A comprehensive multi-snapshot joint Newton orthogonal matching pursuit positioning method

Abstract

The application relates to a comprehensive multi-snapshot joint Newton orthogonal matching pursuit positioning method, and belongs to the field of acoustics. The method estimates all possible grid sound source coordinates through an improved comprehensive orthogonal matching pursuit method, reconstructs a target sound source region by locally refining a grid centering on the coordinates, and constructs a new perception matrix; a plurality of sub-arrays with equal numbers of array elements are randomly selected from an original array, the grid coordinates of the sound source are preliminarily estimated in the reconstructed region based on the multi-sub-array data, and the estimated sound source coordinates and source intensity are compensated and optimized through a multi-snapshot joint Newton method. The application has the advantages that: the process of atom selection is improved, the reconstructed region is maximally ensured to contain all sound sources, the coherence problem caused by a finer grid spacing is avoided, and the grid positioning capability is enhanced; the multi-sub-array joint estimation method can more fully utilize the measurement data, and the sound source positioning precision is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of acoustics and relates to a comprehensive multi-shot joint Newton orthogonal matching pursuit positioning method. BACKGROUND

[0002] To convert the sound source estimation problem into a sparse reconstruction solving problem of compressed sensing, a traditional compressed beamforming method constructs a sparse reconstruction model by griding a target sound source region, which assumes that the target sound source falls on a divided grid point, however, the fact is that no matter how fine the grid division of the target region space is, the sound source may not be on the grid point, and there is always a mismatch problem between the assumed basis and the real basis. For a greedy algorithm such as an orthogonal matching pursuit method, the existence of the basis mismatch problem may even cause positioning errors, although a finer grid can alleviate the basis mismatch problem, but it comes with an increase in the demand for computing resources and an increase in the correlation between dictionary atoms, which causes the estimation difficulty to rise, and the presence of noise also increases the difficulty of correctly selecting atoms.

[0003] In addition, the current common method for improving positioning performance is to increase the observation dimension of the sound source signal, which undoubtedly increases the measurement difficulty, however, in existing research, the measurement data are mostly used only once, and the effective information in the data is obviously not fully utilized. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a comprehensive multi-shot joint Newton orthogonal matching pursuit positioning method, which more fully utilizes the effective information in the existing measurement data and improves the positioning performance of the sound source.

[0005] To achieve the above purpose, the present application provides the following technical scheme:

[0006] A comprehensive multi-shot joint Newton orthogonal matching pursuit positioning method, which comprises the following steps:

[0007] S1: According to the measurement data of the original M-element array, the sound source coordinate candidate set is obtained by an improved comprehensive orthogonal matching pursuit method, and the target sound source region is locally refined by taking it as the center to reconstruct the target sound source region, and a new perception matrix A is constructed sub ;

[0008] S2: According to the total array sparsity principle, a plurality of sub-arrays composed of M sub element are randomly selected from the original array, the measurement data corresponding to each sub-array is constructed, and a maximum likelihood function is constructed;

[0009] S3: Based on the reconstructed target grid region, a rough in-grid estimation result of the sound source is obtained;

[0010] S4: The estimated results are corrected by Newton method to improve the positioning accuracy; it is divided into two steps, first, the single-step correction of the current iteration obtained in the network estimation results, and then the multi-step correction of all estimated results based on the current latest residual;

[0011] S5: Set up a global feedback link to provide an opportunity to make up for the missed sound source in the sound source area reconstruction process of S1.

