A method and system for estimating direction of arrival (DOA) in impulse noise scenarios

Abstract

The present invention discloses a method and system for estimating a direction of arrival in an impulse noise scenario. The method repeatedly iterates and updates impulse noise parameters and a sparse sound source signal matrix according to an overcomplete array manifold matrix, a received data vector matrix, an initial sparse sound source signal matrix, an initial impulse noise matrix, an initial impulse noise distribution accuracy parameter matrix and an initial impulse noise distribution degree of freedom parameter until an iteration condition is satisfied, thereby obtaining a final sparse sound source signal matrix, and estimating a direction of arrival according to the final sparse sound source signal matrix, so as to better adapt to a complex impulse noise environment, thereby improving the estimation accuracy of the direction of arrival, and solving the problem that, in a complex impulse noise statistical characteristic environment, a fitting error matrix with a p value selected by experience cannot fit the impulse noise well, resulting in low impulse noise fitting accuracy, thereby resulting in low estimation accuracy of the direction of arrival.

Description

Technical Field

[0001] The present invention relates to the field of direction of arrival angle estimation, and in particular to a method and system for estimating direction of arrival angle in an impulse noise scenario. Background Art

[0002] Direction of arrival (DOA) estimation is one of the important directions of array signal processing and is widely used in communications, radar, sonar and other fields. Many classic DOA estimation algorithms have been proposed in academia, most of which are only applicable to Gaussian noise scenarios. However, pulse noise is quite common in actual environments. For example, noise in shallow sea environments is particularly suitable for modeling with pulse distributions such as symmetric alpha stable distributions. In DOA estimation, pulse noise will interfere with the signal source and make general detection methods ineffective. Therefore, it is particularly important to detect the target direction in the presence of pulse noise.

[0003] In the current DOA estimation method for impulse noise scenarios, the existing technology uses a sparse class-based fitting error matrix algorithm to achieve arrival direction angle estimation in impulse noise scenarios. The sparse class-based fitting error matrix algorithm utilizes the sparsity characteristics of spatial signals while avoiding a large amount of calculation. However, the current method based on the fitting error matrix mainly selects the p value based on experience to fit the impulse noise. Since the statistical characteristics of impulse noise are quite complex, the p value selected by experience cannot adapt to the complex noise environment, resulting in the problem of low estimation accuracy of the arrival direction angle. Summary of the invention

[0004] The present invention provides a method and system for estimating the direction of arrival angle in an impulse noise scenario, which can solve the problem that in a complex impulse noise statistical characteristic environment, a fitting error matrix with a p value selected by experience cannot fit the impulse noise well, resulting in low impulse noise fitting accuracy, and further resulting in low estimation accuracy of the direction of arrival angle.

[0005] In order to solve the above technical problem, an embodiment of the present invention provides a method for estimating a direction of arrival angle in an impulse noise scenario, comprising:

[0006] Obtaining an overcomplete array manifold matrix, a received data vector matrix, an initial sparse sound source signal matrix, an initial impulse noise distribution precision parameter matrix, and an initial impulse noise distribution degree of freedom parameter corresponding to the array in the detection area;

[0007] Repeat the sparse sound source signal matrix iteration operation until the upper limit threshold of the number of iterations is reached or the change value of the optimized sparse sound source signal matrix is ​​less than the preset change threshold, to obtain the final sparse sound source signal matrix;

[0008] Performing spatial spectrum conversion and peak search according to the final sparse sound source signal matrix to obtain a direction of arrival angle;

[0009] The sparse sound source signal matrix iterative operation includes:

[0010] Calculate the current impulse noise matrix according to the current sparse sound source signal matrix, the complete array manifold matrix and the received data vector matrix; wherein the initial sparse sound source signal matrix is ​​the initial sparse sound source signal matrix;

[0011] According to the current pulse noise matrix and the current pulse noise distribution precision parameter matrix, the current pulse noise distribution degree of freedom parameter is iterated to obtain the iterated pulse noise distribution degree of freedom parameter; wherein the initial pulse noise distribution precision parameter matrix is ​​the initial pulse noise distribution precision parameter matrix; the initial pulse noise distribution degree of freedom parameter is the initial pulse noise distribution degree of freedom parameter;

[0012] Iterate the current pulse noise distribution precision parameter matrix according to the current pulse noise matrix and the iterated pulse noise distribution degree of freedom parameter to obtain the iterated pulse noise distribution precision parameter matrix;

[0013] According to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix, the current sparse sound source signal matrix is ​​iterated to obtain the iterated sparse sound source signal matrix.

[0014] Further, the calculating the current impulse noise matrix according to the current sparse sound source signal matrix, the overcomplete array manifold matrix and the received data vector matrix includes:

[0015] The current impulse noise matrix is ​​calculated by the following expression:

[0016]

[0017] in, represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; Represents the first matrix of received data vectors The received data vector of snapshots; represents an overcomplete array manifold matrix; represents the sparse sound source signal matrix of the q-1th iteration The sparse sound source signal vector of snapshots, q represents the number of iterations.

[0018] Further, the iterating the current pulse noise distribution degree of freedom parameter according to the current pulse noise matrix and the current pulse noise distribution precision parameter matrix to obtain the iterated pulse noise distribution degree of freedom parameter includes:

[0019] The current pulse noise distribution degree of freedom parameter is iterated by the following iterative formula:

[0020]

[0021]

[0022] in, Indicates The qth iteration value of the degree of freedom parameter of the impulse noise distribution of the snapshot; represents the first value of the impulse noise matrix of the q-1th iteration The conjugate transpose of the impulse noise vector of the snapshot; represents the impulse noise distribution precision parameter matrix of the q-1th iteration; M represents the number of array elements in the array; m represents the mth array element; and L represents the total number of snapshots.

