Multi-constraint underwater sound source localization method and system based on improved relative entropy distance

Abstract

The present invention belongs to the technical field of underwater acoustic array signal processing, and discloses a multi-constraint underwater sound source positioning method and system based on improved relative entropy distance, including: obtaining the sound source signal spectrum, extracting the expected signal frequency included in the spectrum, and calculating the covariance matrix after the expected signal frequency is normalized; reconstructing the matrix; dividing the grid points within a preset range; obtaining environmental parameters, performing K random perturbations on the environmental parameters, and obtaining the perturbation matrix at each grid point; calculating and obtaining a multi-constraint copy matrix, and performing normalization processing on the multi-constraint copy matrix to obtain the normalized multi-constraint copy matrix; calculating the improved relative entropy distance between the reconstructed covariance matrix and the normalized multi-constraint copy matrix; obtaining the normalized output power at each grid point based on the improved relative entropy distance; performing spectrum peak search on the normalized output power, and the grid point position corresponding to the maximum peak value is the position of the sound source signal. The present invention has the advantages of small positioning error and high positioning resolution.

Description

Technical Field

[0001] The present invention relates to the technical field of underwater acoustic array signal processing, and in particular to a multi-constraint underwater sound source positioning method and system based on improved relative entropy distance. Background Art

[0002] The sound velocity profile, spatial structure and seabed composition of the ocean waveguide play a decisive role in the sound propagation in the random ocean environment. The signal emitted by the sound source contains the sound field information after propagating in the ocean waveguide. The matching field uses the sound propagation characteristics to compare the sound field of the real sound source with the sound field of the assumed copy sound source, traverse the copy sound sources on all grid points, and select the position with the largest correlation as the position of the underwater sound source.

[0003] However, the matching field still faces many challenges when applied to the actual underwater positioning problem. The characteristics of the underwater acoustic channel that vary with time and space and the measurement errors of environmental information will introduce random factors into the sound field calculation, resulting in environmental parameter mismatch. When there is a difference between the estimated sample covariance matrix and the ideal set covariance matrix, statistical mismatch will occur. These mismatches may cause the fuzzy plane peak of the matching field processor to deviate from the true sound source position, affecting the positioning accuracy and may even lead to positioning failure. Classic matching field methods, including linear processors (Bartlett) and adaptive processors based on minimum variance distortionless response (MVDR), have their own shortcomings. The former may have sidelobe fuzziness and low mainlobe width positioning resolution, while the latter is more sensitive to environmental mismatch. Therefore, it is necessary to propose a matching field positioning method that is robust to environmental mismatch and statistical mismatch. Summary of the invention

[0004] The purpose of the present invention is to provide a multi-constraint underwater sound source localization method and system based on improved relative entropy distance, so as to overcome the defects of the prior art that the prior art is not robust to environmental mismatch and statistical mismatch.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] In a first aspect, the present invention provides a multi-constraint underwater sound source localization method based on improved relative entropy distance, comprising the following steps:

[0007] S1. Obtain the sound source signal, extract the spectrum of the sound source signal, obtain the frequency of the desired signal included in the spectrum, calculate the covariance matrix of the frequency, normalize the covariance matrix, and obtain the normalized covariance matrix K f ;

[0008] S2. Reconstruct the covariance matrix K f , and obtain the reconstructed covariance matrix

[0009] S3. Divide the grid points within the preset sound source search range. Assume that grid point a i is the sound source position, where i=1…J, J is the number of grid points;

[0010] S4. Obtain environmental parameters, perform K random perturbations on the environmental parameters, and obtain K random environmental parameter vectors; calculate K copy field vectors at each grid point based on the K random environmental parameter vectors, and obtain a perturbation matrix at each grid point by combining the K copy field vectors;

[0011] S5. Conjugate multiply the perturbation matrix to obtain the multi-constraint copy matrix, normalize the multi-constraint copy matrix, and obtain the normalized multi-constraint copy matrix for each grid point.

