imaginary number, f is the signal frequency, a(r, θ) is the driving vector, and τ is the arrival time delay of the signal at a focusing distance of r and a focusing azimuth of θ in each sound channel. T For the transpose, X = [x1(f), x2(f), ..., x M [f] represents the Fourier transform of the signal in the element domain at frequency. f Result.

[0022] Preferably, the two-dimensional deconvolution beamforming with respect to distance and angle in the near-field scanning region under the condition of unknown sound source location includes:

[0023] Calculate the scattering function of the sound source at 90° as a function of the focusing distance r:

[0024] ,

[0025] In the formula, rs is the distance from the sound source to each element of the dragline array, rr is the distance from the scanning point to each element of the dragline array, and a is the reciprocal of the distance from the sound source to each element of the dragline array.

[0026] The deconvolution beamforming result P1 is calculated based on the spatial spectrum P′(r, θ) and the scattering function PSF(r, 90°);

[0027] The spatial spectral distribution at the center distance of P1 is taken as the deconvolution spatial spectrum P at a distance of r. deconv Scanning the distance dimension yields the two-dimensional deconvolution beamforming result P. deconv (r, θ).

[0028] Preferably, the calculation of the distribution entropy characteristics of the deconvolution beamforming result in the angular dimension, and the use of a preset threshold for interference localization of associated targets, includes the following steps:

[0029] The result of two-dimensional deconvolution beamforming P deconv Slice (r, θ) into angular dimensions θ = {θ1, θ2, ..., θ} k ...}, where k is the slice index, calculate P deconv The distribution entropy of the (r, θ) angle-dimensional slice, Entropy(θ) k );

[0030] Determine the distribution entropy value Entropy(θ) k If the value is not greater than the threshold TEntropy of the set distribution entropy, then extract P. deconv θ in (r, θ) k The distance r corresponding to the maximum value of the slice k Identify the scan grid (r) k θ k There is near-field interference at point ).

[0031] Preferably, the formula for calculating the orthogonal complement spatial projection matrix of the covariance matrix A of the array manifold matrix is:

[0032] P = IA T (AA T ) -1 A,

[0033] Where T is the transpose and I is the identity matrix.

[0034] Preferably, the formula for calculating the covariance matrix C of the projection matrix P acting on the element domain is: C = PC.

[0035] Preferably, the deconvolution beamforming calculation of the element domain covariance matrix C for suppressing near-field interference using the far-field plane wave model includes:

[0036] The element-domain covariance matrix C, which suppresses near-field interference, is used to calculate the far-field planar beamforming result ISCBF.

[0037] Using the far-field deconvolution beamforming scattering function PSF(r, 90°) and the far-field planar beamforming result ISCBF as inputs to the RL deconvolution algorithm, the deconvolution beamforming result ISDBF after suppressing near-field interference is obtained.

[0038] Preferably, the RL deconvolution algorithm formula is expressed as:

[0039] ,

[0040] Where i is the number of iterations, r is the scanning distance range, and θ is the scanning angle.

[0041] P(r, θ) represents the range-angle two-dimensional spatial spectrum estimation result obtained from the i-th deconvolution iteration, and P(r, θ) represents the range-angle two-dimensional spatial spectrum estimation obtained from conventional beamforming.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0043] Based on the spherical wave model, this invention utilizes a high-precision deconvolution beamforming method combined with distributed entropy information to achieve near-field target localization in the range-angle dimension. Based on the localization results, an array manifold matrix of strong interference in the near-field model is constructed. Then, the covariance matrix of the received signal is projected onto the orthogonal complement space of the manifold matrix as a new covariance matrix and processed for far-field beamforming. This invention can achieve near-field interference localization and interference suppression even when the interference position is unknown, thereby improving target detection performance. Attached Figure Description

[0044] Figure 1 This is a flowchart of the method of the present invention.

[0045] Figure 2 This is a diagram showing the results of dividing the near-field focusing distance.

[0046] Figure 3 A diagram showing the results of focused beamforming with finely divided grids for the near field.

[0047] Figure 4 This is a diagram showing the result of two-dimensional deconvolution beamforming.

[0048] Figure 5 This is a diagram showing the distribution entropy and near-field target localization results.