[0023] Further, the iterating the current pulse noise distribution precision parameter matrix according to the current pulse noise matrix and the iterated pulse noise distribution degree of freedom parameter to obtain the iterated pulse noise distribution precision parameter matrix includes:

[0024] The current pulse noise distribution accuracy parameter matrix is ​​iterated using the following iterative formula:

[0025]

[0026]

[0027]

[0028] in, It means generating a diagonal matrix; It means multiplying the corresponding elements of two vectors; Express Take the conjugate of each element of ; represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; M represents the number of elements in the array.

[0029] Further, the iterating of the current sparse sound source signal matrix according to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix to obtain the iterated sparse sound source signal matrix includes:

[0030] The current sparse sound source signal matrix is ​​iterated using the following iterative formula:

[0031]

[0032] in, Represents the column vector Each element of the matrix Multiply each element of the corresponding row; represents the sparse sound source signal matrix of the qth iteration.

[0033] Furthermore, the column vector Each element includes:

[0034] The column vector is calculated by the following formula Each element of:

[0035]

[0036] Among them, the vector express The nth row of represents the sparse sound source signal matrix of the q-1th iteration; Representation vector The 2-norm of ; represents the regularization parameter; represents the algorithm convergence parameter; N represents the total number of grid points.

[0037] Furthermore, the matrix ,include:

[0038] The matrix is ​​calculated by the following formula :

[0039]

[0040]

[0041]

[0042]

[0043]

[0044] in, represents the sparse sound source signal matrix of the q-1th iteration; represents the regularization parameter; represents the overcomplete array manifold matrix; Y represents the received data vector matrix; represents the impulse noise distribution accuracy parameter matrix of the qth iteration; represents the first value of the impulse noise matrix of the q-1th iteration The conjugate transpose of the impulse noise vector of the snapshot; represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; Indicates multiplying the corresponding elements of two matrices; Represents moment The M-element column vector consisting of the diagonal elements of the array; M represents the number of elements in the array.

[0045] Furthermore, the change value of the optimized sparse sound source signal matrix includes:

[0046] The change value of the optimized sparse sound source signal matrix is ​​calculated by the following calculation formula:

[0047]

[0048] in, represents the sparse sound source signal matrix of the qth iteration; represents the sparse sound source signal matrix of the q-1th iteration; represents the Frobenius norm, which means finding the square root of the sum of the squares of all elements of the matrix.

[0049] Furthermore, performing spatial spectrum conversion and peak search according to the final sparse sound source signal matrix to obtain the direction of arrival angle includes:

[0050] Taking the modulus of all elements in the final sparse sound source signal matrix and adding them row by row to obtain a spatial spectrum vector;

[0051] The maximum peak value of the spatial spectrum vector is searched, and the direction of arrival angle is obtained according to the spatial angle grid point corresponding to the maximum peak value.

[0052] Based on the above method embodiment, the present invention provides a corresponding system embodiment;

[0053] An embodiment of the present invention provides a wave direction angle estimation system in an impulse noise scenario, comprising: a data acquisition module, an iteration module and a wave direction angle calculation module; wherein the iteration module comprises an impulse noise matrix calculation unit, an impulse noise distribution degree of freedom parameter iteration unit, an impulse noise distribution accuracy parameter matrix iteration unit and a sparse sound source signal matrix iteration unit;

[0054] The data acquisition module is used to obtain an overcomplete array manifold matrix, a received data vector matrix, an initial sparse sound source signal matrix, an initial impulse noise distribution precision parameter matrix, and an initial impulse noise distribution degree of freedom parameter corresponding to the array in the detection area;

[0055] The iteration module is used to repeatedly perform the sparse sound source signal matrix iteration operation until the upper limit threshold of the number of iterations is reached or the change value of the optimized sparse sound source signal matrix is ​​less than the preset change threshold, so as to obtain the final sparse sound source signal matrix;

[0056] The wave direction angle calculation module is used to perform spatial spectrum conversion and peak search according to the final sparse sound source signal matrix to obtain the wave direction angle;

[0057] The impulse noise matrix calculation unit is used to calculate the current impulse noise matrix according to the current sparse sound source signal matrix, the complete array manifold matrix and the received data vector matrix; wherein the initial sparse sound source signal matrix is ​​the initial sparse sound source signal matrix;

[0058] The pulse noise distribution degree of freedom parameter iteration unit is used to iterate the current pulse noise distribution degree of freedom parameter according to the current pulse noise matrix and the current pulse noise distribution precision parameter matrix to obtain the iterated pulse noise distribution degree of freedom parameter; wherein the initial pulse noise distribution precision parameter matrix is ​​the initial pulse noise distribution precision parameter matrix; the initial pulse noise distribution degree of freedom parameter is the initial pulse noise distribution degree of freedom parameter;

[0059] The pulse noise distribution precision parameter matrix iteration unit is used to iterate the current pulse noise distribution precision parameter matrix according to the current pulse noise matrix and the pulse noise distribution degree of freedom parameter after iteration to obtain the pulse noise distribution precision parameter matrix after iteration;

[0060] The sparse sound source signal matrix iteration unit is used to iterate the current sparse sound source signal matrix according to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix to obtain the iterated sparse sound source signal matrix.

