A method for detecting failed elements in a MIMO sonar array

The online detection of failed array elements of MIMO sonar arrays is solved by statistical hypothesis testing method, which affects detection performance by array element failure, and realizes efficient and robust array element detection, which is suitable for a variety of application scenarios.

CN118393472BActive Publication Date: 2024-11-26INST OF ACOUSTICS CHINESE ACAD OF SCI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410512650.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-26
Publication Date
2024-11-26
Estimated Expiration
2044-04-26

AI Technical Summary

Technical Problem

Because its sparse structure is easily affected by array element failure, the detection performance is degraded. The existing detection methods have problems such as large iterative solution volume, needing preliminary measurement, or being able to determine the affected output but not being able to determine the location of the failed array element. , especially in long-term unattended application scenarios, it is inconvenient to detect.

Method used

Using statistical hypothesis testing method, by creating phase recognition matrix and detection statistics, the failed array elements of the MIMO sonar array are detected online. The operation is simple and the detection efficiency is high. The location of the failed array elements can be determined and the signal-to-noise ratio or small snapshots are available. Good performance when counting.

Benefits of technology

It realizes efficient and robust online detection of failed array elements in MIMO sonar array, improves detection performance, and is suitable for long-term unattended application scenarios, especially in low signal-to-noise ratio and small snapshot count conditions, performing better than traditional methods .

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118393472B_ABST
    Figure CN118393472B_ABST
Patent Text Reader

Abstract

The present invention relates to a method for detecting failed array elements of a MIMO sonar array, the method comprising: determining a failed output in the output of a matched filter of an array to be detected; using the serial numbers of the array elements to be detected as row numbers and column numbers, creating a phase recognition matrix of the array to be detected; for any array element, the serial numbers of the rows and columns of the phase recognition matrix corresponding to its serial number, in the phase recognition matrix, are the relevant row numbers of the array element; for any array element, executing: querying whether there is a row number in the row where the failed output is located that is the same as the relevant row number and its corresponding state value; judging whether the array element is failed by the sum of the state values ​​corresponding to the relevant row numbers of the array element. The method of the present invention can detect failed array elements of a transceiver-combined MIMO sonar array online.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of underwater acoustic detection, and in particular to a method for detecting failed array elements of a MIMO sonar array. Background Art

[0002] Compared with the MIMO system based on uniform linear array, the multiple-input multiple-output (MIMO) array system based on sparse array can effectively improve the degree of freedom of the array system while using fewer array elements, thereby improving the spatial resolution and parameter estimation performance of the entire detection system. This type of system has important uses in today's rapidly developing Internet of Things and underwater monitoring networks. However, from the perspective of system stability, this type of array is more susceptible to array element failure due to its sparse structure than a uniform linear array system with the same aperture. Array element failure will directly lead to the failure of some traditional detection methods, thereby affecting the detection performance of the entire system. Therefore, it is necessary to study the detection problem of failed array elements in MIMO array systems.

[0003] From the perspective of array test calibration, researchers have conducted extensive research on array diagnosis based on the directional patterns of measured near-field data. First, a matrix method based on measured near-field data was proposed. The amplitude and phase coefficients of unit excitation were reconstructed by solving the linear system associated with the near-field measurement points. Since this method requires iterative solution, the amount of calculation is huge and it is difficult to apply in real time. Subsequently, a researcher proposed a method based on compressed sensing. This method first defines a deviation vector based on the excitation deviation between the reference matrix and the measured matrix, and then uses l1 norm minimization to restore the sparse deviation vector. Although it performs well in detection reliability and reducing the number of measurements, this method requires preliminary measurements of a fault-free reference array, which will bring some inconveniences in application. Moreover, for long-term unattended MIMO sonar detection platforms, for example, in application scenarios such as satellites or underwater monitoring, it is not convenient to remove the array from the operation site for near-field testing.

[0004] The traditional energy detection-based method can determine the matched filter output affected by the array element failure online by performing energy detection on each output of the matched filter. However, this method can only determine the output of the affected matched filter, but cannot determine the position of the failed array element. Moreover, this method has poor detection performance when the signal-to-noise ratio is low or the number of snapshots is small. Summary of the invention

[0005] In view of this, the main purpose of the present invention is to provide a method for detecting failed array elements of a MIMO sonar array, which can detect failed array elements of a transmitting and receiving MIMO sonar array online.

[0006] On the other hand, a MIMO sonar array failure output inspection method is provided, which adopts a statistical hypothesis testing method and can detect the failure output of the MIMO sonar array online. The method has simple operation, high detection efficiency and strong robustness.