[0012] S6. Calculate the reconstructed covariance matrix with the multi-constraint copy matrix at each grid point The improved relative entropy distance includes the following sub-steps:

[0013] S60. Matrix The inverse of the matrix The matrix obtained by multiplication is subjected to eigendecomposition to obtain the maximum eigenvalue λ of the signal component max ;

[0014] S61. Reconstructed covariance matrix with the multi-constraint copy matrix at each grid point The improved relative entropy distance is:

[0015]

[0016] S7. Obtaining the normalized output power at each grid point based on the improved symmetric relative entropy distance;

[0017] S8. Perform spectrum peak search on the normalized output power, and the grid point position corresponding to the maximum peak value is the position of the sound source signal.

[0018] As a possible implementation, the sound source signal is obtained by placing an array composed of hydrophone array elements underwater to receive the sound source signal, and the array is a horizontal, inclined or vertical array.

[0019] As a possible implementation method, S2 specifically includes:

[0020] S20. Covariance matrix K f Perform singular value decomposition: K f =XΛ s Y H ;

[0021] Among them, X is the left singular matrix, Y is the right singular matrix, Λ s is a diagonal matrix;

[0022] S21. Select the largest singular value σ0 and its corresponding columns x1, y1 of the left and right singular value vectors for reconstruction to obtain the reconstructed covariance matrix:

[0023] As a possible implementation method, obtaining environmental parameters specifically includes: obtaining environmental parameters by on-site measurement, or obtaining environmental parameters by simulating a preset environment; the environmental parameters at least include: seawater depth, sound velocity profile, and seabed sediment parameters.

[0024] As a possible implementation method, K copy field vectors at each grid point are calculated based on K random environment parameter vectors. Specifically, the K copy field vectors at each grid point are calculated using a sound field model.

[0025] As a possible implementation, the normalized output power at each grid point is: The reconstructed covariance matrix with the multi-constraint copy matrix at each grid point The inverse of the improved relative entropy distance.

[0026] As a possible implementation manner, the preset sound source search range is (R, Z), wherein R represents the horizontal distance of the search area from the array, and Z represents the depth of the search area from the water surface.

[0027] In a second aspect, the present invention provides a multi-constraint underwater sound source positioning system based on improved relative entropy distance, comprising:

[0028] The sound source signal acquisition unit adopts an array composed of hydrophone array elements and is placed underwater to receive the sound source signal. The array is a horizontal, inclined or vertical array.

[0029] The frequency covariance matrix solving unit of the normalized expected signal extracts the spectrum of the sound source signal, obtains the frequency of the expected signal included in the spectrum, calculates the frequency covariance matrix, normalizes the covariance matrix, and obtains the normalized covariance matrix K f ;

[0030] Covariance matrix K f Reconstruction unit, covariance matrix K f Perform singular value decomposition, select the maximum singular value σ0 and its corresponding columns x1, y1 of the left and right singular value vectors for reconstruction, and obtain the reconstructed covariance matrix:

[0031] The multi-constraint copy matrix solving unit after normalization of each grid point divides the grid points within the preset sound source search range. Assume that the grid point a i is the sound source position, where i=1…J, J is the number of grid points; obtain environmental parameters, perform K random perturbations on the environmental parameters, and obtain K random environmental parameter vectors; calculate K copy field vectors at each grid point based on the K random environmental parameter vectors, and obtain the perturbation matrix at each grid point by combining the K copy field vectors; conjugate multiply the perturbation matrix to obtain a multi-constraint copy matrix, normalize the multi-constraint copy matrix, and obtain the normalized multi-constraint copy matrix at each grid point

[0032] Improve the relative entropy distance solving unit, for the matrix The inverse of the matrix The matrix obtained by multiplication is subjected to eigendecomposition to obtain the maximum eigenvalue λ of the signal component max ; The reconstructed covariance matrix with the multi-constraint copy matrix at each grid point The improved relative entropy distance is:

[0033]

[0034] The location determination unit of the sound source signal obtains the normalized output power at each grid point based on the improved symmetric relative entropy distance, performs spectrum peak search on the normalized output power, and the grid point location corresponding to the maximum peak value is the location of the sound source signal.