[0061] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0062] The present invention repeatedly iterates and updates the pulse noise distribution degree of freedom parameter, the pulse noise distribution precision parameter matrix and the sparse sound source signal matrix according to an overcomplete array manifold matrix, a received data vector matrix, an initial sparse sound source signal matrix, an initial pulse noise matrix, an initial pulse noise distribution precision parameter matrix and an initial pulse noise distribution degree of freedom parameter until an iteration condition is met, thereby obtaining a final sparse sound source signal matrix, and performing wave direction of arrival estimation according to the final sparse sound source signal matrix, that is, iteratively updating the pulse noise parameters and the sparse sound source signal matrix by a method of repeated iterative updating, so as to achieve better adaptation to a complex pulse noise environment, thereby improving the estimation precision of the wave direction of arrival, and solving the problem that, in a complex pulse noise statistical characteristic environment, a fitting error matrix with a p value selected by experience cannot fit the pulse noise well, resulting in low pulse noise fitting precision, thereby resulting in low estimation precision of the wave direction of arrival. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 : A flowchart of a method for estimating a direction of arrival angle in an impulse noise scenario provided by an embodiment of the present invention;

[0064] Figure 2 : A flowchart of the steps of iterative operation of a sparse sound source signal matrix in a method for estimating a direction of arrival angle in an impulse noise scenario provided by an embodiment of the present invention;

[0065] Figure 3 : A system structure diagram of a direction of arrival angle estimation system in an impulse noise scenario provided by an embodiment of the present invention;

[0066] Figure 4 : A DOA simulation result diagram of a direction of arrival estimation method in an impulse noise scenario provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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.

[0068] Embodiment 1:

[0069] Please refer to Figure 1 , is a flowchart of a method for estimating a direction of arrival angle in an impulse noise scenario provided by an embodiment of the present invention, the method comprising at least the following steps:

[0070] Step S1: Obtain an overcomplete array manifold matrix, a received data vector matrix, an initial sparse sound source signal matrix, an initial impulse noise distribution precision parameter matrix, and an initial impulse noise distribution degree of freedom parameter corresponding to the array in the detection area;

[0071] In this embodiment, the array in the detection area includes a plurality of array elements, one array element is selected as a reference array element, and a signal in the form of a far-field flat wave emitted by a target sound source in the detection area is obtained through a plurality of array elements of the array, and the angle range of the entire detection area is evenly divided into a plurality of equally spaced grid points, each grid point includes a corresponding spatial angle, and a received data vector matrix and an overcomplete array manifold matrix are generated according to the received data vectors obtained by the plurality of array elements of the array and the array manifold vectors corresponding to the grid points, respectively; wherein the overcomplete array manifold matrix is ​​generated according to the array manifold vectors corresponding to the grid points, and the expression of the overcomplete array manifold matrix is: , The array manifold vector representing the Nth grid point is expressed as , ƒ is the signal frequency, It represents the time delay between the signal from the nth direction reaching the mth (1≤m≤M)th array element and reaching a pre-selected reference array element. c is the signal propagation speed. represents the distance from the mth array element to the reference array element, M represents the number of array elements in the array, and N represents the total number of grid points; the received data vector matrix is ​​generated according to the received data vector, and the expression of the received data vector matrix is: ;in, represents a set of M×L complex matrices, represents the received data vector of the lth snapshot, 1≤l≤L, L is the total number of snapshots;

[0072] In this embodiment, a sparse sound source signal matrix is ​​constructed, and the sparse sound source signal matrix is ​​initialized to obtain an initialized sparse sound source signal matrix; wherein the expression of the sparse sound source signal matrix is: , represents the sparse sound source signal vector corresponding to the lth snapshot;

[0073] In this embodiment, the impulse noise matrix, the impulse noise distribution precision parameter matrix and the impulse noise distribution degree of freedom parameter are generated by complex Student's T distribution, and the impulse noise distribution precision parameter matrix and the impulse noise distribution degree of freedom parameter are initialized respectively to obtain the initial impulse noise distribution precision parameter matrix and the initial impulse noise distribution degree of freedom parameter; wherein, the impulse noise matrix expression is: , represents the impulse noise vector corresponding to the lth snapshot.

[0074] Step S2: Repeat the sparse sound source signal matrix iteration operation until the upper limit threshold of the number of iterations is reached or the change value of the optimized sparse sound source signal matrix is ​​less than a preset change threshold, to obtain a final sparse sound source signal matrix;

[0075] In this embodiment, the change value of the optimized sparse sound source signal matrix includes:

[0076] The change value of the optimized sparse sound source signal matrix is ​​calculated by the following calculation formula:

[0077]

[0078] in, represents the sparse sound source signal matrix of the qth iteration; represents the sparse sound source signal matrix of the q-1th iteration; represents the Frobenius norm, which means finding the square root of the sum of the squares of all elements of the matrix.

[0079] Step S3: performing spatial spectrum conversion and peak search according to the final sparse sound source signal matrix to obtain a direction of arrival angle;

[0080] In this embodiment, all elements of the final sparse sound source signal matrix are modulo-added row by row to obtain a spatial spectrum vector, a grid point corresponding to the maximum peak is searched according to the spatial spectrum vector, and the direction of arrival angle is obtained according to the spatial angle corresponding to the grid point.

