A sparse direction finding method for underwater unmanned platform sonar array against far and near field interference

Abstract

The application discloses a sparse direction-finding method for resisting far and near field interference suitable for an underwater unmanned platform sonar array, belongs to the underwater acoustic signal processing field, and particularly relates to the sparse direction-finding method for resisting far and near field interference. The application aims to solve the problems of low resolution probability and direction-finding precision of a far field weak target signal under the existing far and near field strong interference conditions. The specific process of the sparse direction-finding method for resisting far and near field interference suitable for the underwater unmanned platform sonar array is as follows: step one: determining a near field region set according to the platform size and far and near field steering vector correlation; step two: performing subspace projection processing on data received by a receiving array to obtain a projection of a covariance matrix in a signal interference subspace; and step three: processing the projection of the covariance matrix in the signal interference subspace based on a sparse reconstruction type method to obtain a direction-finding result.

Description

Technical Field

[0001] This invention belongs to the field of underwater acoustic signal processing, and in particular relates to a sparse direction finding method that resists near and far field interference. Background Technology

[0002] Direction finding for unmanned underwater vehicle (UUV) arrays is a significant research topic in passive sonar signal processing. Unlike arrays mounted on ships, the small size of UUVs makes it difficult to create a vibration-isolated and noise-reducing working environment for the array, resulting in severe near-field interference caused by UUV maneuvers affecting array signal processing. Under the influence of strong near-field interference, direction finding algorithms for far-field target signals suffer significant performance degradation, specifically a decrease in angular resolution and direction finding accuracy, along with the appearance of multiple spurious peaks in the far-field spatial spectrum. Furthermore, when one weak far-field target is selected as the target of interest, other strong far-field targets will act as strong far-field interference, adversely affecting the detection of the weak target.

[0003] Current direction-finding methods can be mainly classified into maximum likelihood (ML), beamforming, subspace, and sparse reconstruction methods. ML algorithms, under model-fitting conditions, often yield results closest to the Cramerau bound (CRB), but their extremely high computational cost is unacceptable for the real-time processing requirements of UUV platforms. Among beamforming methods, conventional beamforming (CBF) is the most common and robust to various mismatch conditions, widely used in sonar signal processing. Its drawbacks include its inability to overcome the Rayleigh limit, its inability to resolve nearby targets within the same beam, and the ability of strong interference components leaked by sidelobes to mask weak targets. Minimum variance distortionless response (MVDR) beamforming algorithms exhibit good suppression of strong interference in both near and far fields and have higher resolution than CBF. However, MVDR is highly sensitive to model mismatch, and its performance degrades significantly when model mismatch exists. Subspace algorithms leverage the orthogonality of the signal and noise subspaces to achieve high-resolution direction finding of target signals. However, in UUV operating environments, received data includes weak far-field target signals, near-field and far-field interference signals, and marine environmental noise. In this case, the basis of the signal and interference subspaces may overlap or superimpose, making accurate estimation of the signal subspace impossible through eigenvalue decomposition. Compared to subspace algorithms, sparse reconstruction methods, beamforming algorithms, and machine learning (ML) algorithms offer higher angular resolution, but their sparse priors often rely on overcomplete dictionaries under ideal conditions. Strong near-field interference caused by UUV platform maneuvers leads to dictionary mismatch, meaning the basis in the far-field dictionary matrix cannot accurately represent the strong near-field interference signals in the array's received signal, resulting in increased fitting errors and degradation of far-field direction finding capabilities. This invention addresses these issues by proposing a sparse direction finding method suitable for small platform sonar arrays that resists near-field and far-field interference. Summary of the Invention

[0004] The purpose of this invention is to solve the problems of low resolution probability and low direction finding accuracy of weak target signals in the far field under strong interference conditions in the near and far fields, and to propose a sparse direction finding method for sonar arrays of underwater unmanned platforms that is resistant to near and far field interference.