[0049] Figure 6 This is the beam pattern after strong interference suppression. Detailed Implementation

[0050] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0051] Referring to Figure 1, a near-field strong interference suppression method based on spherical wave deconvolution beamforming localization includes the following 8 steps.

[0052] Step 1: Use the focusing distance division method of the towed array to coarsely divide the far field and near field of the towed array, and obtain the near field scanning area of ​​the towed array, as shown in Figure 2.

[0053] The process of dividing the far and near fields in this linear array focusing distance division method is as follows:

[0054] Step 1.1, assuming the target is a far-field plane wave, the scanning area R = [r1, r2, ..., r n Based on the signal magnitude SL(r) obtained at different focusing distances after array gain obtained by spherical wave beamforming, the signal loss Loss(r) caused by focusing is calculated as Loss(r) = SL - SL(r), where SL is the signal magnitude obtained by plane wave beamforming.

[0055] Step 1.2: Compare the signal loss Loss(r) with the allowable signal loss σ. When Loss(r) ≥ σ and min(Loss(r) - σ), the corresponding r is the focusing distance when the signal loss Loss is closest to σ. Denote the near-far field critical distance R of the linear array. m =r;

[0056] In step 1, the near and far field critical distance R of the towed linear array is used. m The near-field scanning region can then be obtained as [R]. min R m ], R min = min(R).

[0057] Step 2: Based on the required near-field estimation resolution, perform fine-grained grid division on the near-field scanning area of ​​the linear array.

[0058] Here, the result of this refined grid division is also the beamforming focusing distance [0, R]. m The result of uniform partitioning with a step size of Δr.

[0059] Step 3: According to Equation (1), spherical wave focusing beamforming is performed in the near-field scanning region based on the pairwise domain data to obtain the spatial spectrum P′(r, θ) when the focusing distance is r and the focusing azimuth is θ.

[0060] like Figure 3 As shown,

[0061] (1)

[0062] Where c is the wave velocity, d is the element spacing of the towed array, M is the number of elements in the towed array, r is the focusing distance, θ is the focusing azimuth, x0 is the distance difference between the sound signal to each sound channel and to the reference sound channel, j is an imaginary number, f is the signal frequency, a(r, θ) is the driving vector, and τ is the arrival time delay of the signal at a focusing distance of r and a focusing azimuth of θ in each sound channel. T For the transpose, X = [x1(f), x2(f), ..., x M [f] represents the Fourier transform of the signal in the element domain at frequency. f The result is as follows. In this embodiment, the intermediate acoustic channel is selected as the reference acoustic channel.

[0063] Step 4: Calculate the scattering function PSF(r, 90°) of the sound source at a 90° azimuth as a function of the focusing distance r, and calculate the two-dimensional spatial spectrum estimate P of the deconvolution beamforming distance angle based on the spatial spectrum P′(r, θ) and the scattering function PSF(r, 90°). deconv After (r, θ), slice it into angular dimensions θ = {θ1, θ2, ..., θ}. k , ...}, calculate P deconv The distribution entropy of (r, θ) in the angular dimension is Entropy(θ). k ), k=1, 2, ...

[0064] Here, the formula for calculating the scattering function PSF is shown in equation (2):

[0065] (2)

[0066] Where rs is the distance from the sound source to each element of the drag array, and rr is the distance from the scanning point to each element of the drag array.

[0067] The formula for calculating distribution entropy is based on the methods used in probability theory, which is common knowledge in this field and will not be elaborated upon here.

[0068] The deconvolution beamforming result P is calculated based on the spatial spectrum P′(r, θ) and the scattering function PSF(r, θ). deconv Specifically, (r, θ) involves calculating the deconvolution beamforming result P1 based on the spatial spectrum P′(r, θ) and scattering function PSF(r, θ) obtained in step 3, and using the spatial spectrum distribution at the center distance of P1 as the deconvolution spatial spectrum P at a distance of r. deconv Scanning the distance dimension yields the two-dimensional deconvolution beamforming result P. deconv(r, θ), as shown in Figure 4.

[0069] Step 5: Using a preset threshold based on the distribution entropy characteristics, interference localization of associated targets is performed. Specifically:

[0070] The angle to be determined is θ k Entropy(θ) is the distribution entropy value at a given location. k If the threshold TEntropy for the set distribution entropy is ≤ 0, then it is considered that the angle slice θ is within the range of the set distribution entropy. k There is a near-field interfering target at the location; extract P. deconv θ in (r, θ) k The distance r corresponding to the maximum value of the slice k Identify the scan grid (r) k θ k There is near-field interference at point ), as shown in Figure 5; otherwise, no action is taken.