[0035] In a third aspect, the present invention provides a terminal comprising a processor and a communication interface coupled to the processor, wherein the processor is used to run a computer program or instruction to implement the multi-constraint underwater sound source positioning method based on improved relative entropy distance provided in the first aspect.

[0036] Compared with the prior art, the present invention has the following technical effects:

[0037] 1. The multi-constraint underwater sound source localization method based on improved relative entropy distance provided by the present invention can be applied to the passive localization of underwater sound sources of hydrophone arrays under environmental mismatch and statistical mismatch conditions.

[0038] 2. The multi-constraint underwater sound source localization method based on improved relative entropy distance provided by the present invention maps the matrix to the matrix manifold in the form of points by utilizing the symmetric positive definite geometric structure of the covariance matrix and the copy constraint matrix within the framework of information geometry. The underwater sound source localization is achieved by quantifying the similarity between matrices as the geodesic distance between points on the manifold, which has the advantages of being robust to environmental mismatch and statistical mismatch and having a small positioning error.

[0039] 3. The multi-constraint underwater sound source localization method based on improved relative entropy distance provided by the present invention has simulation and experimental results showing that compared with the traditional matching field positioning method, the present invention has a higher sound source localization success rate and a smaller positioning error in the presence of environmental mismatch and statistical mismatch, is robust to environmental and statistical mismatches, can also reduce side lobes and improve resolution, and has greater positioning advantages in complex ocean environments.

[0040] 4. The existing matching field positioning technology based on relative entropy distance has high side lobes in the fuzzy plane, which is prone to fuzzy positioning results, and is easily affected by environmental mismatch, resulting in positioning errors or even positioning failures. The present invention uses improved relative entropy to make the side lobes of the fuzzy plane lower, which is less prone to positioning ambiguity. It also uses a multi-constraint method to enhance the robustness of the matching field positioning technology to environmental mismatch and improve positioning accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The drawings described herein are used to provide further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0042] Figure 1 A flow chart of a multi-constraint underwater sound source localization method based on improved relative entropy distance provided by an embodiment of the present invention;

[0043] Figure 2 The relationship between the probability of correct positioning and the signal-to-noise ratio of the conventional linear processor, the MV-EPC processor and the MJS-EPC processor when the number of array hydrophones is fixed to 21 in the Monte Carlo simulation of the embodiment of the present invention;

[0044] Figure 3 The relationship between the probability of correct positioning of the conventional linear processor, the MV-EPC processor and the MJS-EPC processor and the number of array hydrophones in the Monte Carlo simulation in the embodiment of the present invention when the signal-to-noise ratio is fixed at 10 dB;

[0045] Figure 4 The relationship between PBR and signal-to-noise ratio of the conventional linear processor, the MV-EPC processor and the MJS-EPC processor when the number of array hydrophones is fixed to 21 in the Monte Carlo simulation of the embodiment of the present invention;

[0046] Figure 5 The relationship between the PBR and the number of array hydrophones of the conventional linear processor, the MV-EPC processor and the MJS-EPC processor when the signal-to-noise ratio is fixed at 10 dB in the Monte Carlo simulation in the embodiment of the present invention;

[0047] Figure 6 It is the fuzzy plane of single sound source localization by conventional linear processor in SWellEx96 test;

[0048] Figure 7 It is the fuzzy plane used by the MV-EPC processor to locate a single sound source in the SWellEx96 experiment;

[0049] Figure 8 It is the fuzzy plane used by the MJS-EPC processor to locate a single sound source in the SWellEx96 experiment;

[0050] Fig. 9 It is the horizontal distance slice of the blur plane of Bartlett processor, MV-EPC processor and MJS-EPC processor in SWellEx96 experiment;

[0051] Fig.10 Depth slices of the blur planes for the Bartlett, MV-EPC, and MJS-EPC processors in the SWellEx96 experiment. DETAILED DESCRIPTION

[0052] In order to clearly describe the technical solutions of the embodiments of the present invention, in the embodiments of the present invention, words such as "first" and "second" are used to distinguish the same items or similar items with basically the same functions and effects. For example, the first threshold and the second threshold are only used to distinguish different thresholds, and their order is not limited. Those skilled in the art can understand that words such as "first" and "second" do not limit the quantity and execution order, and words such as "first" and "second" do not necessarily limit them to be different.