[0012] Further, the S1 comprises the following steps:

[0013] S11, the plane where the sound source is located is divided into Ng grid points with a certain spatial interval d, and a(r ng ) is normalized to obtain a n (r ng ), a(r ng ) is the directional vector (atom) between the spatial coordinates r ng of the ngth grid and the spatial coordinates r mic.k of the array element composed of M array elements, k = 1, 2,..., M;

[0014] a n (r ng ) = a(r ng ) / ||a(r ng )||2, ng = 1, 2,..., Ng

[0015] a(r ng ) = [g(r ng ,r mic.k ), g(r ng ,r mic.2 ),..., g(r ng ,r mic.M )] T

[0016]

[0017] Wherein, g(r ng ,r mic.m ) is the Green function, ||·||2 represents the 2 norm operation, [·] T represents the transpose, f is the sound source frequency, and c is the sound speed;

[0018] S12, in each iteration process, the current residual Y res= [y(1), y(2),..., y(L)], L is the number of snapshots, and is the inner product of each column of the dictionary and the normalized atom, the first three maximum inner product values are selected, and the atom coordinates with the maximum inner product value are taken as the center of a circular region that can just contain its adjacent atoms. The sound source coordinate candidate set is determined by judging whether the other two atoms fall within the circle, i.e., the atom coordinates and the maximum inner product value atom coordinates that fall outside the region together constitute the coordinate candidate set R su ;

[0019]

[0020] R su = R su {r k.1 ,r k.i ⊙(r k.1 )},r k = {r k.1 ,r k.2 ,r k.3}

[0021] wherein ⊙(r k.1 ) represents a circular space covering a certain range with the coordinate r k.1 as the center. Like the orthogonal matching pursuit method, in each iteration process, the coordinate corresponding to the atom with the maximum inner product value is taken as the estimation result to promote the iteration process;

[0022] S13, according to the atom coordinates in the atom candidate set, a square grid space with a radius of 1.5d is radiated with half of the original space grid spacing d as the new grid spacing, and these local grid spaces together constitute our new target sound source region R Ns , and a new perception matrix A sub is constructed, which is composed of the steering vectors corresponding to all grid points in the target region.

[0023] Further, the S2 comprises the following steps:

[0024] S21, according to the array overall sparsity principle, L A groups of subarrays composed of M sub array elements are randomly selected from the original M array elements, and the corresponding array element data together constitute the new measurement signal wherein represents the measurement data corresponding to the l A th subarray;

[0025] S22, for each subarray, based on the maximum likelihood solution of the sound source coordinates and intensity is obtained by minimizing the residual energy , and a maximum likelihood function is constructed

[0026]

[0027] wherein is a perception matrix corresponding to the real sound source coordinate, corresponds to the source strength, is noise, ||·|| is the Frobenius norm, F represents the Frobenius norm, represents taking real numbers, (·) is the conjugate transpose. H represents the conjugate transpose.

[0028] Further, the S3 comprises the following steps:

[0029] S31, according to the least square method, a maximum likelihood ratio test cost function GLRT(·) is established by estimating one sound source estimate per iteration process;

[0030]

[0031] S32, for each subarray, the coordinates corresponding to the first three maximum values of GLRT(·) per iteration are taken as the current possible estimation value, a circular area with a radius just capable of containing its adjacent grid points is radiated with the estimation value as the center, the number of estimation values falling in the area is counted, and the center grid point of the area corresponding to the largest number is taken as the final in-grid estimation coordinate

[0032]

[0033] count(g) represents the size of the count satisfying the condition;

[0034] S33, the in-grid estimated sound source strength is calculated

[0035]

[0036] Further, the S4 comprises the following steps:

[0037] S41, the current in-grid estimation result is single-step corrected by Newton method;

[0038]

[0039] wherein, is the Hessian matrix, is the Jacobian matrix;

[0040] S42, the residual Y is updated res , the atom support set A sup , the coordinate support set R supand source strength support set Q sup ;

[0041]

[0042] S43. Based on the latest residuals, perform multi-step corrections on all estimated results in sequence; first, update the source strength using the least squares method to complete the orthogonalization operation, and then correct the results using Newton's method;

[0043]

[0044] Furthermore, S5 includes the following steps:

[0045] S51. Set up a global feedback loop to provide an opportunity to compensate for missed sound sources during the S1 sound source region reconstruction process; update using the least squares method. Y estimated and Y res Calculate Tol1 and Tol2, and determine if the iteration stopping condition is met. The iteration terminates when (Tol1≤ε1) or (Tol2≤ε2); where Y res0 The measurement residual before each iteration;

[0046]

[0047] Tol1=Σ(|Y res0 | 2 -|Y res | 2 ) / ∑(|Y res0 | 2 )

[0048] Tol2=∑(|Y res0 | 2 -|Y res | 2 ) / ∑(|Y Estimated | 2 )

[0049] After the iteration stops, the output coordinate support set R is... sup Heyuan Strong Support Set Q sup This yields the estimated coordinates of the sound source and its corresponding intensity.