[0005] The specific process of a sparse direction finding method for resisting near-field and far-field interference for sonar arrays of underwater unmanned platforms is as follows:

[0006] Step 1: Determine the near-field region set based on the platform size and the correlation between near and far-field guidance vectors;

[0007] Step 2: Perform subspace projection processing on the data received by the receiving array to obtain the projection R of the covariance matrix in the signal interference subspace. s ;

[0008] Step 3: Process the projection R of the covariance matrix in the signal interference subspace based on sparse reconstruction methods. s The direction finding results were obtained.

[0009] The beneficial effects of this invention are as follows:

[0010] The purpose of this invention is to construct a sparse representation model for joint far and near fields, and to reconstruct the covariance matrix of the far-field and near-field parts of the received signal respectively, thereby achieving high-precision and high-resolution direction finding of weak target signals in the far field under strong interference conditions in both the far and near fields.

[0011] This invention determines the range of the near-field dictionary set based on the platform size and the correlation between near and far-field steering vectors; performs subspace projection processing on the received data to obtain the projection of the covariance matrix in the signal interference subspace; processes the projection of the covariance matrix in the signal interference subspace based on a sparse reconstruction method, and outputs the direction finding results.

[0012] The core technology of this invention lies in using the sparse prior of near and far field sound sources in the spatial domain to separate the near and far field covariance matrices. Furthermore, to address the problem of large differences between the power of interference and target signals, a subspace projection method is used to normalize each source.

[0013] This invention is based on sparse reconstruction technology. It utilizes the sparse priors of near and far field sound sources in the spatial domain to separate far field target signals and near field interference signals, thereby reducing the impact of near field interference caused by platform self-noise on far field direction finding.

[0014] This invention preprocesses the input covariance matrix based on subspace projection technology and normalizes the power of each source, thus suppressing strong interference in both near and far fields.

[0015] This invention constructs a direction finding method suitable for underwater small platform arrays, achieving high-precision and high-resolution direction finding for weak targets in the far field under finite array element conditions;

[0016] This invention considers the impact of far-field interference and near-field interference caused by the self-noise of underwater unmanned platform maneuvers on the direction finding of weak far-field targets. It combines sparse reconstruction and subspace method to suppress far- and near-field interference, thereby improving the direction finding capability of weak far-field target signals in scenarios with strong far- and near-field interference. Attached Figure Description

[0017] Figure 1 This is a flowchart of the present invention;

[0018] Figure 2A schematic diagram of the near and far field division applicable to UUV platforms, (a) a correlation diagram of near and far field steering vectors, (b) a schematic diagram of the array coordinate system;

[0019] Figure 3 A situational map of China's tests in the South China Sea;

[0020] Figure 4 Here are the true bearing and heading angle of the target ship: (a) True bearing of the target ship, (b) Heading angle of the target ship.

[0021] Figure 5 The diagram shows the time-location history before and after near-field interference suppression: (a) before near-field interference suppression, (b) after near-field interference suppression.

[0022] Figure 6 This is a comparison chart of the direction finding results of the improved method and the original method. Detailed Implementation

[0023] Specific implementation method one: Combining Figure 1 This embodiment describes a sparse direction finding method for resisting near-field and far-field interference applicable to sonar arrays of underwater unmanned platforms. The specific process is as follows:

[0024] Step 1: Determine the near-field region set based on the platform size and the correlation between near and far-field guidance vectors;

[0025] Step 2: Perform subspace projection processing on the data received by the receiving array to obtain the projection R of the covariance matrix in the signal interference subspace. s ;

[0026] Step 3: Process the projection R of the covariance matrix in the signal interference subspace based on sparse reconstruction methods. s The direction finding results were obtained.

[0027] Specific Implementation Method Two: This implementation method differs from Specific Implementation Method One in that: in step one, the near-field region set is determined based on the platform size and the correlation between near and far-field guidance vectors;

[0028] The specific process is as follows:

[0029] Step 11: Determine the correlation between near and far field steering vectors;

[0030] Steps 1 and 2: Define the set Ω1 based on the correlation of near and far field steering vectors;

[0031] Step 13: Define set Ω2 based on platform dimensions;

[0032] Step 14: Determine the near-field region set based on set Ω1 and set Ω2.

[0033] The other steps and parameters are the same as in Specific Implementation Method 1.