[0053] It should be noted that, in the present invention, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0054] In the present invention, "at least one" means one or more, and "plurality" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, the combination of a and b, the combination of a and c, the combination of b and c, or the combination of a, b and c, where a, b, c can be single or multiple.

[0055] When the matching field is applied to actual underwater positioning problems, the positioning accuracy will be affected due to mismatch, and it may even lead to positioning failure. In order to solve the defects of the prior art that it is not robust to environmental mismatch and statistical mismatch, the present invention provides a multi-constraint underwater sound source localization method and system based on improved relative entropy distance, which has smaller positioning error, lower sidelobe level and higher positioning resolution.

[0056] In a first aspect, the present invention provides a multi-constraint underwater sound source localization method based on improved relative entropy distance, see Figure 1 , including the following steps:

[0057] S1. Obtain the sound source signal, extract the spectrum of the sound source signal, obtain the frequency of the desired signal included in the spectrum, calculate the covariance matrix of the frequency, normalize the covariance matrix, and obtain the normalized covariance matrix K f ;

[0058] As a possible implementation, the sound source signal is obtained by placing an array composed of hydrophone array elements underwater to receive the sound source signal, and the array is a horizontal, inclined or vertical array.

[0059] As an example, the spectrum of the sound source signal is extracted by short-time Fourier transform or spectrum estimation method, the frequency of the desired signal included in the spectrum is obtained, the covariance matrix K of the frequency is calculated, and the covariance matrix K is normalized by Frobenius norm to obtain the normalized covariance matrix K f It is expressed as:

[0060]

[0061] Among them, ||·|| F represents the Frobenius norm, k m,c represents the element in the mth row and cth column of the matrix K, where M is the dimension of the matrix K;

[0062] S2. Reconstruct the covariance matrix K f , and obtain the reconstructed covariance matrix

[0063] As a possible implementation method, S2 specifically includes:

[0064] S20. Covariance matrix K f Perform singular value decomposition: K f =XΛ s Y H ;

[0065] Among them, X is the left singular matrix, Y is the right singular matrix, Λ s is a diagonal matrix;

[0066] S21. Select the largest singular value σ0 and its corresponding columns x1, y1 of the left and right singular value vectors for reconstruction to obtain the reconstructed covariance matrix:

[0067] S3. Divide the grid points within the preset sound source search range. Assume that grid point a i is the sound source position, where i=1…J, J is the number of grid points;

[0068] As a possible implementation manner, the preset sound source search range is (R, Z), wherein R represents the horizontal distance of the search area from the array, and Z represents the depth of the search area from the water surface.

[0069] S4. Obtain environmental parameters, perform K random perturbations on the environmental parameters, and obtain K random environmental parameter vectors; calculate K copy field vectors at each grid point based on the K random environmental parameter vectors, and obtain a perturbation matrix at each grid point by combining the K copy field vectors;

[0070] As a possible implementation method, K copy field vectors at each grid point are calculated based on K random environment parameter vectors. Specifically, the K copy field vectors at each grid point are calculated using a sound field model.

[0071] As an example, let the environmental parameter be ψ k Based on the simple normal wave acoustic field model, the copy field vector at the grid point ai = (ri, zi) is calculated as g(r i ,z i ,ψ k ), where ri represents the horizontal distance between the grid point ai and the array, z i Represents the grid point a i At the depth in the seawater, i=1…J, where J is the number of grid points; the perturbation matrix at each grid point is obtained by combining the K copy field vector sets:

[0072] C(r i ,z i )=[g(r i ,z i ,ψ1),g(r i ,z i ,ψ2),...,g(r i ,z i ,ψ K )];

[0073] As a possible implementation method, obtaining environmental parameters specifically includes: obtaining environmental parameters by on-site measurement, or obtaining environmental parameters by simulating a preset environment; the environmental parameters at least include: seawater depth, sound velocity profile, and seabed sediment parameters.