[0081] Please refer to Figure 2 , the sparse sound source signal matrix iterative operation includes:

[0082] Step S201: Calculate the current impulse noise matrix according to the current sparse sound source signal matrix, the complete array manifold matrix and the received data vector matrix; wherein the initial sparse sound source signal matrix is ​​the initial sparse sound source signal matrix;

[0083] In this embodiment, the current impulse noise matrix is ​​calculated according to the current sparse sound source signal matrix, the overcomplete array manifold matrix and the received data vector matrix, including:

[0084] The current impulse noise matrix is ​​calculated by the following expression:

[0085]

[0086] in, represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; Represents the first matrix of received data vectors The received data vector of snapshots; represents an overcomplete array manifold matrix; represents the sparse sound source signal matrix of the q-1th iteration The sparse sound source signal vector of snapshots, q represents the number of iterations.

[0087] Step S202: iterating the current pulse noise distribution degree of freedom parameters according to the current pulse noise matrix and the current pulse noise distribution precision parameter matrix to obtain the iterated pulse noise distribution degree of freedom parameters; wherein the initial pulse noise distribution precision parameter matrix is ​​the initial pulse noise distribution precision parameter matrix; the initial pulse noise distribution degree of freedom parameters are the initial pulse noise distribution degree of freedom parameters;

[0088] In this embodiment, the current pulse noise distribution degree of freedom parameter is iterated according to the current pulse noise matrix and the current pulse noise distribution precision parameter matrix to obtain the iterated pulse noise distribution degree of freedom parameter, including:

[0089] The current pulse noise distribution degree of freedom parameter is iterated by the following iterative formula:

[0090] (1)

[0091]

[0092] in, Indicates The qth iteration value of the degree of freedom parameter of the impulse noise distribution of the snapshot; represents the first value of the impulse noise matrix of the q-1th iteration The conjugate transpose of the impulse noise vector of the snapshot; represents the pulse noise distribution precision parameter matrix of the q-1th iteration; M represents the number of array elements in the array; m represents the mth array element; L represents the total number of snapshots;

[0093] In this embodiment, the steps of iterating the current impulse noise distribution degree of freedom parameter through the impulse noise distribution degree of freedom parameter iteration formula are as follows:

[0094] Step a1: Determine the value range of the freedom parameter of the impulse noise distribution as ;

[0095] Step a2: Calculate the function value and derivative value of the current pulse noise distribution degree of freedom parameter according to the current pulse noise distribution degree of freedom parameter and formula (1); wherein the initial pulse noise distribution degree of freedom parameter is the initial pulse noise distribution degree of freedom parameter;

[0096] Step a3: updating the current pulse noise distribution degree of freedom parameter according to the current pulse noise distribution degree of freedom parameter, the function value and the derivative value of the current pulse noise distribution degree of freedom parameter to obtain the updated pulse noise distribution degree of freedom parameter; when the updated pulse noise distribution degree of freedom parameter is less than or equal to a, the updated pulse noise distribution degree of freedom parameter is assigned a value; when the updated pulse noise distribution degree of freedom parameter is greater than or equal to b, the updated pulse noise distribution degree of freedom parameter is assigned b value;

[0097] Repeat step a1, step a2 and step a3 until the upper limit threshold of the pulse noise distribution freedom parameter update times is reached or the solution function value of the current pulse noise distribution freedom parameter in formula (1) is less than or equal to the preset tolerance, and the iterated pulse noise distribution freedom parameter is obtained.

[0098] Step S203: iterating the current pulse noise distribution precision parameter matrix according to the current pulse noise matrix and the iterated pulse noise distribution degree of freedom parameter to obtain the iterated pulse noise distribution precision parameter matrix;

[0099] In this embodiment, the current pulse noise distribution precision parameter matrix is ​​iterated according to the current pulse noise matrix and the iterated pulse noise distribution degree of freedom parameter to obtain the iterated pulse noise distribution precision parameter matrix, including:

[0100] The current pulse noise distribution accuracy parameter matrix is ​​iterated using the following iterative formula:

[0101]

[0102]

[0103]

[0104] in, It means generating a diagonal matrix; Multiply corresponding elements of two vectors; Express Take the conjugate of each element of ; represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; M represents the number of elements in the array.

[0105] Step S204: iterating the current sparse sound source signal matrix according to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix to obtain the iterated sparse sound source signal matrix;

[0106] In this embodiment, the current sparse sound source signal matrix is ​​iterated according to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix to obtain the iterated sparse sound source signal matrix, including:

[0107] The current sparse sound source signal matrix is ​​iterated using the following iterative formula:

[0108]

[0109] in, Represents the column vector Each element of the matrix Multiply each element of the corresponding row; represents the sparse sound source signal matrix of the qth iteration.

[0110] In this embodiment, the column vector Each element includes:

[0111] The column vector is calculated by the following formula Each element of:

[0112]

[0113] Among them, the vector express The nth row of represents the sparse sound source signal matrix of the q-1th iteration; Representation vector The 2-norm of ; represents the regularization parameter; represents the algorithm convergence parameter; N represents the total number of grid points.

