Underwater sound target orientation estimation method and device based on line spectrum focusing and storage medium

Abstract

The invention discloses an underwater acoustic target orientation estimation method and device based on line spectrum focusing and a storage medium, and relates to the technical field of signal processing. The method comprises the following steps: receiving a broadband radiation noise signal of a target through a sensor array, selecting a reference array element to receive the signal, and performing short-time Fourier transform on the signal to extract a line spectrum signal frequency generated by target movement; an array receiving signal is converted to a frequency domain, and a sub-band covariance matrix with each line spectrum as the center is calculated; constructing a focusing transformation matrix based on the estimated incident angle and the reference frequency, mapping the signal subspace under each line spectrum frequency to the signal subspace of the reference frequency, and calculating the geometric average of a covariance matrix after focusing; and finally, performing spectrum peak search by using a MUSIC algorithm to obtain a target orientation. Therefore, not only is the accuracy of azimuth estimation significantly improved and the phenomenon of multi-target confusion in the BTR map avoided, but also the calculation complexity is greatly reduced and the operation time is saved by executing the focusing transformation only in the narrow band adjacent to the line spectrum.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of signal processing, in particular to a method and device for estimating the azimuth of an underwater acoustic target based on line spectrum focusing, an electronic device and a storage medium. BACKGROUND

[0002] In recent years, ocean unmanned platforms have developed rapidly. With the development of unmanned technology, ocean unmanned platforms have become a field that navies of various countries are competing to develop and develop. Compared with other underwater platforms, unmanned platforms have the advantages of small size, high flexibility and weak radiated noise, and can be widely used in target reconnaissance, communication relay, environmental monitoring and other fields. Entering sensitive sea areas to collect key information, the platform itself uses low-noise propulsion and optimized shape design, which brings challenges to passive detection. Therefore, in the face of the potential challenges brought by the rapid development of unmanned platforms, it is necessary to carry out research on the target azimuth estimation method for the acoustic characteristics of the platform itself and optimize the passive detection strategy, which has become an inevitable trend to counter the activities of unmanned platforms. Not only can it improve the passive detection capability of ocean unmanned platforms, but also is conducive to protecting the safety of key regional ports and maintaining maritime rights and interests.

[0003] The radiated noise of underwater targets is the primary basis for passive sonar detection. The radiated noise of both surface ships and underwater vehicles consists of a broadband continuous spectrum and a series of line spectra. Therefore, current location estimation for underwater acoustic targets primarily relies on broadband processing. Classical broadband high-resolution direction-finding algorithms mainly fall into two categories: the Incoherent Signal Subspace Method (ISM) and the Coherent Signal Subspace Method (CSM). The ISM method decomposes the broadband signal into several narrowband signals, processes each narrowband signal, and finally weights and synthesizes the processed results to obtain a broadband DOA estimate. Because this method only utilizes a portion of the broadband signal information in each frequency band, its estimation performance is low and its resolution is poor. The CSM method, on the other hand, proposes a "focusing" concept. Through a focusing transformation, it transforms the signal subspaces of each narrowband to the same reference frequency, forming a covariance matrix at a single frequency point, and then uses the narrowband subspace method to achieve DOA estimation. However, both the joint averaging of frequency division points in ISM and the frequency focusing processing in CSM assume that the incident signal has a consistent distribution range in the frequency domain. In practical applications, the exact frequency range of a broadband target is often difficult to define accurately in advance, and the same frequency band often contains the energy of other targets and environmental noise, leading to deviations in azimuth estimation. Specifically, this manifests as false peaks in the azimuth spectrum, target obscuration, or difficulty in distinguishing the target. Furthermore, since the covariance matrix needs to be calculated separately for each sub-band, the above methods also face challenges of high computational complexity and limited real-time performance. Despite their different approaches, these two methods share common limitations: they use the entire broadband data during signal processing, failing to distinguish the energy distribution differences between the target and interference in the frequency domain, and failing to utilize the acoustic characteristics of the target in the frequency dimension. Therefore, they are difficult to effectively extract the azimuth information of a specific target in scenarios with strong interference or multiple targets.