[0074] S5. Conjugate multiply the perturbation matrix to obtain the multi-constraint copy matrix, normalize the multi-constraint copy matrix, and obtain the normalized multi-constraint copy matrix for each grid point.

[0075] As an example, conjugate multiply the perturbation matrix to obtain the multi-constraint copy matrix:

[0076]

[0077] Perform Frobenius norm normalization on the multi-constraint copy matrix to obtain the normalized multi-constraint copy matrix for each grid point

[0078]

[0079] At this time, the multi-constraint copy matrix is a symmetric positive definite matrix; d m,c Representation Matrix The element in the mth row and cth column of Dimension.

[0080] S6. Calculate the reconstructed covariance matrix with the multi-constraint copy matrix at each grid point The improved relative entropy distance includes the following sub-steps:

[0081] S60. Matrix The inverse of the matrix The matrix obtained by multiplication is subjected to eigendecomposition to obtain the maximum eigenvalue λ of the signal component max ;

[0082] S61. Reconstructed covariance matrix with the multi-constraint copy matrix at each grid point The improved relative entropy distance is:

[0083]

[0084] The existing matching field positioning technology based on relative entropy distance has high side lobes in the fuzzy plane, which is prone to fuzzy positioning results, and is easily affected by environmental mismatch, resulting in positioning errors or even positioning failures. The present invention uses improved relative entropy to make the side lobes of the fuzzy plane lower, which is less prone to positioning ambiguity. It also uses a multi-constraint method to enhance the robustness of the matching field positioning technology to environmental mismatch and improve positioning accuracy.

[0085] S7. Obtaining the normalized output power at each grid point based on the improved symmetric relative entropy distance;

[0086] As a possible implementation, the normalized output power at each grid point is: The reconstructed covariance matrix with the multi-constraint copy matrix at each grid point The inverse of the improved relative entropy distance:

[0087]

[0088] S8. Perform spectrum peak search on the normalized output power, and the grid point position corresponding to the maximum peak value is the position of the sound source signal.

[0089] In a possible implementation, the multi-constraint underwater sound source localization method based on improved relative entropy distance provided by the present invention is referred to as MJS-EPC processor. In the Monte Carlo simulation, the waveguide model of the SWellEx-96 test and 21 sensors are used to form a vertical array. The sound source is located 3500m away from the array and 56m below the water surface. In order to simulate the situation where the number of snapshots is insufficient in the experiment and causes statistical mismatch, the number of snapshots is set to 21. Determine the search range (R, Z) of the sound source, where the horizontal search range is R=0.02km-10km, and the depth search range is Z=2m-200m. A grid is divided every 20m in horizontal distance, with a total of 500 grid points, and a grid is divided every 2m in depth, with a total of 100 grid points. The grid point a i =(r i ,z i ), i=1…J, with a total number of J=50000. Under different signal-to-noise ratios and numbers of array hydrophones, random perturbations were imposed on environmental parameters to simulate randomly changing ocean environments. 3 In order to maintain a single variable, the same multi-constraint copy matrix is ​​used for the minimum variance environmental perturbation constraint matching field (MV-EPC) method and the multi-constraint based on improved relative entropy distance (MJS-EPC) method. The simulation was carried out under the above conditions and the results are shown in Figures 2 to 3 .

[0090] Figure 2 The relationship between the probability of correct positioning and the signal-to-noise ratio of the conventional linear processor, MV-EPC processor and MJS-EPC processor when the number of array hydrophones is fixed to 21; Figure 3 The figure shows the relationship between the probability of correct positioning of the conventional linear processor, MV-EPC processor and MJS-EPC processor and the number of array hydrophones when the signal-to-noise ratio is fixed at 10 dB. It can be seen that the MJS-EPC processor proposed in the present invention has the highest positioning success rate among the three processors when the signal-to-noise ratio is greater than 0 dB and the number of hydrophones is greater than 15, and the probability of correct positioning when the signal-to-noise ratio is 0 dB is about 10% higher than that of the other two methods.