[0114] In this embodiment, the matrix ,include:

[0115] The matrix is ​​calculated by the following formula :

[0116]

[0117]

[0118]

[0119]

[0120]

[0121] in, represents the sparse sound source signal matrix of the q-1th iteration; represents the regularization parameter; represents the overcomplete array manifold matrix; Y represents the received data vector matrix; represents the impulse noise distribution accuracy parameter matrix of the qth iteration; represents the first value of the impulse noise matrix of the q-1th iteration The conjugate transpose of the impulse noise vector of the snapshot; represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; Indicates multiplying the corresponding elements of two matrices; Represents moment The M-element column vector consisting of the diagonal elements of the array; M represents the number of elements in the array.

[0122] Example 2, a simulation experiment was performed using MATLAB, and the experimental steps are as follows:

[0123] Step B1: Assume that there are two far-field targets K, and the target direction angle vector is , the target signal frequency is f=10KHz, the signal sampling rate is 192KHZ, and the underwater acoustic signal propagation rate is c=1500 m / s; assuming that the array is a uniform linear array, the number of array elements is M=10, and the spacing between array elements is , select array element 1 as the reference array element, the total number of snapshots L = 100; divide the spatial grid into , setting the number of grid points N = 181, the overcomplete array manifold matrix is , grid points The corresponding array of manifold vectors , It represents the time delay between the signal from the nth direction reaching the mth (1≤m≤M)th array element and reaching a pre-selected reference array element; the received data vector matrix is , 1≤l≤L represents the received data vector of the lth snapshot; the sparse sound source signal matrix is , represents the sparse sound source signal vector corresponding to the lth snapshot; the impulse noise matrix, impulse noise distribution precision parameter matrix and impulse noise distribution freedom parameter are generated using the complex Student T distribution with a degree of freedom parameter of 5. The expression of the impulse noise matrix is: , represents the impulse noise vector corresponding to the lth snapshot; using Define the ratio of the target signal to the noise power, where represents the average power of the target signal, Represents the average power of noise, and sets the ratio of target signal to noise power SNR=10;

[0124] Step B2: Initialize the sparse sound source signal matrix, the impulse noise distribution precision parameter matrix and the impulse noise distribution degree of freedom parameter; wherein the initial sparse sound source signal matrix is , that is, a 181×100-order all-zero matrix; the initial impulse noise distribution accuracy parameter matrix is , represents the 10th-order unit matrix; the initial impulse noise distribution degree of freedom parameter is ;

[0125] Step B3: Repeat the sparse sound source signal matrix iteration operation until the upper limit threshold of the number of iterations is reached or the change value of the optimized sparse sound source signal matrix is ​​less than the preset change threshold, and obtain the final sparse sound source signal matrix; wherein the upper limit threshold of the number of iterations is set to 200; the initial number of iterations q=0; the preset change threshold is set to 0.001;

[0126] Step B4: modulo all elements of the final sparse sound source signal matrix and add them row by row to obtain a spatial spectrum vector, search for the grid point corresponding to the maximum peak value based on the spatial spectrum vector, and obtain the direction of arrival angle based on the spatial angle corresponding to the grid point.

[0127] In this embodiment, the change value of the optimized sparse sound source signal matrix includes:

[0128] The change value of the optimized sparse sound source signal matrix is ​​calculated by the following calculation formula:

[0129]

[0130] in, represents the sparse sound source signal matrix of the qth iteration; represents the sparse sound source signal matrix of the q-1th iteration; represents the Frobenius norm, which means finding the square root of the sum of the squares of all elements of the matrix.

[0131] In this embodiment, the search is performed by taking the modulus of all elements of the final sparse sound source signal matrix and adding them row by row to obtain a spatial spectrum vector, searching for a grid point corresponding to the maximum peak value according to the spatial spectrum vector, and obtaining the direction of arrival angle according to the spatial angle corresponding to the grid point, including:

[0132] The spatial spectrum vector is obtained by the following calculation formula:

[0133]

[0134] According to the space spectrum vector Find the grid point set corresponding to the peak value in the spatial spectrum, that is ,according to The grid point corresponding to the largest K=2 peaks in the equation is used to obtain the direction of arrival angle;

[0135] In this embodiment, refer to Figure 4 , Figure 4 : A DOA simulation result diagram of a direction of arrival estimation method in an impulse noise scenario provided by an embodiment of the present invention; According to the DOA simulation result diagram, it can be seen that the spatial spectrum vector There are the largest K=2 peaks in the , and then the direction of arrival angle is obtained according to the grid points corresponding to the peaks.

[0136] Please refer to Figure 2 In this embodiment, the sparse sound source signal matrix iterative operation includes:

[0137] Step S201: Calculate the current impulse noise matrix according to the current sparse sound source signal matrix, the overcomplete array manifold matrix and the received data vector matrix; wherein the initial sparse sound source signal matrix is ​​the initial sparse sound source signal matrix;

[0138] In this embodiment, the current impulse noise matrix is ​​calculated according to the current sparse sound source signal matrix, the overcomplete array manifold matrix and the received data vector matrix, including:

[0139] The current impulse noise matrix is ​​calculated by the following expression:

[0140]

[0141] in, represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; Represents the first matrix of received data vectors The received data vector of snapshots; represents an overcomplete array manifold matrix; represents the sparse sound source signal matrix of the q-1th iteration The sparse sound source signal vector of snapshots, q represents the number of iterations.