[0034] Specific Implementation Method Three: This implementation method differs from Specific Implementation Method One or Two in that: in step one, the correlation between the near and far field steering vectors is determined; the specific process is as follows:

[0035] The receiving array is a uniform linear array (ULA), consisting of M uniformly arranged array elements with an element spacing of d = λ / 2, where λ is the wavelength corresponding to the center frequency of the narrowband signal; M takes a value of 4 ≤ M ≤ 50; a Cartesian coordinate system is established with the array center as the origin, the platform bow direction as the positive horizontal axis, and the array center normal direction as the positive vertical axis. The coordinates of each array element are (d... m ,0), where d m Let d represent the coordinates of each array element on the horizontal axis in the Cartesian coordinate system. m =[m-1-(M-1) / 2]d,m=1...M;

[0036] The distribution of near-field interference sources studied in this invention is related to the size of the UUV platform. The interference sources are often not distributed in the Fresnel region. Therefore, the approximate conclusions of the Fresnel region and the near-field division criteria are not applicable.

[0037] The correlation expression for the near and far field steering vectors is shown below:

[0038]

[0039] Wherein, κ(α,β) represents the correlation between the near and far field steering vectors. The larger the value of κ(α,β), the stronger the correlation between the near field steering vector and the far field steering vector when the near field source coordinates are (α,β). This indicates that θ is the variable, and the maximum value among all values ​​is taken; |·| represents the modulo operation; a(θ) is the far-field steering vector, θ is the direction of arrival of the far-field signal, and the value range of θ is any angle in the far-field spatial domain; a(α,β) is the near-field steering vector, (α,β) is the near-field source coordinate (in the Cartesian coordinate system mentioned above), and the value range of (α,β) is: {(α,β)|α 2 +β 2 ≤4D 4 / λ 2}, where D is the array aperture, λ is the wavelength corresponding to the center frequency of the narrowband signal; |||2 is the L2 norm; the superscript H indicates conjugate;

[0040] The mathematical expressions for a(θ) and a(α,β) are:

[0041]

[0042] Where d1 is the x-coordinate of the first array element in the Cartesian coordinate system; dM Let (α,β) be the x-coordinate of the Mth element in Cartesian coordinates; the polar coordinate position corresponding to (α,β) is (r,δ), where r is the polar radius and δ is the polar angle. sin -1 (·) denotes the arcsine operation; r1 is the distance from coordinate (α,β) to the first element; r M Let be the distance from coordinates (α, β) to the Mth element; j is the imaginary unit, j 2 =-1; (·) T This indicates the transpose operation.

[0043] When κ(α,β)→1 in equation (1), the sound source is considered to be more similar to a far-field sound source. Consider the following scenario: a 10-element ULA with an element spacing of 0.5m, the trend of κ(α,β) is as follows. Figure 2 As shown in Figure (a), contour lines were drawn for positions with κ values ​​of 0.6, 0.8, and 0.9, respectively. Sound sources located in the region where κ(α,β) < 0.9 were considered as near-field sound sources, while sound sources in the region where κ(α,β) > 0.9 were considered to have characteristics closer to far-field sound sources and were therefore not considered.

[0044] Other steps and parameters are the same as in specific implementation method one or two.

[0045] Specific Implementation Method Four: This implementation method differs from one of the specific implementation methods one to three in that: in steps one and two, the correlation definition set Ω1 is based on the near and far field guidance vectors;

[0046] The expression is as follows:

[0047]

[0048] in, Define the symbol.

[0049] The other steps and parameters are the same as those in one of the specific implementation methods one to three.

[0050] Specific Implementation Method Five: This implementation method differs from Specific Implementation Methods One to Four in that: in step one to three, the set Ω2 is defined based on the platform size; the specific process is as follows:

[0051] Meanwhile, the near-field interference signal received by the UUV array has a near-field interference distribution range that is related to the platform size. For example... Figure 2 As shown in (b), in the established Cartesian coordinate system, the length of the UUV is X. u Width is Y u Assume the horizontal axis of the platform in the Cartesian coordinate system has a range of... The vertical axis range is

[0052] in This represents the left boundary of the UUV in the Cartesian coordinate system. This represents the right boundary of the UUV in the Cartesian coordinate system. This represents the upper boundary of the UUV in the Cartesian coordinate system. This represents the lower boundary of the UUV in the Cartesian coordinate system.