[0091] See also Figures 4 to 5 , Figure 4 The relationship between the peak-to-background ratio (PBR) and the signal-to-noise ratio of the conventional linear processor, MV-EPC processor, and MJS-EPC processor when the number of array hydrophones is fixed at 21; Figure 5 The relationship between the PBR and the number of array hydrophones of the conventional linear processor, MV-EPC processor and MJS-EPC processor when the signal-to-noise ratio is fixed at 10 dB. It can be seen that the MJS-EPC processor proposed in the present invention has the highest peak-to-background ratio and the best sidelobe performance among the three processors. The performance is more prominent at low signal-to-noise ratio, and the peak-to-background ratio is about 1 dB higher than that of MV-EPC when the signal-to-noise ratio is high.

[0092] In another possible implementation, conventional linear processors, MV-EPC processors, and MJS-EPC processors are used to process the SWellEx96 test data. The SWellEx96 test uses a vertical array of 21 sensors, and the sound source is located 3500m away from the array and 56m below the water surface. After extracting the frequency information of the 201Hz sound source signal, due to the limitation of the sampling frequency, 21 snapshots are used to calculate the covariance matrix. Determine the search range (R, Z) of the sound source, where the horizontal search range is R = 0.02km-10km, and the depth search range is Z = 2m-200m. A grid is divided every 20m in horizontal distance, with a total of 500 grid points. A grid is divided every 2m in depth, with a total of 100 grid points. The grid point a is i =(r i ,z i ), i=1…J, where the total number J=50000.

[0093] See also Figures 6 to 8 , Figure 6 It is the fuzzy plane of single sound source localization by conventional linear processor in SWellEx96 test; Figure 7It is the fuzzy plane used by the MV-EPC processor to locate a single sound source in the SWellEx96 experiment; Figure 8 It is the fuzzy plane of the MJS-EPC processor for single sound source positioning in the SWellEx96 test. The white box in the figure represents the actual sound source position. It can be seen that the positioning result of the MJS-EPC processor proposed in the present invention is closest to the actual sound source position. The horizontal distance positioning error is reduced by about 200m compared with the other two processors, and the depth positioning error is reduced by about 10m compared with the other two processors.

[0094] See also Figures 9 and 10 , Fig. 9 It is the horizontal distance slice of the blur plane of Bartlett processor, MV-EPC processor and MJS-EPC processor in SWellEx96 experiment; Fig.10 The depth sections of the blur planes of the Bartlett processor, MV-EPC processor and MJS-EPC processor in the SWellEx96 test are shown in Figure 1. It can be seen that the MJS-EPC processor proposed in the present invention has the lowest side lobe, the narrowest main lobe and the highest resolution.

[0095] In a second aspect, the present invention provides a multi-constraint underwater sound source positioning system based on improved relative entropy distance, comprising:

[0096] The sound source signal acquisition unit adopts an array composed of hydrophone array elements and is placed underwater to receive the sound source signal. The array is a horizontal, inclined or vertical array.

[0097] The frequency covariance matrix solving unit of the normalized expected signal extracts the spectrum of the sound source signal, obtains the frequency of the expected signal included in the spectrum, calculates the frequency covariance matrix, normalizes the covariance matrix, and obtains the normalized covariance matrix K f ;

[0098] Covariance matrix K f Reconstruction unit, covariance matrix K f Perform singular value decomposition, select the maximum singular value σ0 and its corresponding columns x1, y1 of the left and right singular value vectors for reconstruction, and obtain the reconstructed covariance matrix:

[0099] The multi-constraint copy matrix solving unit after normalization of each grid point divides the grid points within the preset sound source search range. Assume that the grid point a iis the sound source position, where i=1…J, J is the number of grid points; obtain environmental parameters, perform K random perturbations on the environmental parameters, and obtain K random environmental parameter vectors; calculate K copy field vectors at each grid point based on the K random environmental parameter vectors, and obtain the perturbation matrix at each grid point by combining the K copy field vectors; conjugate multiply the perturbation matrix to obtain a multi-constraint copy matrix, normalize the multi-constraint copy matrix, and obtain the normalized multi-constraint copy matrix at each grid point

[0100] Improve the relative entropy distance solving unit, for the matrix The inverse of the matrix The matrix obtained by multiplication is subjected to eigendecomposition to obtain the maximum eigenvalue λ of the signal component max ; The reconstructed covariance matrix with the multi-constraint copy matrix at each grid point The improved relative entropy distance is:

[0101]

[0102] The location determination unit of the sound source signal obtains the normalized output power at each grid point based on the improved symmetric relative entropy distance, performs spectrum peak search on the normalized output power, and the grid point location corresponding to the maximum peak value is the location of the sound source signal.