[0142] Step S202: iterating the current pulse noise distribution degree of freedom parameters according to the current pulse noise matrix and the current pulse noise distribution precision parameter matrix to obtain the iterated pulse noise distribution degree of freedom parameters; wherein the initial pulse noise distribution precision parameter matrix is ​​the initial pulse noise distribution precision parameter matrix; the initial pulse noise distribution degree of freedom parameters are the initial pulse noise distribution degree of freedom parameters;

[0143] In this embodiment, the current pulse noise distribution degree of freedom parameter is iterated according to the current pulse noise matrix and the current pulse noise distribution precision parameter matrix to obtain the iterated pulse noise distribution degree of freedom parameter, including:

[0144] The current pulse noise distribution degree of freedom parameter is iterated by the following iterative formula:

[0145] (1)

[0146]

[0147] in, Indicates The qth iteration value of the degree of freedom parameter of the impulse noise distribution of the snapshot; represents the first value of the impulse noise matrix of the q-1th iteration The conjugate transpose of the impulse noise vector of the snapshot; represents the pulse noise distribution precision parameter matrix of the q-1th iteration; M represents the number of array elements in the array; m represents the mth array element; L represents the total number of snapshots;

[0148] In this embodiment, the steps of iterating the current impulse noise distribution degree of freedom parameter through the impulse noise distribution degree of freedom parameter iteration formula are as follows:

[0149] Step a1: Determine the value range of the freedom parameter of the impulse noise distribution as ; Among them, a=0.001; b=100000; U=100000;

[0150] Step a2: Calculate the function value and derivative value of the current pulse noise distribution degree of freedom parameter according to the current pulse noise distribution degree of freedom parameter and formula (1); wherein the initial pulse noise distribution degree of freedom parameter is the initial pulse noise distribution degree of freedom parameter;

[0151] Step a3: updating the current pulse noise distribution degree of freedom parameter according to the current pulse noise distribution degree of freedom parameter, the function value and the derivative value of the current pulse noise distribution degree of freedom parameter to obtain the updated pulse noise distribution degree of freedom parameter; when the updated pulse noise distribution degree of freedom parameter is less than or equal to a, the updated pulse noise distribution degree of freedom parameter is assigned a value; when the updated pulse noise distribution degree of freedom parameter is greater than or equal to b, the updated pulse noise distribution degree of freedom parameter is assigned b value;

[0152] Repeat step a1, step a2 and step a3 until the upper limit threshold of the pulse noise distribution freedom parameter update times is reached or the solution function value of the current pulse noise distribution freedom parameter in formula (1) is less than or equal to the preset tolerance, and the iterative pulse noise distribution freedom parameter is obtained; wherein the upper limit threshold of the pulse noise distribution freedom parameter update times is set to 200; the initial pulse noise distribution freedom parameter update times is 0; and the preset tolerance is 0.0001.

[0153] Step S203: iterating the current pulse noise distribution precision parameter matrix according to the current pulse noise matrix and the iterated pulse noise distribution degree of freedom parameter to obtain the iterated pulse noise distribution precision parameter matrix;

[0154] In this embodiment, the current pulse noise distribution precision parameter matrix is ​​iterated according to the current pulse noise matrix and the iterated pulse noise distribution degree of freedom parameter to obtain the iterated pulse noise distribution precision parameter matrix, including:

[0155] The current pulse noise distribution accuracy parameter matrix is ​​iterated using the following iterative formula:

[0156]

[0157]

[0158]

[0159] in, It means generating a diagonal matrix; It means multiplying the corresponding elements of two vectors; Express Take the conjugate of each element of ; represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of each snapshot; M=10 represents the number of elements in the array.

[0160] Step S204: iterating the current sparse sound source signal matrix according to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix to obtain the iterated sparse sound source signal matrix;

[0161] In this embodiment, the current sparse sound source signal matrix is ​​iterated according to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix to obtain the iterated sparse sound source signal matrix, including:

[0162] The current sparse sound source signal matrix is ​​iterated using the following iterative formula:

[0163]

[0164] in, Represents the column vector Each element of the matrix Multiply each element of the corresponding row; represents the sparse sound source signal matrix of the qth iteration.

[0165] In this embodiment, the column vector Each element includes:

[0166] The column vector is calculated by the following formula Each element of:

[0167]

[0168] Among them, the vector express The nth row of represents the sparse sound source signal matrix of the q-1th iteration; Representation vector The 2-norm of ; It means that the regularization parameter is set to 0.0001; indicates that the algorithm convergence parameter is set to 0.000001; N indicates that the total number of grid points is set to 181.

[0169] In this embodiment, the matrix ,include:

[0170] The matrix is ​​calculated by the following formula :

[0171]

[0172]

[0173]

[0174]

[0175]

[0176] in, represents the sparse sound source signal matrix of the q-1th iteration; represents the regularization parameter; represents the overcomplete array manifold matrix; Y represents the received data vector matrix; represents the impulse noise distribution accuracy parameter matrix of the qth iteration; represents the first value of the impulse noise matrix of the q-1th iteration The conjugate transpose of the impulse noise vector of the snapshot; represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; Indicates multiplying the corresponding elements of two matrices; Represents moment The diagonal elements of the array are composed of an M-element column vector; M indicates that the number of elements in the array is set to 10.