[0053] Define set Ω2:

[0054]

[0055] The other steps and parameters are the same as those in specific implementation methods one through four.

[0056] Specific Implementation Method Six: This implementation method differs from Specific Implementation Methods One through Five in that: in step one four, the near-field region set is determined based on set Ω1 and set Ω2; expressed as:

[0057] Ω=Ω1∩Ω2 (5)

[0058] Where ∩ represents the intersection.

[0059] The other steps and parameters are the same as those in specific implementation methods one through five.

[0060] Specific Implementation Method Seven: This implementation method differs from Specific Implementation Methods One through Six in that: in step two, the data received by the receiving array undergoes subspace projection processing to obtain the projection R of the covariance matrix in the signal interference subspace. s The specific process is as follows:

[0061] Step 21: Estimate the covariance matrix of the received data:

[0062]

[0063] in, Y represents the sample covariance matrix estimated by the receiving snapshot, and T represents the original data received by the receiving array.

[0064] Step 22: Estimate the sample covariance matrix Perform eigenvalue decomposition; represented as:

[0065]

[0066] Among them, U s The subspace matrix representing the far-field target signal plus near-field interference is simply called the signal-interference subspace matrix; Σ s Represents matrix U s eigenvalue matrix; U e Let Σ represent the noise subspace matrix. eRepresents the noise subspace matrix U e eigenvalue matrix;

[0067] Steps two and three: Based on the signal interference subspace matrix U s Define the projection R of the covariance matrix into the signal interference subspace. s .

[0068] The other steps and parameters are the same as those in one of the specific implementation methods one to six.

[0069] Specific Implementation Method Eight: This implementation method differs from Specific Implementation Methods One through Seven in that: in steps two and three, the signal interference subspace matrix U is used as the basis for... s Define the projection R of the covariance matrix into the signal interference subspace. s ;

[0070] The expression is:

[0071]

[0072] Among them, R s U represents the projection of the covariance matrix onto the signal interference subspace. s This represents the signal interference subspace matrix.

[0073] The other steps and parameters are the same as those in any of the specific implementation methods one to seven.

[0074] Specific Implementation Method Nine: This implementation method differs from Specific Implementation Methods One through Eight in that: in step three, the projection R of the covariance matrix in the signal interference subspace is processed based on the sparse reconstruction method. s The direction finding results are obtained;

[0075] The specific process is as follows:

[0076] Step 31: Divide the near and far field meshes and generate a hybrid overcomplete dictionary set Ψ;

[0077] The specific process is as follows:

[0078] For the far field, the entire far field space angle from 0° to 180° is divided into L1 grid points, represented as a set.

[0079] θ1 represents the angle of the first grid point in the far field, and θ2 represents the angle of the second grid point in the far field. This represents the angle of the L1-th grid point in the far field;

[0080] For the near field, the near field region set Ω is uniformly divided into L2 grid points, denoted as set Ω.

[0081] (α1,β1) represents the position of the first grid point in the near field, and (α2,β2) represents the position of the second grid point in the near field. Indicates the position of the L2-th grid point in the near field;

[0082] Define grid In the following text, the near-field mesh and the far-field mesh will be collectively referred to as mesh 1 to L;

[0083] Substituting sets Θ and χ into equation (2) respectively, we construct a mixed overcomplete dictionary set Ψ, denoted as:

[0084]

[0085] Where a(θ1) is the steering vector corresponding to the first far-field grid angle, The guiding vector corresponding to the L1th far-field grid angle;

[0086] a(α1,β1) is the guiding vector corresponding to the first near-field grid position. This is the guiding vector corresponding to the L2th near-field grid position;

[0087] I M This represents an identity matrix of dimension M×M, where M represents the total number of matrix elements;

[0088] This represents the first column of the Ψ matrix. This represents the second column in the Ψ matrix. This represents the L+Mth column in the Ψ matrix;

[0089] Step 32: Based on the projection R of the covariance matrix into the signal interference subspace s The far-field spatial spectrum estimate is obtained by mixing the overcomplete dictionary set Ψ.