[0103] In a third aspect, the present invention provides a terminal comprising a processor and a communication interface coupled to the processor, wherein the processor is used to run a computer program or instruction to implement the multi-constraint underwater sound source positioning method based on improved relative entropy distance provided in the first aspect.

[0104] Although the present invention is described herein in conjunction with various embodiments, in the process of implementing the claimed invention, those skilled in the art may understand and implement other variations of the disclosed embodiments by viewing the drawings, the disclosure, and the drawings, etc. In the specification, the word "comprising" does not exclude other components or steps, and "one" or "an" does not exclude multiple situations. A single processor or other unit can implement several functions listed in the specification. Certain measures are recorded in different embodiments, but this does not mean that these measures cannot be combined to produce good results.

[0105] Although the present invention has been described in conjunction with specific features and embodiments thereof, it is apparent that various modifications and combinations may be made thereto without departing from the spirit and scope of the present invention. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present invention and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present invention. Obviously, those skilled in the art may make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, the present invention is intended to include such modifications and variations if they fall within the scope of the present invention and its equivalents.

Claims

1. A multi-constraint underwater sound source localization method based on improved relative entropy distance, characterized in that: The steps include: S1. Obtain a sound source signal, extract the spectrum of the sound source signal, obtain the frequency of the desired signal included in the spectrum, calculate the covariance matrix of the frequency, normalize the covariance matrix, and obtain the normalized covariance matrix K f ; S2. Reconstruct the covariance matrix K f , and obtain the reconstructed covariance matrix S3. Divide the grid points within the preset sound source search range. Assume that grid point a i is the sound source position, where i=1…J, J is the number of grid points; S4. Obtain environmental parameters, perform K random perturbations on the environmental parameters to obtain K random environmental parameter vectors; calculate K copy field vectors at each grid point based on the K random environmental parameter vectors, and obtain a perturbation matrix at each grid point by combining the K copy field vectors: C(r i ,z i )=[g(r i ,z i ,ψ1),g(r i ,z i ,ψ2),...,g(r i ,z i ,ψ K )]; Among them, ψ k represents environmental parameters, r i Represents the grid point a i Horizontal distance from the array, z i Represents the grid point a i Depth in seawater, i = 1…J, J is the number of grid points, g(r i ,z i ,ψ k ) represents the copy field vector at the grid point ai=(ri,zi); S5. Conjugate multiply the perturbation matrix to obtain a multi-constraint copy matrix: The multi-constraint copy matrix is ​​normalized to obtain the normalized multi-constraint copy matrix of each grid point. Among them, d m,c Representation Matrix The element in the mth row and cth column of Dimensions; S6. Calculate the reconstructed covariance matrix with the multi-constraint copy matrix at each grid point The improved relative entropy distance includes the following sub-steps: S60. Reconstructed covariance matrix The inverse of and the normalized multi-constraint copy matrix at each grid point The matrix obtained by multiplication is subjected to eigendecomposition to obtain the maximum eigenvalue λ containing the signal component max ; S61. Reconstructed covariance matrix The normalized multi-constraint copy matrix for each grid point The improved relative entropy distance is: S7. Obtaining a normalized output power at each of the grid points based on the improved relative entropy distance; S8. Perform spectrum peak search on the normalized output power, and the grid point position corresponding to the maximum peak value is the position of the sound source signal.

2. The multi-constraint underwater sound source localization method based on improved relative entropy distance according to claim 1 is characterized in that: The acquisition of the sound source signal is performed by placing an array composed of hydrophone array elements underwater to receive the sound source signal, and the array is a horizontal, inclined or vertical array.