[0177] Embodiment 3:

[0178] Please refer to Figure 3 , is a system structure diagram of a wave direction angle estimation system in an impulse noise scenario provided by an embodiment of the present invention, the system comprising: a data acquisition module, an iteration module and a wave direction angle calculation module; wherein the iteration module comprises an impulse noise matrix calculation unit, an impulse noise distribution degree of freedom parameter iteration unit, an impulse noise distribution accuracy parameter matrix iteration unit and a sparse sound source signal matrix iteration unit;

[0179] The data acquisition module is used to obtain an overcomplete array manifold matrix, a received data vector matrix, an initial sparse sound source signal matrix, an initial impulse noise distribution precision parameter matrix, and an initial impulse noise distribution degree of freedom parameter corresponding to the array in the detection area;

[0180] The iteration module is used to repeatedly perform the sparse sound source signal matrix iteration operation until the upper limit threshold of the number of iterations is reached or the change value of the optimized sparse sound source signal matrix is ​​less than the preset change threshold, so as to obtain the final sparse sound source signal matrix;

[0181] The wave direction angle calculation module is used to perform spatial spectrum conversion and peak search according to the final sparse sound source signal matrix to obtain the wave direction angle;

[0182] The impulse noise matrix calculation unit is used to calculate the current impulse noise matrix according to the current sparse sound source signal matrix, the complete array manifold matrix and the received data vector matrix; wherein the initial sparse sound source signal matrix is ​​the initial sparse sound source signal matrix;

[0183] The pulse noise distribution degree of freedom parameter iteration unit is used to iterate the current pulse noise distribution degree of freedom parameter according to the current pulse noise matrix and the current pulse noise distribution precision parameter matrix to obtain the iterated pulse noise distribution degree of freedom parameter; wherein the initial pulse noise distribution precision parameter matrix is ​​the initial pulse noise distribution precision parameter matrix; the initial pulse noise distribution degree of freedom parameter is the initial pulse noise distribution degree of freedom parameter;

[0184] The pulse noise distribution precision parameter matrix iteration unit is used to iterate the current pulse noise distribution precision parameter matrix according to the current pulse noise matrix and the pulse noise distribution degree of freedom parameter after iteration to obtain the pulse noise distribution precision parameter matrix after iteration;

[0185] The sparse sound source signal matrix iteration unit is used to iterate the current sparse sound source signal matrix according to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix to obtain the iterated sparse sound source signal matrix.

[0186] It should be noted that the above-mentioned terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will understand that the above-mentioned terminal device is merely an example and does not constitute a limitation on the terminal device. It may include more or fewer components, or a combination of certain components, or different components.

[0187] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for estimating a direction of arrival angle in an impulse noise scenario, characterized in that: include: Obtaining an overcomplete array manifold matrix, a received data vector matrix, an initial sparse sound source signal matrix, an initial impulse noise distribution precision parameter matrix, and an initial impulse noise distribution degree of freedom parameter corresponding to the array in the detection area; Repeat the sparse sound source signal matrix iteration operation until the upper limit threshold of the number of iterations is reached or the change value of the optimized sparse sound source signal matrix is ​​less than the preset change threshold, to obtain the final sparse sound source signal matrix; Performing spatial spectrum conversion and peak search according to the final sparse sound source signal matrix to obtain a direction of arrival angle; The sparse sound source signal matrix iterative operation includes: Calculate the current impulse noise matrix according to the current sparse sound source signal matrix, the overcomplete array manifold matrix and the received data vector matrix; wherein the initial sparse sound source signal matrix is ​​the initial sparse sound source signal matrix; According to the current pulse noise matrix and the current pulse noise distribution precision parameter matrix, the current pulse noise distribution degree of freedom parameter is iterated to obtain the iterated pulse noise distribution degree of freedom parameter; wherein the initial pulse noise distribution precision parameter matrix is ​​the initial pulse noise distribution precision parameter matrix; the initial pulse noise distribution degree of freedom parameter is the initial pulse noise distribution degree of freedom parameter; the iterative formula for iterating the current pulse noise distribution degree of freedom parameter is: in, Indicates The qth iteration value of the degree of freedom parameter of the impulse noise distribution of the snapshot; represents the first value of the impulse noise matrix of the q-1th iteration The conjugate transpose of the impulse noise vector of the snapshot; represents the pulse noise distribution precision parameter matrix of the q-1th iteration; M represents the number of array elements in the array; m represents the mth array element; L represents the total number of snapshots; Iterate the current pulse noise distribution precision parameter matrix according to the current pulse noise matrix and the iterated pulse noise distribution degree of freedom parameter to obtain the iterated pulse noise distribution precision parameter matrix; According to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix, the current sparse sound source signal matrix is ​​iterated to obtain the iterated sparse sound source signal matrix.

2. The method for estimating direction of arrival in an impulse noise scenario according to claim 1, characterized in that: The step of calculating the current impulse noise matrix according to the current sparse sound source signal matrix, the overcomplete array manifold matrix and the received data vector matrix includes: The current impulse noise matrix is ​​calculated by the following expression: in, represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; Represents the first matrix of received data vectors The received data vector of snapshots; represents an overcomplete array manifold matrix; represents the sparse sound source signal matrix of the q-1th iteration The sparse sound source signal vector of snapshots, q represents the number of iterations.

3. The method for estimating direction of arrival in an impulse noise scenario according to claim 2, characterized in that: The method of iterating the current pulse noise distribution precision parameter matrix according to the current pulse noise matrix and the iterated pulse noise distribution degree of freedom parameter to obtain the iterated pulse noise distribution precision parameter matrix includes: The current pulse noise distribution accuracy parameter matrix is ​​iterated using the following iterative formula: in, It means generating a diagonal matrix; It means multiplying the corresponding elements of two vectors; Express Take the conjugate of each element of ; represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; M represents the number of elements in the array.