[0090] This invention requires an iterative approach to DOA estimation. During the iteration process, this invention is indicated by superscript (·). (i) Indicates the number of iterations.

[0091] The other steps and parameters are the same as those in one of the specific implementation methods one to eight.

[0092] Specific Implementation Method Ten: This implementation method differs from Specific Implementation Methods One to Nine in that: Step Three Two: Based on the projection R of the covariance matrix in the signal interference subspace s The far-field spatial spectrum estimate is obtained by mixing the overcomplete dictionary set Ψ;

[0093] This invention requires an iterative approach to DOA estimation. During the iteration process, this invention is indicated by superscript (·). (i) Indicates the number of iterations.

[0094] The specific process is as follows:

[0095] Input data:

[0096] The input data for this step is as follows: R obtained from equation (8) s The hybrid overcomplete dictionary set Ψ obtained by equation (9);

[0097] Step 321: Set the maximum number of iterations to iter max The tolerance threshold is th;

[0098] Step 322: Initialize the iteration count i = 0, and initialize the tolerance initial value ε = 1;

[0099] initialization and σ (0) :

[0100]

[0101] in, This represents the initial value of the spatial spectrum when the grid number is l, with the subscript values ​​ranging from l to 1, 2, ..., L. (0) Indicates the initial value of the noise power; This represents the l-th column in the Ψ matrix;

[0102] definition for The M smallest values ​​in the middle, This represents a set, with indices ranging from 1 to M. This represents a set, with indices ranging from 1 to L;

[0103] Represents the set of columns that have been mixed with a complete dictionary set;

[0104] Step 323: Determine if i < iter max Or ε≤th;

[0105] If not, then the far-field spatial spectrum estimate is obtained.

[0106] If so, update the fitted covariance matrix R using the following formula. (i) :

[0107] R (i) =ΨP (i) Ψ H (11)

[0108] Among them, P (i) Let P be a diagonal matrix. (i) Represented as:

[0109]

[0110] in This is the spatial spectrum value corresponding to the first grid in the i-th iteration (obtained from the previous iteration or initialization); This represents the spatial spectrum value corresponding to the second grid in the i-th iteration; σ is the spatial spectrum value corresponding to the Lth grid in the i-th iteration; (i) This represents the noise power value in the i-th iteration (obtained from the previous iteration or initialization);

[0111] Based on the fitted covariance matrix R (i) The projection R of the sum covariance matrix in the signal interference subspace s Update the corresponding iteration for the (i+1)th iteration and σ (i+1) :

[0112]

[0113] Where, p l (i+1) This represents the spatial spectrum value corresponding to the l-th grid in the (i+1)-th iteration; This represents the spatial spectrum value corresponding to the l-th grid in the i-th iteration;

[0114] R (i) Represents the covariance matrix obtained by reconstruction in the i-th iteration;

[0115] w represents an intermediate variable, w = [w1, ..., w] l ,…,w L ], where w1 is w l The value of l=1 in the middle, w L For w l The value of l = L, w l Indicates intermediate variables.

[0116] ρ (i) This represents an intermediate variable in the i-th iteration;

[0117] l represents the spatial spectrum index, l = 1, 2, ..., L, and L represents the total number of grid points in the near and far fields;

[0118] σ (i+1) σ represents the noise power value in the (i+1)th iteration. (i) This represents the noise power value in the i-th iteration;

[0119] (·) -1 || represents the matrix inversion operation; ||2 represents the L2 norm;

[0120] Update the tolerance value ε using the following formula:

[0121]

[0122] Where, p s (i+1) The mixed spatial spectrum value in the (i+1)th iteration is represented as: in This represents the spatial spectrum value corresponding to the first grid in the (i+1)th iteration. p represents the spatial spectrum value corresponding to the Lth grid in the (i+1)th iteration; s (i) This represents the mixed spatial spectrum value in the i-th iteration. in This represents the spatial spectrum value corresponding to the first grid in the i-th iteration. Let represent the spatial spectrum value corresponding to the Lth grid in the i-th iteration;

[0123] Step 3-2-4: Let i = i + 1, and repeat Step 3-2-3 until i < iter max Or ε≤th; thus, the far-field spatial spectrum estimate is obtained.