3. The multi-constraint underwater sound source localization method based on improved relative entropy distance according to claim 1 is characterized in that: The S2 specifically includes: S20. For the covariance matrix K f Perform singular value decomposition: K f =XΛ s Y H ; Among them, X is the left singular matrix, Y is the right singular matrix, Λ s is a diagonal matrix; S21. Select the largest singular value σ0 and its corresponding columns x1, y1 of the left and right singular value vectors for reconstruction to obtain the reconstructed covariance matrix:

4. The multi-constraint underwater sound source localization method based on improved relative entropy distance according to claim 1 is characterized in that: Acquiring environmental parameters specifically includes: obtaining environmental parameters by on-site measurement, or obtaining environmental parameters by simulating a preset environment; the environmental parameters at least include: seawater depth, sound velocity profile, and seabed sediment parameters.

5. The multi-constraint underwater sound source localization method based on improved relative entropy distance according to claim 1, characterized in that: The K copy field vectors at each grid point are calculated based on the K random environment parameter vectors, specifically: the K copy field vectors at each grid point are calculated using a sound field model.

6. The multi-constraint underwater sound source localization method based on improved relative entropy distance according to claim 1, characterized in that: The normalized output power at each grid point is: with the multi-constraint copy matrix at each grid point The inverse of the improved relative entropy distance.

7. The multi-constraint underwater sound source localization method based on improved relative entropy distance according to claim 2 is characterized in that: The preset sound source search range is (R, Z), where R represents the horizontal distance of the search area from the array, and Z represents the depth of the search area from the water surface.

8. A multi-constraint underwater sound source localization system based on improved relative entropy distance, characterized in that: include: The frequency covariance matrix solving unit of the normalized expected signal obtains the sound source signal, extracts the spectrum of the sound source signal, obtains the frequency of the expected signal included in the spectrum, calculates the covariance matrix of the frequency, normalizes the covariance matrix, and obtains the normalized covariance matrix K f ; The multi-constraint copy matrix solving unit after normalization of each grid point divides the grid points within the preset sound source search range. Assume that the grid point a i is the sound source position, where i=1…J, J is the number of grid points; obtain environmental parameters, perform K random perturbations on the environmental parameters to obtain K random environmental parameter vectors; calculate K copy field vectors at each grid point based on the K random environmental parameter vectors, and obtain the perturbation matrix at each grid point by combining the K copy field vectors: C(r i ,z i )=[g(r i ,z i ,ψ1),g(r i ,z i ,ψ2),...,g(r i ,z i ,ψ K )]; Among them, ψ k represents environmental parameters, r i Represents the grid point a i Horizontal distance from the array, z i Represents the grid point a i Depth in seawater, i = 1…J, J is the number of grid points, g(r i ,z i ,ψ k ) represents the copy field vector at the grid point ai=(ri,zi); The perturbation matrix is ​​conjugate multiplied to obtain the multi-constraint copy matrix: The multi-constraint copy matrix is ​​normalized to obtain the normalized multi-constraint copy matrix of each grid point. Among them, d m,c Representation Matrix The element in the mth row and cth column of Dimensions; Improve the relative entropy distance solving unit and reconstruct the covariance matrix K f , and obtain the reconstructed covariance matrix The reconstructed covariance matrix The inverse of and the normalized multi-constraint copy matrix at each grid point The matrix obtained by multiplication is subjected to eigendecomposition to obtain the maximum eigenvalue λ containing the signal component max ; The reconstructed covariance matrix The normalized multi-constraint copy matrix for each grid point The improved relative entropy distance is: The location determination unit of the sound source signal obtains the normalized output power at each of the grid points based on the improved relative entropy distance, performs a spectrum peak search on the normalized output power, and the grid point location corresponding to the maximum peak value is the location of the sound source signal.

9. A terminal comprising a processor and a communication interface coupled to the processor, wherein the processor is used to run a computer program or instruction to implement the multi-constraint underwater sound source localization method based on improved relative entropy distance as described in any one of claims 1 to 7.

Images