[0007] To achieve the above objectives, on the one hand, the present application provides a method for detecting failed elements in a MIMO sonar array, comprising:

[0008] Determine the matched filter outputs of the array matched filters to be detected, which are affected by the failed array elements;

[0009] The phase recognition matrix of the array to be detected is created with the serial numbers of the array elements to be detected as the row numbers and column numbers; for any array element, the serial numbers of the rows and columns of the phase recognition matrix corresponding to its serial number in the phase recognition matrix are the row numbers of the outputs of the matched filter associated with the array element in the outputs of the matched filter of the array to be detected, hereinafter referred to as the relevant row numbers of the array element;

[0010] Execute on any array:

[0011] For any relevant row number of the array element, query whether there is a row number identical to the relevant row number in the row number of the matched filter output affected by the failed array element, and if not, record the corresponding state value as 0; if so, record the corresponding state value as 1;

[0012] The state values ​​corresponding to the relevant row numbers of the array element are summed up. If the sum is not less than the set threshold, the array element is considered to be invalid; otherwise, the array element is considered to be not invalid;

[0013] The threshold is greater than 0 and less than or equal to the total number of related row numbers of the array element.

[0014] In one possible implementation, determining the matched filter output affected by the failed array element includes:

[0015] The output of the array matched filter to be detected is taken as the sample, hereinafter referred to as array output;

[0016] The detection statistic is defined as follows:

[0017]

[0018] T n Divided into two subsets: the non-failed subset, Failure subset,

[0019] definition: Null hypothesis H0: ND = n-1; alternative hypothesis H1: ND> n-1; group detection is performed starting from n = 1. If T n<τ n , then accept the H0 hypothesis and let D=N-n+1; if T n >τ n , then accept the H1 hypothesis, set n = n + 1, and continue testing until the H0 hypothesis is accepted or n = N;

[0020] Find l separately n ,…,l N exist In the array output, the row with the same row number as the found serial number is the matched filter output affected by the failed array element, hereinafter referred to as the failed output;

[0021] Among them, the diagonal elements of the sample covariance matrix of the array output The (N-n+1)×(N-n+1)-dimensional trailing main sub-matrix of the array obtained after arranging in non-ascending order, l j are the diagonal elements of the trailing main submatrix; is the additive Gaussian noise variance; N is the number of array matched filters to be detected; D represents the number of rows of failed outputs in the array output; τ n Represents the detection threshold, and the formula is expressed as:

[0022]

[0023] in, Represents a random variable The inverse of the cumulative distribution function, P fa is the set false alarm probability.

[0024] In another possible implementation, the formula for determining the relevant row numbers of the array element is expressed as:

[0025]

[0026] in, is the set of relevant row numbers of the array element with sequence number m; is the set of serial numbers of the elements in the mth row of the phase identification matrix; is the set of serial numbers of the elements in the mth column of the phase recognition matrix; the serial number of each element in the phase recognition matrix is ​​represented by the serial number of the element in the phase recognition matrix according to the column number.

[0027] In another possible implementation, the formula for determining the state value corresponding to the relevant row number is expressed as:

[0028]

[0029] in, u nis the state value corresponding to the row number n of the mth array element.

[0030] In another possible implementation, the sum of the state values ​​corresponding to the relevant row numbers of the array element is used to determine whether it is invalid, and the formula is expressed as:

[0031] Not expired

[0032] Invalidation

[0033] Wherein, X is the threshold, X=ξ(2M-1), ξ(0<ξ≤1) is a balance coefficient, which is set according to the requirements of the array to be detected in the application scenario. The stricter the requirement, the larger the value of ξ; M is the number of elements of the array to be detected, and 2M-1 is the total number of related row numbers of the m-th element.