[0124] The other steps and parameters are the same as those in any of the specific implementation methods one to nine.

[0125] Figure 3 A situational map of China's tests in the South China Sea;

[0126] Figure 4 Given the true bearing and heading angle of the target ship, Figure 4 (a) represents the true bearing of the target ship. Figure 4 (b) represents the target ship's heading angle;

[0127] Figure 5 This is a time-location history diagram before and after near-field interference suppression. Figure 5 (a) shows the area before near-field interference suppression. Figure 5 (b) shows the results after suppressing near and far field interference;

[0128] Figure 6 This is a comparison chart of the direction finding results of the improved method and the original method.

[0129] The pseudocode for the process of this invention is shown in the table below:

[0130]

[0131]

[0132] This invention may have other embodiments. Without departing from the spirit and essence of this invention, those skilled in the art can make various corresponding changes and modifications according to this invention, but these corresponding changes and modifications should all fall within the protection scope of the appended claims.

Claims

1. A sparse direction finding method for resisting near-field and far-field interference in sonar arrays of underwater unmanned platforms, characterized in that: The specific process of the method is as follows: Step 1: Determine the near-field region set based on the platform size and the correlation between near and far-field guidance vectors; Step 2: Perform subspace projection processing on the data received by the receiving array to obtain the projection R of the covariance matrix in the signal interference subspace. s ; Step 3: Process the projection R of the covariance matrix in the signal interference subspace based on sparse reconstruction methods. s The direction finding results are obtained; the specific process is as follows: Step 31: Generate a mixed overcomplete dictionary set Ψ; The specific process is as follows: For the far field, the entire far field space angle from 0° to 180° is divided into L1 grid points, represented as a set. θ1 represents the angle of the first grid point in the far field, and θ2 represents the angle of the second grid point in the far field. This represents the angle of the L1-th grid point in the far field; For the near field, the near field region set Ω is uniformly divided into L2 grid points, denoted as set Ω. (α1,β1) represents the position of the first grid point in the near field, and (α2,β2) represents the position of the second grid point in the near field. Indicates the position of the L2-th grid point in the near field; Define grid Substitute sets Θ and χ into... In the middle, a mixed overcomplete dictionary set Ψ is formed, denoted as Where a(θ1) is the steering vector corresponding to the first far-field grid angle, The guiding vector corresponding to the L1th far-field grid angle; a(α1,β1) is the guiding vector corresponding to the first near-field grid position. This is the guiding vector corresponding to the L2th near-field grid position; I M This represents an identity matrix of dimension M×M, where M represents the total number of matrix elements; This represents the first column of the Ψ matrix. This represents the second column in the Ψ matrix. This represents the L+Mth column in the Ψ matrix; a(θ) is the far-field steering vector, where θ is the direction of arrival of the far-field signal wave; a(α,β) is the near-field steering vector, (α,β) is the near-field source coordinates, and λ is the wavelength corresponding to the center frequency of the narrowband signal; d1 is the x-coordinate of the first array element in Cartesian coordinates; d M Let r be the x-coordinate of the Mth element in Cartesian coordinates; r is the polar radius; r1 is the distance from coordinate (α,β) to the 1st element; r M Let be the distance from coordinates (α, β) to the Mth element; j is the imaginary unit, j 2 =-1; (·) T Indicates the transpose operation; Step 32: Based on the projection R of the covariance matrix into the signal interference subspace s The far-field spatial spectrum estimate is obtained by mixing the overcomplete dictionary set Ψ.