[0050] The beneficial effects of this invention are as follows:

[0051] Since the target area of the reconstruction range of the application is greatly reduced, all sound sources are guaranteed to fall in the reconstruction area according to the correlation as much as possible, the correlation problem caused by the finer grid is avoided, and the grid positioning ability is improved; meanwhile, the application combines the estimation results of multiple sub-arrays, so that the in-grid coordinates estimated in each iteration process are closest to the real sound source coordinates, and the joint use of multiple groups of sub-array data in the off-grid correction process of the Newton method makes the correction direction of the Newton method more stable and reliable, so that the accurate positioning of the sound source is more reliably realized.

[0052] The reconstruction sound source target area method proposed by the method improves the sound source estimation ability and can avoid estimating false results to a certain extent; the multi-sub-array joint estimation model introduced by the method can more fully utilize the useful information in each measurement data, improve the noise robustness, improve the positioning accuracy and reduce the positioning deviation.

[0053] Other advantages, objects and features of the application will be set forth in part in the following specification, and in part will become apparent to those skilled in the art from a consideration of the following specification, or can be learned from practice of the application. The objects and other advantages of the application can be realized and attained by the below description. BRIEF DESCRIPTION OF DRAWINGS

[0054] In order to make the objects, technical solutions and advantages of the application clearer, the preferred detailed description of the application will be combined with the drawings to describe the application, in which:

[0055] Figure 1 The figure is a flowchart of the method of the application;

[0056] Figure 2 The figure is a schematic diagram of the principle, Figure 2 (a) is a schematic diagram of the principle of the improved orthogonal matching pursuit method, Figure 2 (b) is a schematic diagram of the reconstruction sound source area;

[0057] Figure 3 The figure is a simulation result diagram of 3 sound sources under 20kHz; Figure 3 (a) is the positioning result of the MNOMP method under the spaced sound sources; Figure 3 (b) is the positioning result of the COMP-MJNOMP method under the spaced sound sources; Figure 3 (c) is the positioning result of the MNOMP method under the adjacent sound sources; Figure 3 (d) is the positioning result of the COMP-MJNOMP method under the adjacent sound sources;

[0058] Figure 4 The figure is a Monte Carlo random test result; Figure 4(a) Correct Recognition Rate (CRR) for sound source, which measures the probability of single correct positioning; Figure 4 (b) Root Mean Square Error (RMSE) for positioning coordinates, which measures the deviation size of positioning coordinates; Figure 4 (c) Cumulative Correct Recognition Rate (CCRR) for sound source, which measures the probability of single correct estimation; Figure 4 (d) Mean Amplitude Error (MAE) for source strength, which measures the deviation size of estimated source strength;

[0059] Figure 5 Fig. 1 is a diagram of positioning results of the test; Figure 5 (a) is the positioning result of the multiple-shot orthogonal matching pursuit method; Figure 5 (b) is the positioning result of the multiple-shot Newton orthogonal matching pursuit method; Figure 5 (c) is the positioning result of the method of the present application. DETAILED DESCRIPTION