[0034] On the other hand, the present application provides a MIMO sonar array failure output inspection method, comprising:

[0035] The output of the array matched filter to be detected is taken as the sample, hereinafter referred to as array output;

[0036] The detection statistic is defined as follows:

[0037]

[0038] T n Divided into two subsets: the non-failed subset, Failure subset,

[0039] definition: Null hypothesis H0: ND = n-1; alternative hypothesis H1: ND> n-1; group detection is performed starting from n = 1. If T n <τ n , then accept the H0 hypothesis and let D=N-n+1; if T n >τ n , then accept the H1 hypothesis, set n = n + 1, and continue testing until the H0 hypothesis is accepted or n = N;

[0040] Find l separately n ,…,l N exist In the array output, the row with the same row number as the found serial number is the failure output of the array to be detected. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A schematic diagram of a flow chart of a method for detecting failed elements in a MIMO sonar array according to a first embodiment of the present invention;

[0042] Figure 2 Schematic diagram of a curve showing a change in the probability of detecting a failed array element as a function of the signal-to-noise ratio according to an embodiment of the present invention;

[0043] Figure 3 FIG. 4 is a curve showing a change in the detection probability of a failed array element versus the number of targets according to an embodiment of the present invention. DETAILED DESCRIPTION

[0044] Specifically, the process of a method for detecting failed array elements in a MIMO sonar array according to Embodiment 1 of the present invention is as follows: Figure 1 As shown, steps 101 to 104 are included.

[0045] Step 101: determining the matched filter outputs of the array matched filters to be detected, which are affected by the failed array elements;

[0046] Step 102: using the serial numbers of the array elements to be detected as the row numbers and column numbers, a phase recognition matrix of the array to be detected is created; for any array element, the serial numbers of the rows and columns of the phase recognition matrix corresponding to its serial number, in the phase recognition matrix, are the row numbers of the output of the matched filter associated with the array element in the output of the matched filter of the array to be detected, hereinafter referred to as the relevant row numbers of the array element, and the rows in the array output with the serial numbers as the row numbers are the relevant rows of the array element;

[0047] After step 102, steps 103 to 104 are performed on any array element:

[0048] Step 103: for any relevant row number of the array element, query whether there is a row number identical to the relevant row number in the row number of the matched filter output affected by the failed array element, if not, record its corresponding state value as 0; if so, record its corresponding state value as 1;

[0049] Step 104: summing the state values ​​corresponding to the relevant row numbers of the array element. If the sum is not less than a set threshold, the array element is considered to be invalid; otherwise, the array element is considered to be not invalid.

[0050] The threshold is greater than 0 and less than or equal to the total number of related row numbers of the array element.

[0051] In a possible implementation, in step 101, determining the matched filter output affected by the failed array element includes:

[0052] The output of the array matched filter to be detected is taken as the sample, hereinafter referred to as array output;

[0053] The detection statistic is defined as follows:

[0054]

[0055] T nDivided into two subsets: the non-failed subset, Failure subset,

[0056] definition: Null hypothesis H0: ND = n-1; alternative hypothesis H1: ND> n-1; group detection is performed starting from n = 1. If T n <τ n , then accept the H0 hypothesis and let D=N-n+1; if T n >τ n , then accept the H1 hypothesis, set n = n + 1, and continue testing until the H0 hypothesis is accepted or n = N;

[0057] Find l separately n ,…,l N exist In the array output, the row with the same row number as the found serial number is the matched filter output affected by the failed array element, hereinafter referred to as the failed output;

[0058] Among them, the diagonal elements of the sample covariance matrix of the array output The (N-n+1)×(N-n+1)-dimensional trailing main sub-matrix of the array obtained after arranging in non-ascending order, l j are the diagonal elements of the trailing main submatrix; is the additive Gaussian noise variance; N is the number of array matched filters to be detected; D represents the number of rows of failed outputs in the array output; τ n Represents the detection threshold, and the formula is expressed as:

[0059]

[0060] in, Represents a random variable The inverse of the cumulative distribution function, P fa is the set false alarm probability.

[0061] The estimated value calculation formula is expressed as:

[0062]

[0063] Among them, λ i is the sample covariance matrix of the matched filter output of the array to be detected Perform eigenvalue decomposition and arrange the eigenvalues ​​in non-descending order to obtain λ1≤λ2≤…≤λ N The first P values ​​in The maximum number of eigenvalues ​​in the noise subspace;

[0064] The P formula is expressed as:

[0065]

[0066] Wherein, L is the number of targets in the observation space, which is obtained by using the prediction feature threshold method and spatial smoothing technology; DOF is the degree of freedom of the array to be detected and the cooperative array or virtual array.

[0067] The following describes the implementation principle of determining the matched filter output affected by the failed array element in this implementation:

[0068] Without loss of generality, we assume that the failed array element can neither transmit signals to the target in the observation space nor receive the echo signal reflected by the target. In other words, the output of the failed array element does not contain any useful signal, and its output will only contain additive Gaussian noise. Therefore, the joint distribution of each element in the sample covariance matrix of the failed output is a central complex Wishart distribution, and then the detection statistic is defined based on the properties of the trace of the central complex Wishart distribution. The formula reasoning is expressed as:

[0069]

[0070] in, The diagonal elements of the sample covariance matrix output as an array The (N-n+1)×(N-n+1)-dimensional trailing main sub-matrix of the array obtained after arranging in non-ascending order.