4. The method for estimating direction of arrival in an impulse noise scenario according to claim 3, characterized in that: The method iterates the current sparse sound source signal matrix according to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix to obtain the iterated sparse sound source signal matrix, including: The current sparse sound source signal matrix is ​​iterated using the following iterative formula: in, Represents the column vector Each element of the matrix Multiply each element of the corresponding row; represents the sparse sound source signal matrix of the qth iteration.

5. The method for estimating direction of arrival in an impulse noise scenario according to claim 4, characterized in that: The column vector Each element includes: The column vector is calculated by the following formula Each element of: Among them, the vector express The nth row of represents the sparse sound source signal matrix of the q-1th iteration; Representation vector The 2-norm of ; represents the regularization parameter; represents the algorithm convergence parameter; N represents the total number of grid points.

6. The method for estimating direction of arrival in an impulse noise scenario according to claim 5, characterized in that: The matrix ,include: The matrix is ​​calculated by the following formula : in, represents the sparse sound source signal matrix of the q-1th iteration; represents the regularization parameter; represents the overcomplete array manifold matrix; Y represents the received data vector matrix; represents the impulse noise distribution accuracy parameter matrix of the qth iteration; represents the first value of the impulse noise matrix of the q-1th iteration The conjugate transpose of the impulse noise vector of the snapshot; represents the first value of the impulse noise matrix of the q-1th iteration The impulse noise vector of the snapshots; Indicates multiplying the corresponding elements of two matrices; Represents taking the matrix An M-element column vector consisting of the diagonal elements of ; M represents the number of elements in the array.

7. The method for estimating direction of arrival in an impulse noise scenario according to claim 6, characterized in that: The change value of the optimized sparse sound source signal matrix includes: The change value of the optimized sparse sound source signal matrix is ​​calculated by the following calculation formula: in, represents the sparse sound source signal matrix of the qth iteration; represents the sparse sound source signal matrix of the q-1th iteration; represents the Frobenius norm, which means finding the square root of the sum of the squares of all elements of the matrix.

8. The method for estimating direction of arrival in an impulse noise scenario according to claim 7, characterized in that: The performing spatial spectrum conversion and peak search according to the final sparse sound source signal matrix to obtain the direction of arrival angle includes: Taking the modulus of all elements in the final sparse sound source signal matrix and adding them row by row to obtain a spatial spectrum vector; The maximum peak value of the spatial spectrum vector is searched, and the direction of arrival angle is obtained according to the spatial angle grid point corresponding to the maximum peak value.

9. A direction of arrival angle estimation system in an impulse noise scenario, characterized in that: include: A data acquisition module, an iteration module and a wave direction of arrival angle calculation module; wherein the iteration module includes an impulse noise matrix calculation unit, an impulse noise distribution degree of freedom parameter iteration unit, an impulse noise distribution precision parameter matrix iteration unit and a sparse sound source signal matrix iteration unit; The data acquisition module is used to obtain an overcomplete array manifold matrix, a received data vector matrix, an initial sparse sound source signal matrix, an initial impulse noise distribution precision parameter matrix, and an initial impulse noise distribution degree of freedom parameter corresponding to the array in the detection area; The iteration module is used to repeatedly perform the sparse sound source signal matrix iteration operation until the upper limit threshold of the number of iterations is reached or the change value of the optimized sparse sound source signal matrix is ​​less than the preset change threshold, so as to obtain the final sparse sound source signal matrix; The wave direction angle calculation module is used to perform spatial spectrum conversion and peak search according to the final sparse sound source signal matrix to obtain the wave direction angle; The impulse noise matrix calculation unit is used to calculate the current impulse noise matrix according to the current sparse sound source signal matrix, the overcomplete array manifold matrix and the received data vector matrix; wherein the initial sparse sound source signal matrix is ​​the initial sparse sound source signal matrix; The pulse noise distribution degree of freedom parameter iteration unit is used to iterate the current pulse noise distribution degree of freedom parameter according to the current pulse noise matrix and the current pulse noise distribution precision parameter matrix to obtain the iterated pulse noise distribution degree of freedom parameter; wherein the initial pulse noise distribution precision parameter matrix is ​​the initial pulse noise distribution precision parameter matrix; the initial pulse noise distribution degree of freedom parameter is the initial pulse noise distribution degree of freedom parameter; the iteration formula for iterating the current pulse noise distribution degree of freedom parameter is: in, Indicates The qth iteration value of the degree of freedom parameter of the impulse noise distribution of the snapshot; represents the first value of the impulse noise matrix of the q-1th iteration The conjugate transpose of the impulse noise vector of the snapshot; represents the pulse noise distribution precision parameter matrix of the q-1th iteration; M represents the number of array elements in the array; m represents the mth array element; L represents the total number of snapshots; The pulse noise distribution precision parameter matrix iteration unit is used to iterate the current pulse noise distribution precision parameter matrix according to the current pulse noise matrix and the pulse noise distribution degree of freedom parameter after iteration to obtain the pulse noise distribution precision parameter matrix after iteration; The sparse sound source signal matrix iteration unit is used to iterate the current sparse sound source signal matrix according to the current impulse noise matrix, the iterated impulse noise distribution degree of freedom parameter, the iterated impulse noise distribution precision parameter matrix, the overcomplete array manifold matrix and the received data vector matrix to obtain the iterated sparse sound source signal matrix.

Images