[0060] The present application will be described in greater detail by way of specific embodiments, and as such, those skilled in the art can easily understand other advantages and purposes of the present application from the contents disclosed in the specification. The present application can also be implemented or applied in other different embodiments, and the details in the specification can be modified or changed based on different views and applications without departing from the spirit of the present application. It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present application in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0061] In the drawings, only the diagrams are used for illustrative purposes, and the representations are only schematic diagrams, not physical diagrams, and should not be understood as limitations of the present application; in order to better illustrate the embodiments of the present application, some components in the drawings may be omitted, enlarged or reduced, and do not represent the actual size of the product; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0062] The same or similar reference numerals in the drawings of the embodiments of the present application correspond to the same or similar components; in the description of the present application, it should be understood that the orientations or positional relationships indicated by terms such as "upper", "lower", "left", "right", "front", "back" and the like are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, therefore the terms describing the positional relationship in the drawings are only used for illustrative purposes, and should not be understood as limitations of the present application, and for those skilled in the art, the specific meanings of the above terms can be understood according to the specific circumstances.

[0063] Please refer to Figure 1 , Figure 2It is a schematic diagram of principle, Figure 2 (a) is a schematic diagram of principle of improved orthogonal matching pursuit method, Figure 2 (b) is a schematic diagram of reconstructed sound source region. The application is a comprehensive multi-shot joint Newton orthogonal matching pursuit positioning method, comprising the following steps:

[0064] S1: according to the measurement data of the original M-element array, the sound source coordinate candidate set is obtained by the improved comprehensive orthogonal matching pursuit method, and the target sound source region is reconstructed by local refinement grid centering on the sound source coordinate candidate set, and a new perception matrix A is constructed sub :

[0065] S11, the atoms in the original sound source plane are normalized to obtain a n (r ng );

[0066] a n (r ng )=a(r ng ) / ||a(r ng )||2,ng=1,2,...,Ng

[0067] S12, in each iteration process, the inner product of the current residual Y res and each column of the normalized atoms in the dictionary is calculated, the three atoms corresponding to the maximum inner product values are selected, the atom coordinate with the maximum inner product value is taken as the center of a circular region which can just contain its adjacent atoms, and the sound source coordinate candidate set is determined by judging whether the other two atoms fall within the circle, that is, the atom coordinates and the maximum inner product value atom coordinates which fall outside the region jointly constitute the coordinate candidate set R su .

[0068]

[0069] R su =R su {r k.1 ,r k.i ⊙(r k.1 )},r k ={r k.1 ,r k.2 ,r k.3}

[0070] In the formula, ⊙(r k.1 ) represents a circular space covering a certain range with r k.1 as the center, and like the orthogonal matching pursuit method, in each iteration process, the coordinate corresponding to the atom with the maximum inner product value is taken as the estimation result, and the iteration process is promoted.

[0071] S13, according to the atomic coordinates in the atomic candidate set, a square grid space with a radius of 1.5d is radiated with half of the original space grid interval d as the new grid interval, and these local grid spaces together constitute our new target sound source region R Ns .

[0072] S2: randomly select multiple groups of sub-arrays composed of M sub array elements from the original array according to the array overall sparse principle, construct the measurement data corresponding to each sub-array, and construct the maximum likelihood function:

[0073] S21, randomly select L A groups of sub-arrays composed of M sub array elements from the original M array element array according to the array overall sparse principle, and the corresponding array element data jointly constitute the new measurement signal wherein

[0074] S22, for each sub-array, based on obtain the maximum likelihood solution of the sound source coordinates and intensity by minimizing the residual energy construct the maximum likelihood function wherein is the perception matrix corresponding to the true sound source coordinates, corresponds to the source intensity, is noise, and ||·|| represents the Frobenius norm. F

[0075]

[0076] S3: based on the reconstructed target grid region, obtain the rough in-grid estimation result of the sound source:

[0077] S31, by estimating one sound source estimate in each iteration process, establish the maximum likelihood ratio test cost function GLRT(·) according to the least squares method;

[0078]

[0079] S32, for each sub-array, take the coordinates corresponding to the first three maximum values of GLRT(·) in each iteration as the current possible estimation value, and preferentially radiate a circular region with a radius that can just contain its adjacent grid points with the estimation value with the most number of occurrences in all possible estimation values as the center, count the number of estimation values falling in the region, and take the center grid point of the region with the most number of regions as the final in-grid estimation coordinate