[0071] Step 6: Based on the identified near-field interference targets, construct the first near-field target array manifold matrix, denoted as A. l The covariance matrix A is obtained by summing the array manifold matrices formed by all targets.

[0072] Step 7: Calculate the projection matrix P of the covariance matrix A of the array manifold matrix onto the orthogonal complement space, apply matrix P to the covariance matrix C of the element domain, update the covariance matrix C, and obtain the covariance matrix that suppresses near-field interference.

[0073] Here, the formula for calculating the projection matrix P of the covariance matrix A of the array manifold matrix onto the orthogonal complement space is:

[0074] P = IA T (AA T ) -1 A (3)

[0075] Where I is the identity matrix.

[0076] Step 8: Use the far-field plane wave model to perform deconvolution beamforming calculations on the covariance matrix for suppressing near-field interference, obtaining the passive detection results for suppressing strong near-field interference. Specifically:

[0077] (8-1) Construct the far-field deconvolution beamforming scattering function PSF, and use the updated covariance matrix C to calculate the far-field planar beamforming result ISCBF;

[0078] (8-2) Using the far-field deconvolution beamforming scattering function PSF and the far-field planar beamforming result ISCBF as inputs to the RL deconvolution algorithm, the deconvolution beamforming result after suppressing near-field interference is obtained; wherein, the formula of the RL deconvolution algorithm is expressed as:

[0079] (4)

[0080] Where i is the number of iterations, r is the scanning distance range, and θ is the scanning angle.

[0081] P(r, θ) represents the range-angle two-dimensional spatial spectrum estimation result obtained from the i-th deconvolution iteration, P(r, θ) represents the range-angle two-dimensional spatial spectrum estimation obtained from conventional beamforming, Ω and Θ represent the integration ranges of the angle dimension and the range dimension, respectively, and α and β are the corresponding integration variables.

[0082] In this embodiment, the calculated deconvolution beamforming result (ISDBF) after suppressing near-field interference is as follows: Figure 6 As shown.

[0083] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A near-field strong interference suppression method based on spherical wave deconvolution beamforming localization, characterized in that, The method includes: Fine-scale meshing is performed on the near-field scanning area of ​​the towed linear array; Spherical wave focusing beamforming and two-dimensional deconvolution beamforming with respect to distance and angle in the case of unknown sound source location are performed in the near-field scanning region. Calculate the distribution entropy characteristics of the deconvolution beamforming result in the angular dimension, and use a preset threshold to locate the interference of associated targets; Construct the array manifold matrix A for each near-field target l And generate the covariance matrix A of the array manifold matrix; Calculate the orthogonal complement projection matrix of the covariance matrix A, and apply the projection matrix to the element domain covariance matrix C to suppress near-field interference. The far-field plane wave model is used to perform deconvolution beamforming calculations on the element domain covariance matrix C for suppressing near-field interference, resulting in passive detection results for suppressing strong near-field interference.

2. The near-field strong interference suppression method based on spherical wave deconvolution beamforming localization as described in claim 1, characterized in that, The detailed mesh division of the near-field scanning region of the linear array includes: The far field and near field of the towed linear array are coarsely divided to obtain the near field scanning region of the towed linear array. Based on the required near-field estimation resolution, a fine grid is made for the near-field scanning region of the linear array.

3. The near-field strong interference suppression method based on spherical wave deconvolution beamforming localization as described in claim 2, characterized in that, The coarse division of the far and near fields of the towed array adopts the towed array focusing distance division method, including: Assuming the target is a far-field plane wave, the scanning region R = [r1, r2, ..., r n Based on the signal magnitude SL(r) obtained at different focusing distances after array gain obtained by spherical wave beamforming, the signal loss Loss(r) caused by focusing is calculated as Loss(r) = SL - SL(r), where SL is the signal magnitude obtained by plane wave beamforming. Compare the signal loss Loss(r), r = 1, 2, ..., n, with the allowable signal loss σ. When Loss(r) ≥ σ and min(Loss(r) - σ), the corresponding r is the focusing distance when the signal loss Loss is closest to σ. Let R be the critical distance between the near and far fields of the linear array. m =r.