[0004] Analysis of extensive experimental data reveals that line spectrum characteristics are prevalent in the broadband radiated noise of underwater targets such as surface ships and unmanned underwater vehicles (UAVs), primarily originating from mechanical and propeller noise. Line spectra, characterized by frequency stability and concentrated energy, effectively characterize the energy distribution and inherent properties of targets, thus distinguishing them from background interference acoustically. However, current research on underwater acoustic target processing lacks a clearly defined high-resolution orientation estimation method based on the target's own line spectrum characteristics. Existing technologies have explored the application of harmonic focusing techniques in UAV positioning, but due to the complexity of the underwater environment and the lack of a systematic understanding of target noise characteristics, the effectiveness and applicability of such methods in passive underwater acoustic detection remain to be verified. Furthermore, existing technologies have also conducted research on target orientation estimation based on line spectrum characteristics, but these have not been organically combined with broadband high-resolution direction-finding algorithms and lack sufficient experimental data support; their engineering practicality and robustness require further investigation.

[0005] Existing broadband high-resolution algorithms typically decompose the array-received signal into multiple sub-bands using Discrete Fourier Transform (DFT) based on a pre-defined broadband frequency band. ISM (Independent Frequency Sampling) performs narrowband DOA estimation independently in each sub-band and averages the results across all frequency points to obtain the final azimuth estimate. CSM (Concentrated Frequency Sampling), on the other hand, uses a focusing transformation to map the signal subspace at each frequency point within the full bandwidth to the same reference frequency, and then uses a narrowband subspace method to achieve high-resolution azimuth estimation. However, both ISM's frequency-point joint averaging and CSM's frequency-point focusing processing assume that the incident signal has a consistent distribution range in the frequency domain. In practical applications, the exact frequency range of a broadband target is often difficult to define accurately in advance, and the same frequency band often contains energy from other targets and environmental noise, leading to deviations in azimuth estimation. These deviations manifest as false peaks in the azimuth spectrum, target obscuration, or difficulty in distinguishing the target. Furthermore, since the covariance matrix needs to be calculated separately for each sub-band, these methods also face challenges of high computational complexity and limited real-time performance. Summary of the Invention

[0006] This application provides a method, apparatus, electronic device, and storage medium for estimating the azimuth of underwater acoustic targets based on line spectrum focusing, thereby addressing the technical problems of existing broadband direction finding methods, such as weak noise resistance, low azimuth resolution, and high computational complexity in environments with strong interference.

[0007] To achieve the above objectives, the technical solution of this invention is as follows: In a first aspect, embodiments of the present invention provide a method for estimating the azimuth of an underwater acoustic target based on line spectrum focusing, comprising: receiving a broadband radiated noise signal of an underwater acoustic target through a sensor array, converting the received underwater acoustic signal into an electrical signal, processing it to obtain a discrete time-domain signal, and selecting an array element as a reference array element; A short-time Fourier transform is performed on the discrete-time domain signal of the reference array element, and the line spectrum frequency generated during the target motion is extracted through the time-frequency diagram. The discrete time-domain signal received by the array is subjected to discrete Fourier transform, the broadband frequency domain data is divided into multiple sub-bands, and the narrowband signal centered on each line spectrum frequency is extracted. Based on the estimated incident angle and preset reference frequency of the broadband signal, a focusing transformation matrix corresponding to each line spectrum frequency is constructed to minimize the error between the focused array manifold and the array manifold at the reference frequency. Based on the focusing transformation matrix, the narrowband signal corresponding to each line spectrum frequency is focused and transformed, and the signal subspace of each line spectrum frequency is mapped to the signal subspace of the reference frequency. The covariance matrix of the transformed signal is calculated and the geometric mean is taken to obtain the covariance matrix at the reference frequency. The covariance matrix at the reference frequency is decomposed into eigenvalues. After the eigenvalues ​​are sorted in descending order, the signal subspace and noise subspace are divided according to the number of signal sources and array elements. The MUSIC algorithm is used to search for spectral peaks to obtain the azimuth estimation result of the underwater acoustic target.