2. The sparse direction finding method for resisting near-field and far-field interference of sonar arrays for underwater unmanned platforms according to claim 1, characterized in that: In step one, the near-field region set is determined based on the platform size and the correlation between near and far-field guidance vectors; The specific process is as follows: Step 11: Determine the correlation between near and far field steering vectors; Steps 1 and 2: Define the set Ω1 based on the correlation of near and far field steering vectors; Step 13: Define set Ω2 based on platform dimensions; Step 14: Determine the near-field region set based on set Ω1 and set Ω2.

3. The sparse direction finding method for resisting near-field and far-field interference for sonar arrays of underwater unmanned platforms according to claim 2, characterized in that: The correlation between the near and far field steering vectors is determined in step one by one; the specific process is as follows: The receiving array is a uniform linear array, consisting of M array elements evenly arranged with an element spacing of d = λ / 2, where λ is the wavelength corresponding to the center frequency of the narrowband signal. A Cartesian coordinate system is established with the array center as the origin, the platform bow direction as the positive horizontal axis, and the array center normal direction as the positive vertical axis. The coordinates of each element are (d... m ,0), where d m Let d represent the coordinates of each array element on the horizontal axis in the Cartesian coordinate system. m =[m-1-(M-1) / 2]d,m=1...M; The correlation expression for the near and far field steering vectors is shown below: Where κ(α,β) represents the correlation between the near and far field steering vectors; This indicates that θ is the variable, and the maximum value among all values ​​is taken; |·| represents the modulo operation; a(θ) is the far-field steering vector, and θ is the direction of arrival of the far-field signal; a(α,β) is the near-field steering vector, (α,β) is the near-field source coordinate, and the range of (α,β) is: {(α,β)|α 2 +β 2 ≤4D 4 / λ 2 }, where D is the array aperture, λ is the wavelength corresponding to the center frequency of the narrowband signal; |||2 is the L2 norm; the superscript H indicates conjugate; The mathematical expressions for a(θ) and a(α,β) are: Where d1 is the x-coordinate of the first array element in the Cartesian coordinate system; d M Let (α,β) be the x-coordinate of the Mth element in Cartesian coordinates; the polar coordinate position corresponding to (α,β) is (r,δ), where r is the polar radius and δ is the polar angle. sin -1 (·) denotes the arcsine operation; r1 is the distance from coordinate (α,β) to the first element; r M Let be the distance from coordinates (α, β) to the Mth element; j is the imaginary unit, j 2 =-1; (·) T This indicates the transpose operation.

4. The sparse direction finding method for resisting near-field and far-field interference for sonar arrays of underwater unmanned platforms according to claim 3, characterized in that: The correlation definition set Ω1 based on the near and far field guidance vectors in steps one and two; The expression is as follows: in, Define the symbol.

5. A sparse direction finding method for resisting near-field and far-field interference for sonar arrays of underwater unmanned platforms according to claim 4, characterized in that: In steps one and three, the set Ω2 is defined based on the platform size; the specific process is as follows: In the established Cartesian coordinate system, the length of the UUV is X. u Width is Y u Assume the horizontal axis of the platform in the Cartesian coordinate system has a range of... The vertical axis range is in This represents the left boundary of the UUV in the Cartesian coordinate system. This represents the right boundary of the UUV in the Cartesian coordinate system. This represents the upper boundary of the UUV in the Cartesian coordinate system. This represents the lower boundary of the UUV in the Cartesian coordinate system. Define set Ω2:

6. A sparse direction finding method for resisting near-field and far-field interference for sonar arrays of underwater unmanned platforms according to claim 5, characterized in that: In step one of the four steps, the near-field region set is determined based on set Ω1 and set Ω2; represented as follows: Ω=Ω1∩Ω2 (5) Where ∩ represents the intersection.

7. A sparse direction finding method for resisting near-field and far-field interference for sonar arrays of underwater unmanned platforms according to claim 6, characterized in that: In step two, the data received by the receiving array is subjected to subspace projection processing to obtain the projection R of the covariance matrix in the signal interference subspace. s The specific process is as follows: Step 21: Estimate the covariance matrix of the received data: in, Y represents the sample covariance matrix estimated by the receiving snapshot, and T represents the original data received by the receiving array. Step 22: Estimate the sample covariance matrix Perform eigenvalue decomposition; represented as: Among them, U s Represents the signal interference subspace matrix; Σ s Represents matrix U s eigenvalue matrix; U e Let Σ represent the noise subspace matrix. e Represents the noise subspace matrix U e eigenvalue matrix; Steps two and three: Based on the signal interference subspace matrix U s Define the projection R of the covariance matrix into the signal interference subspace. s .