[0071] When n-1 rows of array output are undamaged, the joint distribution of the elements in the sample covariance matrix of the remaining N-n+1 rows of damaged output is a central complex Wishart distribution, that is, we can deduce T n It follows a central chi-square distribution with 2T(N-n+1) degrees of freedom.

[0072] Based on this, the statistical hypothesis testing method is used to n The system is divided into two subsets: the non-failure subset and the failure subset, and group detection is performed. According to the Neyman-Pearson criterion, the false alarm probability P fa Derived threshold τ n , the export process formula is expressed as:

[0073]

[0074]

[0075] Among them, P fa The requirements of the array to be detected are set according to the application scenario. The stricter the requirements, the higher the P fa The smaller it is, generally one percent or one thousandth.

[0076] The following specifically describes the first embodiment of the present invention by taking a MIMO sonar array composed of M array elements and having both transmitting and receiving elements as an array to be detected as an example.

[0077] The array to be detected has its element positions determined by Given, the signal model when the array element is damaged is given below.

[0078] Let a(θ) be the M×1-dimensional steering vector related to the incoming wave direction θ. Assuming that a total of Q array elements are damaged, then their corresponding positions in a(θ) are replaced by 0, and a new steering vector is obtained as a d (θ). Assume that there are L observation spaces located at θ l ,l=1,2,…,L. For a MIMO array system with combined transmission and reception, each transmitting array element independently transmits an orthogonal waveform, and at the receiving end, the output of each array element is processed by M-way matched filters. The output of the matched filter of the array to be detected is vectorized and expressed as:

[0079] y(t)=B(θ)α(t)+n(t)

[0080] B(θ)=[b(θ1),b(θ2),…,b(θ L )]

[0081]

[0082] α(t)=[α1(t),α2(t),…,α L (t)] T

[0083] n(t)=[n1(t),n2(t),…,n N (t)] T

[0084] Where t is the slow time index, i.e., the number of pulses; n(t) represents the N×1-dimensional additive Gaussian noise that is unrelated to the signal. Assume that any element in n(t) satisfies the mean of zero and the variance of N=M 2 ; α l (t) represents the reflection coefficient of the target, and its variance is

[0085] At sampling time t, let y n (t) represents the nth output of the received signal of the array to be detected. The different pulse data collected within the coherent processing interval are regarded as the data of T snapshots received by the receiving end. n (t) The nth row of the output of the array matched filter to be detected corresponding to is expressed as:

[0086]

[0087] Then the N matched filter outputs in T snapshots are expressed as an N×T matrix:

[0088]

[0089] Step 101 determines the output of the N matched filters. Figure 2. The output of the matched filter affected by the failed array element.

[0090] In step 102, the formula for determining the relevant row numbers of the array element is expressed as:

[0091]

[0092] in, is the array element with serial number m, that is, the set of relevant row numbers of the mth array element, and m corresponds to the array element position set Medium m order; is the set of serial numbers of the elements in the mth row of the phase identification matrix; is the set of serial numbers of the elements in the mth column of the phase recognition matrix; the serial number of each element in the phase recognition matrix is ​​represented by the serial number of the element in the phase recognition matrix according to the column number.

[0093] The principle of determining the relevant row numbers of the array element is as follows:

[0094] The phase identification matrix is ​​denoted as Γ y , Γ y Any element of is defined as: the sum of the position of the array element whose serial number is the row number where the element is located and the position of the array element whose serial number is the column number where the element is located, divided by the minimum spacing between two adjacent array elements in the array to be detected, and the formula is expressed as:

[0095]

[0096] Among them, γ i,j Represents Γ y The element in the i-th row and j-th column; d represents the minimum distance between two adjacent array elements in the array to be detected, d i represents the position of the i-th array element, d j represents the position of the jth array element; M is the number of array elements to be detected.

[0097] From γ i,j It can be seen that Γ y The union of the mth row and mth column of gives the phase centers associated with the mth element in the array output, Γ y The elements of the mth row and mth column are in Γ yThe set of serial numbers in is the set of row numbers of these phase centers in the array output, that is, the set of row numbers of the matched filter output associated with the array element in the array output.