[0008] In some possible implementations, time-frequency analysis is performed on the discrete-time domain signal of the reference array elements to extract the line spectrum frequencies generated during target motion, including: A discrete short-time Fourier transform is performed on the time-domain signal of the reference array element to obtain the time-frequency diagram of the signal; where the time-domain signal of the reference array element... Its discrete short-time Fourier transform is expressed as: ; in, For discrete time intervals, The x-axis is the coordinate of the time-frequency graph. These are the corresponding frequency coordinates. Represents the window function. Using a sliding window function of length denoted as the time, the signal is analyzed piecewise along the time axis to obtain the time-frequency plot of the signal. Identify the line spectrum trajectory of a stable narrowband signal from the time-frequency plot; Extract one or more line spectrum frequencies within the optimal processing frequency band of the array.

[0009] In some possible implementations, the optimal processing frequency band of the array is predetermined through array performance testing, and line spectrum trajectory identification is achieved by screening the continuity features of multiple frames of power spectrum.

[0010] In some possible implementations, the construction of the focusing transformation matrix is ​​based on a rotation subspace algorithm, which satisfies the following relationship: ; in, To focus the transformation matrix, Line spectrum frequency The array manifold at that location, Reference frequency The array manifold at that location, For the estimated incident angle of broadband signals, The reference frequency, also called the focusing frequency, For line spectrum frequencies; Mapping the signal subspace of each line spectrum frequency onto the reference frequency, and then solving using the following formula: ; in, Denotes the Frobenius norm. Let be the identity matrix; the solution to the above formula is: ; in, They are The left and right singular vector matrices of the singular value decomposition.

[0011] In some possible implementations, the covariance matrix at the reference frequency is calculated using the following formula: ; in, The covariance matrix at the reference frequency, Let be the covariance matrix at each line spectrum frequency. This is the conjugate transpose of the focusing transformation matrix at each line spectrum frequency. This is the symbol for the conjugate transpose.

[0012] In some possible implementations, the covariance matrix at the reference frequency is decomposed into eigenvalues, as follows: ; in, The number of array elements For eigenvalues, For feature vectors, This is the conjugate transpose of the eigenvectors. The eigenvectors corresponding to the signal subspace. This is the diagonal matrix corresponding to the signal subspace. This is the conjugate transpose of the eigenvectors corresponding to the signal subspace. The feature vector corresponding to the noise subspace. This is the diagonal matrix corresponding to the noise subspace. This is the conjugate transpose of the eigenvectors corresponding to the noise subspace; Arrange the eigenvalues ​​in descending order, and then construct the signal subspaces of the corresponding eigenvectors according to the magnitude of their eigenvalues. and noise subspace , Let be the number of signal sources, then the spectral estimation formula for the MUSIC algorithm is: ; in, The azimuth spectrum is calculated using the MUSIC method. This is the conjugate transpose of the array manifold vector. For the noise subspace, This is the conjugate transpose of the noise subspace. For array manifold vectors; Through the By performing a spectral peak search, the azimuth estimation result of the underwater acoustic target signal can be obtained.