8. A sparse direction finding method for resisting near-field and far-field interference for sonar arrays of underwater unmanned platforms according to claim 7, characterized in that: In steps two and three, the signal interference subspace matrix U is used as a basis. s Define the projection R of the covariance matrix into the signal interference subspace. s ; The expression is: Among them, R s U represents the projection of the covariance matrix onto the signal interference subspace. s This represents the signal interference subspace matrix.

9. A sparse direction finding method for resisting near-field and far-field interference for sonar arrays of underwater unmanned platforms according to claim 8, characterized in that: In step three, the projection R of the covariance matrix into the signal interference subspace is used. s The far-field spatial spectrum estimate is obtained by mixing the overcomplete dictionary set Ψ; The specific process is as follows: Step 321: Set the maximum number of iterations to iter max The tolerance threshold is th; Step 322: Initialize the iteration count i = 0, and initialize the tolerance initial value ε = 1; initialization and σ (0) : in, This represents the initial value of the spatial spectrum when the grid number is l, with the subscript values ​​ranging from l to 1, 2, ..., L. (0) Indicates the initial value of the noise power; This represents the l-th column in the Ψ matrix; definition for The M smallest values ​​in the middle, This represents a set, with indices ranging from 1 to M. This represents a set, with indices ranging from 1 to L; Represents the set of columns that have been mixed with a complete dictionary set; Step 323: Determine if i < iter max Or ε≤th; If not, then the far-field spatial spectrum estimate is obtained. If so, update the fitted covariance matrix R using the following formula. (i) : R (i) =ΨP (i) P H (11) Among them, P (i) Let P be a diagonal matrix. (i) Represented as: in This represents the spatial spectrum value corresponding to the first grid in the i-th iteration; This represents the spatial spectrum value corresponding to the second grid in the i-th iteration; σ is the spatial spectrum value corresponding to the Lth grid in the i-th iteration; (i) Let be the value of the noise power in the i-th iteration; Based on the fitted covariance matrix R (i) The projection R of the sum covariance matrix in the signal interference subspace s Update p corresponding to the (i+1)th iteration l (i+1) and σ (i+1) : Where, p l (i+1) This represents the spatial spectrum value corresponding to the l-th grid in the (i+1)-th iteration; This represents the spatial spectrum value corresponding to the l-th grid in the i-th iteration; R (i) Represents the covariance matrix obtained by reconstruction in the i-th iteration; w represents an intermediate variable, w = [w1, ..., w] l ,…,w L ], where w1 is w l The value of l=1 in the middle, w L For w l The value of l = L, w l Indicates intermediate variables. ρ (i) This represents an intermediate variable in the i-th iteration; l represents the spatial spectrum index, l = 1, 2, ..., L, and L represents the total number of grid points in the near and far fields; σ (i+1) σ represents the noise power value in the (i+1)th iteration. (i) This represents the noise power value in the i-th iteration; (·) -1 || represents the matrix inversion operation; ||2 represents the L2 norm; Update the tolerance value ε using the following formula: Where, p s (i+1) The mixed spatial spectrum value in the (i+1)th iteration is represented as: in This represents the spatial spectrum value corresponding to the first grid in the (i+1)th iteration. p represents the spatial spectrum value corresponding to the Lth grid in the (i+1)th iteration; s (i) This represents the mixed spatial spectrum value in the i-th iteration. in This represents the spatial spectrum value corresponding to the first grid in the i-th iteration. Let represent the spatial spectrum value corresponding to the Lth grid in the i-th iteration; Step 3-2-4: Let i = i + 1, and repeat Step 3-2-3 until i < iter max Or ε≤th; thus, the far-field spatial spectrum estimate is obtained.

Images