[0080]

[0081] S33, calculate the in-grid estimation sound source intensity​

[0082]

[0083] S4: Correct the estimated results by Newton method to improve the positioning accuracy. It is divided into two steps, first, one-step correction is made to the in-network estimation results obtained in the current iteration, and then multi-step correction is made to all the estimated results based on the current latest residual:

[0084] S41, one-step correction is made to the current in-network estimation results by Newton method;

[0085]

[0086]

[0087] In the formula, is the Hessian matrix, is the Jacobian matrix.

[0088] S42, update the residual Y res , the atomic support set A sup , the coordinate support set R sup and the source intensity support set Q sup

[0089]

[0090] S43, multi-step correction is made to all the estimated results in turn based on the current latest residual; first, the source intensity is updated by least square method to complete the orthogonalization operation, and then the results are corrected by Newton method;

[0091]

[0092] S5: Set a global feedback link to provide a remedy opportunity when a missed sound source occurs in the sound source area reconstruction process of S1:

[0093] S51, set a global feedback link to provide a remedy opportunity when a missed sound source occurs in the sound source area reconstruction process of S1. Update Y estimated and Y res , calculate Tol1 and Tol2, and judge whether the iteration stopping condition is met. When (Tol1≤ε1) or (Tol2≤ε2), the iteration is terminated.

[0094]

[0095] Tol1 = ∑(|Y res0 | 2 -Y res |​2 ) / ∑(|Y res0 | 2 )

[0096] Tol2=∑(|Y res0 | 2 -|Y res | 2 ) / ∑(|Y Estimated | 2 )

[0097] After the iteration is stopped, the coordinate support set R sup and the source strength support set Q sup are output, that is, the estimation results of the sound source coordinates and the corresponding sound source strengths are obtained.

[0098] In this example, the beneficial effects of the present application are verified through simulation test verification and field test verification:

[0099] Simulation test: a microphone array with a diameter of 0.32m and M=128 array elements is used for simulation test, the target sound source is located in a target area with an area of 16m 2 in front of the array at a distance of 2m, the target area is divided by taking a pitch d=0.1m, and a total of 1681 grid points are obtained. Assuming that there are 3 sound sources with a sound pressure level of 100dB at a frequency of f=20kHz in the target area, sampling is performed at a sampling rate of 192k. In order to fully reflect the positioning performance of the three algorithms, we set two types of positioning situations of distant sound sources and adjacent sound sources according to the distance of the sound sources, and without loss of generality, we assume that the coordinates of the distant sound sources are (-1.1, -1.3, 2), (0.16, 0.93, 2), (1.07, -0.73, 2), and the coordinates of the adjacent sound sources are (-1.1, -1.3, 2), (0.16, 0.93, 2), (0.16, 0.83, 2). The positioning test is performed in a simulated environment with a signal-to-noise ratio SNR=0dB and a signal length of 100 shots.

[0100] The above sound sources are identified by using the present method and the Newton orthogonal matching pursuit algorithm respectively, and the results are shown in Figure 3 , in which "*" and "O" respectively represent the estimated sound source position and the real sound source position. As can be seen from Figure 3 (a) and Figure 3 (b), for distant sound sources, the present method and the Newton orthogonal matching pursuit algorithm can correctly locate the real sound sources. As can be seen from Figure 3 (c) and Figure 3 (d), the off-grid compensation effect of the Newton orthogonal matching pursuit algorithm is affected when the sound sources are adjacent, while the positioning effect of the present method is still good. Therefore, the present method has more stable and accurate positioning effect.

[0101] Monte Carlo random simulation test: simulate the random test of different algorithms under various conditions. For each algorithm, the SNR value is taken at 5dB intervals in the range of -10dB~30dB, and the positioning test is carried out under the condition of the number of snapshots L=100. For each case, 500 random tests are carried out. In order to evaluate the performance of the algorithm, four evaluation parameters are set, including correct recognition rate CRR, root mean square error RMSE, cumulative correct recognition rate CCRR and mean amplitude error MAE.