[0013] Secondly, embodiments of the present invention provide an underwater acoustic target location estimation device based on line spectrum focusing, comprising: The signal acquisition module is used to receive broadband radiated noise signals from underwater acoustic targets through a sensor array, convert the received underwater acoustic signals into electrical signals, process them to obtain discrete time-domain signals, and select an array element as a reference array element. The line spectrum frequency extraction module is used to perform short-time Fourier transform on the discrete time-domain signal of the reference array element and extract the line spectrum frequency generated when the target moves through the time-frequency diagram. The narrowband signal extraction module is used to perform discrete Fourier transform on the discrete time-domain signal received by the array, divide the broadband frequency-domain data into multiple sub-bands, and extract the narrowband signal centered on each line spectrum frequency. The focusing transformation matrix construction module is used to construct the focusing transformation matrix corresponding to each line spectrum frequency based on the estimated incident angle and preset reference frequency of the broadband signal, so as to minimize the error between the focused array manifold and the array manifold at the reference frequency. The focusing transformation module is used to perform focusing transformation on the narrowband signals corresponding to each line spectrum frequency according to the focusing transformation matrix, map the signal subspace of each line spectrum frequency to the signal subspace of the reference frequency, calculate the covariance matrix of the transformed signal and take the geometric mean to obtain the covariance matrix at the reference frequency. The azimuth estimation module is used to perform eigenvalue decomposition on the covariance matrix at the reference frequency. After arranging the eigenvalues ​​in descending order, the signal subspace and noise subspace are divided according to the number of signal sources and array elements. The MUSIC algorithm is used to search for spectral peaks to obtain the azimuth estimation result of the underwater acoustic target.

[0014] Thirdly, embodiments of the present invention provide an electronic device, including: a memory for storing executable instructions; and a processor for implementing the method provided in the first aspect of the present invention when executing the executable instructions or computer program stored in the memory.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing executable instructions for implementing the method provided in the first aspect of the present invention when executed by a processor.

[0016] One or more technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages: In this embodiment of the invention, interference suppression and extraction of a specified target trajectory are achieved by utilizing the specificity of different target radiation line spectra in the frequency domain. Unlike traditional methods that simply perform incoherent accumulation and averaging of frequency domain energy, this invention fully considers the inherent frequency domain characteristics of the target, focusing the energy near the target's own line spectrum frequency, thereby effectively eliminating interference and noise components in the frequency domain. This method not only significantly improves the accuracy of azimuth estimation and avoids the phenomenon of multiple targets being confused in BTR diagrams, but also greatly reduces computational complexity and saves computation time by performing focusing transformation only within a narrow band near the line spectrum. Attached Figure Description

[0017] To more clearly illustrate the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic flowchart of an embodiment of an underwater acoustic target orientation estimation method based on line spectrum focusing provided for the implementation of the present invention; Figure 2a A diagram showing the movement of a speedboat; Figure 2b This is a schematic diagram illustrating the change in distance between the speedboat target and the receiving array. Figure 3a This is a time-domain waveform diagram; Figure 3b This is a time-frequency graph; Figure 4a BTR diagram for the broadband MUSIC method; Figure 4b BTR diagram of the method provided in the embodiments of the present invention; Figure 5 This is a schematic diagram of the structure of an underwater acoustic target orientation estimation device based on line spectrum focusing in an embodiment of the present invention; Figure 6 This is a schematic diagram of an electronic device structure according to an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0020] In the relevant descriptions of this embodiment, the terms "including," "containing," and "possessing" are all open terms and are generally understood to include but not be limited to; the term "at least one" is generally understood to mean one or more, where "multiple" refers to two or more; the term "at least one of the following" or similar expressions refer to any combination of these items, including any combination of single or multiple items, for example, "at least one of a, b, or c", or "at least one of a, b, and c", which can all mean: a, b, c, ab (i.e., a and b), ac, bc, or abc, where a, b, and c can be single or multiple; the symbol "A / B" is used to describe the selection relationship of associated objects, generally indicating an "or" relationship.

[0021] In the following description of the embodiments, the terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms "a" and "the" as used in the embodiments of this application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0022] Those skilled in the art should understand that, in the following description of the embodiments of this application, the sequence of numbers does not imply the order of execution. Some or all steps may be executed in parallel or sequentially. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0023] Those skilled in the art will understand that the numerical ranges in the embodiments of this application should be understood to specifically disclose each intermediate value between the upper and lower limits of the range. Any stated value or intermediate value within a stated range, as well as any other stated value or each smaller range between intermediate values ​​within a range, are also included within this invention. The upper and lower limits of these smaller ranges may be independently included or excluded from the range.