4. The near-field strong interference suppression method based on spherical wave deconvolution beamforming localization as described in claim 1, characterized in that, The spherical wave focusing beamforming algorithm is expressed as follows: ， Where P′(r, θ) is the spatial spectrum at a focusing distance of r and a focusing azimuth of θ, c is the wave velocity, d is the element spacing of the towed array, M is the number of elements in the towed array, r is the focusing distance, θ is the focusing azimuth, x0 is the distance difference between the sound signal to each sound channel and to the reference sound channel, j is an imaginary number, f is the signal frequency, a(r, θ) is the driving vector, and τ is the arrival time delay of the signal at a focusing distance of r and a focusing azimuth of θ in each sound channel. T For the transpose, X = [x1(f), x2(f), ..., x M [f] represents the Fourier transform of the signal in the element domain at frequency. f Result.

5. The near-field strong interference suppression method based on spherical wave deconvolution beamforming localization as described in claim 4, characterized in that, The two-dimensional deconvolution beamforming with respect to distance and angle in the near-field scanning region with unknown sound source location includes: Calculate the scattering function of the sound source at 90° as a function of the focusing distance r: ， In the formula, rs is the distance from the sound source to each element of the dragline array, rr is the distance from the scanning point to each element of the dragline array, and a is the reciprocal of the distance from the sound source to each element of the dragline array. The deconvolution beamforming result P1 is calculated based on the spatial spectrum P′(r, θ) and the scattering function PSF(r, 90°); The spatial spectral distribution at the center distance of P1 is taken as the deconvolution spatial spectrum P at a distance of r. deconv Scanning the distance dimension yields the two-dimensional deconvolution beamforming result P. deconv (r, θ).

6. The near-field strong interference suppression method based on spherical wave deconvolution beamforming localization as described in claim 1, characterized in that, The calculation of the distribution entropy characteristics of the deconvolution beamforming result in the angular dimension, and the use of a preset threshold for interference localization of associated targets, includes the following steps: The result of two-dimensional deconvolution beamforming P deconv Slice (r, θ) into angular dimensions θ = {θ1, θ2, ..., θ} k ...}, where k is the slice index, calculate P deconv The distribution entropy of the (r, θ) angle-dimensional slice, Entropy(θ) k ); Determine the distribution entropy value Entropy(θ) k If the value is not greater than the threshold TEntropy of the set distribution entropy, then extract P. deconv θ in (r, θ) k The distance r corresponding to the maximum value of the slice k Identify the scan grid (r) k θ k There is near-field interference at point ).

7. The near-field strong interference suppression method based on spherical wave deconvolution beamforming localization as described in claim 1, characterized in that, The formula for calculating the orthogonal complement spatial projection matrix of the covariance matrix A of the array manifold matrix is: P=I-A T (AA T ) -1 A, Where T is the transpose and I is the identity matrix.

8. The near-field strong interference suppression method based on spherical wave deconvolution beamforming localization as described in claim 1, characterized in that, The formula for calculating the covariance matrix C of the array element domain under the projection matrix P is: C = PC.

9. The near-field strong interference suppression method based on spherical wave deconvolution beamforming localization as described in claim 1, characterized in that, The deconvolution beamforming calculation using the far-field plane wave model to suppress near-field interference on the element-domain covariance matrix C includes: The element-domain covariance matrix C, which suppresses near-field interference, is used to calculate the far-field planar beamforming result ISCBF. Using the far-field deconvolution beamforming scattering function PSF(r, 90°) and the far-field planar beamforming result ISCBF as inputs to the RL deconvolution algorithm, the deconvolution beamforming result ISDBF after suppressing near-field interference is obtained.

10. The near-field strong interference suppression method based on spherical wave deconvolution beamforming localization as described in claim 9, characterized in that, The formula for the RL deconvolution algorithm is expressed as follows: ， Where i is the number of iterations, r is the scanning distance range, and θ is the scanning angle. P(r, θ) represents the range-angle two-dimensional spatial spectrum estimation result obtained from the i-th deconvolution iteration, and P(r, θ) represents the range-angle two-dimensional spatial spectrum estimation obtained from conventional beamforming.

Images