[0102] From Figure 4 (a) it can be seen that the CRR of the method invented is close to 95% at 0dB, compared with the existing orthogonal matching pursuit algorithm and Newton orthogonal matching pursuit algorithm, the correct recognition rate of the method is significantly increased, and the noise robustness is significantly better. In addition, Figure 4 (c) the size of CCRR also shows the superiority of the method invented. Figure 4 (b) the positioning deviation of the orthogonal matching pursuit algorithm is about 4.3cm under high SNR, the positioning deviation of the Newton orthogonal matching pursuit algorithm is about 3.3cm, and the positioning deviation of the method invented is about 2cm, which shows that the algorithm has good positioning accuracy. As Figure 4 (d), the source intensity estimation result of the algorithm is closer to the true source intensity in the whole SNR space.

[0103] Field test verification: a 128-element array with a sampling rate of 192k and a diameter of about 0.16m arranged in a spiral structure is used for positioning test. Three speakers with uniform intensity and random position are placed 1.5m away from the array panel, and the speakers are emitted by Philips S1009. The positioning of the three algorithms under 20k frequency is compared, and 100 sampling points, i.e. 100 snapshots, are used for single positioning, and the positioning results are shown in Figure 5 , Figure 5 (a) is the positioning result by the multi-snapshot orthogonal matching pursuit method; Figure 5 (b) is the positioning result by the multi-snapshot Newton orthogonal matching pursuit method; Figure 5 (c) is the positioning result by the method invented. From the results, it can be seen that in this experimental scene, the orthogonal matching pursuit algorithm has obvious grid mismatch problem, the Newton orthogonal matching pursuit algorithm has limited correction effect on the grid positioning deviation, and compared with the method invented, the method invented can basically locate the center of the real sound source.

[0104] From the simulation test and field test results, it is proved that the sound source positioning ability of the method invented is better than that of the prior art.

[0105] Finally, it is to be explained that the above embodiments are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the purpose and scope of the present application, and all should be covered in the scope of the claims of the present application.

Claims

1. A method for integrated multi-shot joint Newton orthogonal matching pursuit positioning, characterized in that: The method comprises the following steps: S1: According to the measurement data of the original M-element array, the sound source coordinate candidate set is obtained by the improved comprehensive orthogonal matching pursuit method, and the target sound source region is reconstructed by locally refining the grid centering on the candidate set, and a new perception matrix A is constructed sub ; S2: randomly select multiple groups of sub-arrays composed of M sub array elements from the original array according to the array overall sparsity principle, construct the measurement data corresponding to each sub-array, and build a maximum likelihood function; S3: obtaining a rough grid estimation result of the sound source based on the reconstructed target grid area; S4: correcting the estimated result by using the Newton method to improve the positioning accuracy; the correction is divided into two steps, i.e., firstly, single-step correction is performed on the grid estimation result obtained in the current iteration, and then multi-step correction is performed on all the estimated results based on the latest residual; S5: setting a global feedback link to provide a compensation opportunity for the S1 reconstruction of the sound source area when a missed sound source occurs.