[0098] According to the above principle, in actual implementation, it is clear that Γ y The serial numbers of the array elements to be detected are used as row and column numbers to determine the serial numbers of each element, without having to complete the definition and calculation of its elements. For example, a MIMO sonar array with 4 elements based on the optimal second-order nested array for transmission and reception, that is, an optimal second-order nested array composed of M = 4 elements, is composed of 1 internal array and 1 external array cascaded, where the internal array contains The outer array contains L = MJ = 2 elements. The spacing between two adjacent elements in the inner array is That is, half wavelength; the distance between two adjacent elements in the external array is (J+1)d=3d. Then the array element position S={d1,d2,d3,d4}={0d,1d,2d,5d}. Therefore, its identification matrix is ​​expressed as:

[0099]

[0100] Assume that the second array element is damaged, Γ y The serial numbers of the items in bold are the relevant row numbers of the array element, expressed as:

[0101]

[0102] In step 103, the formula for determining the state value corresponding to the relevant row number is expressed as:

[0103]

[0104] in, u n is the state value corresponding to the row number n of the mth array element.

[0105] In step 104, the sum of the state values ​​corresponding to the relevant row numbers of the array element is used to determine whether it is invalid, and the formula is expressed as:

[0106] Not expired

[0107] Invalidation

[0108] Wherein, X is the threshold, X=ξ(2M-1), ξ(0<ξ≤1) is a balance coefficient, which is set according to the requirements of the array to be detected in the application scenario. The stricter the requirement, the larger the value of ξ; M is the number of elements of the array to be detected, and 2M-1 is the total number of related row numbers of the m-th element.

[0109] The process of a MIMO sonar array failure output inspection method according to the second embodiment of the present invention is as follows:

[0110] The output of the array matched filter to be detected is taken as the sample, hereinafter referred to as array output;

[0111] The detection statistic is defined as follows:

[0112]

[0113] T n Divided into two subsets: the non-failed subset, Failure subset,

[0114] definition: Null hypothesis H0: ND = n-1; alternative hypothesis H1: ND> n-1; group detection is performed starting from n = 1. If T n <τ n , then accept the H0 hypothesis and let D=N-n+1; if T n >τ n , then accept the H1 hypothesis, set n = n + 1, and continue testing until the H0 hypothesis is accepted or n = N;

[0115] Find l separately n ,…,l N exist In the array output, the row with the same row number as the found serial number is the failure output of the array to be detected;

[0116] Among them, the diagonal elements of the sample covariance matrix of the array output The (N-n+1)×(N-n+1)-dimensional trailing main sub-matrix of the array obtained after arranging in non-ascending order, l j are the diagonal elements of the trailing main submatrix; is the additive Gaussian noise variance; N is the number of array matched filters to be detected; D represents the number of rows of failed outputs in the array output; τ n Represents the detection threshold, and the formula is expressed as:

[0117]

[0118] in, Represents a random variable The inverse of the cumulative distribution function, P fa is the set false alarm probability.

[0119] The second embodiment of the present invention is not limited to being used for failure output inspection of a co-located MIMO sonar array.

[0120] Experimental conclusion:

[0121] The effect of the present invention is verified by computer simulation. A nested MIMO sonar array with M=6 array elements is used as the array to be detected, in which the second and fifth array elements fail. fa =0.001. For the sake of comparison, the failed array element is further detected based on the classic energy detection method to detect the failed output and the method of the second embodiment of the present invention to check the failed output.

[0122] Assume that there is one target in the observation space from -60° to 60°. Figure 2 The curves of the failure output detection probability varying with the signal-to-noise ratio are given when the number of snapshots is T=5000 and T=50, using two failure output detection methods. Figure 3 The curves of the detection probability of failed array elements changing with the number of targets when the signal-to-noise ratio is -20dB and the number of snapshots is T=5000, using two failed output detection methods.

[0123] Depend on Figure 2 , Figure 3 It can be seen that the method of the present invention can detect the failed array element of the combined transceiver MIMO sonar array online. However, when the signal-to-noise ratio is low, the number of snapshots or the number of targets is small, the method of the second embodiment of the present invention is used to check the failed output, and the detection effect is obviously better than the classic energy detection method. However, with the improvement of the signal-to-noise ratio, the increase of the number of snapshots and the increase of the number of targets, this advantage becomes less obvious.

[0124] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention.