2. The integrated multi-snapshot joint Newton orthogonal matching pursuit positioning method according to claim 1, characterized in that: The S1 comprises the following steps: S11. Divide the plane containing the sound source into Ng grid points according to a certain spatial spacing d. For each grid point, a(r) ng Normalization yields a n (r ng ), a(r ng ) is the spatial coordinate r of the ng-th grid. ng To the array element spatial coordinates r of the array consisting of M elements mic.k The guiding vectors between k = 1, 2, ..., M, i.e., atoms; a n (r ng )=a(r ng ) / ||a(r ng )||2,ng=1,2,...,Ng a(r ng ) = [g(r ng ,r mic.1 ), g(r ng ,r mic.2 ),..., g(r ng ,r mic.M )] T where g(r ng ,r mic.k ) is the Green function, ||·||2represents the 2-norm operation, [·] T represents the transpose, f is the sound source frequency, and c is the sound speed; S12, in each iteration process, the current residual Y is calculated by res = [y(1), y(2),..., y(L)], L is the number of snapshots, and the inner product of each column of the normalized atoms in the dictionary is calculated. The first three maximum inner product values are selected, and the atom coordinates with the maximum inner product value are taken as the center of a circular region that can just contain its adjacent atoms. By judging whether the other two atoms fall within the circle, the candidate set of sound source coordinates is determined, that is, the atom coordinates falling outside the region and the atom coordinates with the maximum inner product value together constitute the candidate set R of coordinates su ; R su = R su {r k.1 , r k.i e(r k.1 ), r k = {r k.1 , r k.2 , r k.3} where e(r k.1 ) represents a circular space centered at coordinate r k.1 , with a certain range. Like the OMP method, the estimated result in each iteration is the coordinate corresponding to the atom with the largest inner product value, which drives the iteration process. S13, according to the atomic coordinates in the atomic candidate set, take half of the original space grid interval d as the new grid interval, radiate a square grid space with a radius of 1.5d, and these local grid spaces together constitute our new target sound source area R Ns And construct a new perception matrix A sub Which is composed of the steering vectors corresponding to all the grid points in the target area.

3. The method of claim 1, wherein: The S2 comprises the following steps: S21, randomly selecting L A The group is composed of M sub subarray, the corresponding array data jointly constitute a new measurement signal Wherein represent the measurement data corresponding to the l A th subarray; S22, for each subarray, based on By minimizing the residual energy Obtain the maximum likelihood solution of the sound source coordinates and intensity, construct the maximum likelihood function in It is the perception matrix corresponding to the actual sound source coordinates. Corresponding source strength, It's noise. F Represents the Frobenius norm. Represents taking real numbers, (·) H This represents the conjugate transpose.

4. The integrated multi-snapshot joint Newton orthogonal matching pursuit positioning method according to claim 3, characterized in that: The S3 comprises the following steps: S31: establishing a maximum likelihood ratio test cost function GLRT(·) according to the least square method in the manner of estimating one sound source estimation result through each iteration process; S32. For each subarray, in each iteration, take the coordinates corresponding to the top 3 maximum values ​​of GLRT(·) as the current possible estimate. Take the estimate that appears most frequently among all possible estimates as the center and radiate a circular region with a radius that just includes its neighboring grid points. Count the number of estimates falling within this region, and take the center grid point of the region with the most estimates as the final on-grid estimated coordinates. count(·) represents the size of the count satisfying the condition; S33, calculate the in-network estimated sound source source strength 5. The integrated multi-snapshot joint Newton orthogonal matching pursuit positioning method according to claim 4, characterized in that: The S4 comprises the following steps: S41: performing single-step correction on the current grid estimation result by using the Newton method; wherein is the Hessian matrix, is the Jacobian matrix; S42, update the residual Y res , the atomic support set A sup , the coordinate support set R sup , and the source strength support set Q sup ; S43: performing multi-step correction on all the estimated results based on the latest residual; firstly, orthogonalization operation is completed by updating the source strength through the least square method, and then the result is corrected by using the Newton method; 6. The integrated multi-snapshot joint Newton orthogonal matching pursuit positioning method according to claim 5, characterized in that: The S5 comprises the following steps: S51, setting a global feedback link, providing an opportunity to make up for the missed sound source in the sound source area reconstruction process of S1; updating by least square method Y estimated and Y res , calculate Tol1 and Tol2, judge whether the iteration stopping condition is satisfied, when Tol1≤ε1 or Tol2≤ε2, the iteration is terminated; wherein, Y res0 indicates the measurement residual before each iteration; After the iteration stops, output the coordinate support set R sup and the source strength support set Q sup That is, the estimation result of the sound source coordinate and its corresponding sound source intensity is obtained.

Images