Claims

1. A method for detecting failed array elements in a MIMO sonar array, characterized in that: include: Determine the matched filter outputs of the array matched filters to be detected, which are affected by the failed array elements; The phase recognition matrix of the array to be detected is created with the serial numbers of the array elements to be detected as the row numbers and column numbers; for any array element, the serial numbers of the rows and columns of the phase recognition matrix corresponding to its serial number in the phase recognition matrix are the row numbers of the outputs of the matched filter associated with the array element in the outputs of the matched filter of the array to be detected, hereinafter referred to as the relevant row numbers of the array element; Execute on any array: For any relevant row number of the array element, query whether there is a row number identical to the relevant row number in the row number of the matched filter output affected by the failed array element, and if not, record the corresponding state value as 0; if so, record the corresponding state value as 1; The state values ​​corresponding to the relevant row numbers of the array element are summed up. If the sum is not less than the set threshold, the array element is considered to be invalid; otherwise, the array element is considered to be not invalid; The threshold is greater than 0 and less than or equal to the total number of related row numbers of the array element.

2. The method according to claim 1, characterized in that The determining of the matched filter output affected by the failed array element comprises: The output of the array matched filter to be detected is taken as the sample, hereinafter referred to as array output; The detection statistic is defined as follows: T n Divided into two subsets: the non-failed subset, Failure subset, definition: Null hypothesis H0: ND = n-1; alternative hypothesis H1: ND> n-1; group detection is performed starting from n = 1. If T n <τ n , then accept the H0 hypothesis and let D=N-n+1; if T n >τ n , then accept the H1 hypothesis, set n = n + 1, and continue testing until the H0 hypothesis is accepted or n = N; Find l separately n ,…,l N exist In the array output, the row with the same row number as the found serial number is the matched filter output affected by the failed array element, hereinafter referred to as the failed output; Among them, the diagonal elements of the sample covariance matrix of the array output The (N-n+1)×(N-n+1)-dimensional trailing main sub-matrix of the array obtained after arranging in non-ascending order, l j are the diagonal elements of the trailing main submatrix; is the additive Gaussian noise variance; N is the number of array matched filters to be detected; D represents the number of rows of failed outputs in the array output; τ n Represents the detection threshold, and the formula is expressed as: in, Represents a random variable The inverse of the cumulative distribution function, P fa is the set false alarm probability.

3. The method according to claim 1, characterized in that The formula for determining the relevant row numbers of the array element is expressed as: in, is the set of relevant row numbers of the array element with sequence number m; is the set of serial numbers of the elements in the mth row of the phase identification matrix; is the set of serial numbers of the elements in the mth column of the phase recognition matrix; the serial number of each element in the phase recognition matrix is ​​represented by the serial number of the element in the phase recognition matrix according to the column number.

4. The method according to claim 3, characterized in that The formula for determining the state value corresponding to the relevant row number is expressed as: in, u n is the state value corresponding to the row number n of the mth array element.

5. The method according to claim 4, characterized in that The sum of the state values ​​corresponding to the relevant row numbers of the array element is used to determine whether it is invalid. The formula is expressed as: Wherein, X is the threshold, X=ξ(2M-1), ξ(0<ξ≤1) is a balance coefficient, which is set according to the requirements of the array to be detected in the application scenario. The stricter the requirement, the larger the value of ξ; M is the number of elements of the array to be detected, and 2M-1 is the total number of related row numbers of the m-th element.

6. A MIMO sonar array failure output inspection method, characterized in that: include: The output of the array matched filter to be detected is taken as the sample, hereinafter referred to as array output; The detection statistic is defined as follows: T n Divided into two subsets: the non-failed subset, Failure subset, definition: Null hypothesis H0: ND = n-1; alternative hypothesis H1: ND> n-1; group detection is performed starting from n = 1. If T n <τ n , then accept the H0 hypothesis and let D=N-n+1; if T n >τ n , then accept the H1 hypothesis, set n = n + 1, and continue testing until the H0 hypothesis is accepted or n = N; Find l separately n ,…,l N exist In the array output, the row with the same row number as the found serial number is the failure output of the array to be detected; Among them, the diagonal elements of the sample covariance matrix of the array output The (N-n+1)×(N-n+1)-dimensional trailing main sub-matrix of the array obtained after arranging in non-ascending order, l j are the diagonal elements of the trailing main submatrix; is the additive Gaussian noise variance; N is the number of array matched filters to be detected; D represents the number of rows of failed outputs in the array output; τ n Represents the detection threshold, and the formula is expressed as: in, Represents a random variable The inverse of the cumulative distribution function, P fa is the set false alarm probability.