[0024] Unless otherwise stated, the technical / scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. While this application describes only preferred methods and materials, any methods and materials similar or equivalent to those described herein may be used in the implementation or testing of this application. All references to this specification are incorporated by way of citation to disclose and describe the methods and / or materials associated with those references. In the event of any conflict with any incorporated reference, the content of this specification shall prevail.

[0025] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0026] Existing broadband azimuth estimation methods, when processing underwater acoustic target signals, typically perform signal subspace analysis based on a pre-defined broadband band, dividing the entire frequency band into multiple sub-bands for separate processing. However, in real-world marine environments, broadband interference sources (such as other targets, environmental noise, or active sonar signals) often coexist with the target signal in the same frequency band, making it difficult for traditional methods to effectively distinguish between the target and the interference. This can easily lead to problems such as false peaks, main lobe broadening, or target masking. Furthermore, because the covariance matrix needs to be calculated independently for each sub-band and then processed, the overall computational complexity of the algorithm is high, affecting real-time performance. Simultaneously, the target's own acoustic characteristics are not fully utilized, limiting further improvements in azimuth resolution.

[0027] Based on this, embodiments of the present invention provide a method for estimating the azimuth of underwater acoustic targets based on line spectrum focusing, which solves the technical problems of existing broadband direction finding methods having weak noise resistance, low azimuth estimation accuracy, and high computational complexity in strong interference environments.

[0028] Figure 1 A schematic flowchart illustrating an embodiment of the underwater acoustic target location estimation method based on line spectrum focusing, provided for implementation of this invention, is shown below. Figure 1 As shown, the above method may include: S101 receives broadband radiated noise signals from underwater acoustic targets through a sensor array, converts the received underwater acoustic signals into electrical signals, processes them to obtain discrete time-domain signals, and selects an array element as a reference array element. The sensor array is used to receive broadband radiated noise signals emitted by underwater acoustic targets. This array can be of any spatial arrangement, including but not limited to uniform linear arrays, uniform circular arrays, planar arrays, or polygonal arrays. Each element in the array is a sensor responsible for converting the received sound pressure fluctuations into electrical signals. The number of elements can be flexibly set according to the application scenario, generally no fewer than four to ensure basic spatial resolution. The array can be deployed in near-shore shallow waters, on a fixed seabed platform, or on a mobile carrier, suitable for static or dynamic detection missions. During reception, the target may be in uniform linear motion, variable speed motion, or maneuvering, and the array continuously samples to obtain time-series signals. The reference element can be a specific element selected from the underwater acoustic array, used as the base channel for time-frequency analysis. Its selection can be arbitrary, such as the first physically numbered element, or the channel with the highest signal-to-noise ratio.

[0029] For example, using the number of array elements is radius is A uniform circular array is used as the receiving array, fixed in the water near the shore, to receive broadband radiated noise signals from underwater acoustic targets. When the target moves, the sensors on the receiving circular array convert the received underwater acoustic signals into electrical signals, which are then processed by a preamplifier circuit and a data acquisition unit to obtain a discrete time-domain signal, denoted as . The first array element is selected as the reference array element.

[0030] S102 performs a short-time Fourier transform on the discrete-time domain signal of the reference array element and extracts the line spectrum frequency generated when the target moves through the time-frequency diagram; In some embodiments, step S102 may specifically include: The time-frequency diagram of the signal is obtained by performing a discrete short-time Fourier transform on the time-domain signal of the reference array element. Identify the line spectrum trajectory formed by a stable narrowband signal from the time-frequency plot; Extract one or more line spectrum frequencies within the optimal processing frequency band of the array.

[0031] Specifically, the time-domain signal of the reference array element. Its discrete short-time Fourier transform is expressed as: ; in, For discrete time intervals, The x-axis is the coordinate of the time-frequency graph. These are the corresponding frequency coordinates. Represents the window function. Given a window length, the signal is analyzed piecewise by sliding a window function on the time axis to obtain the time-frequency diagram of the signal.

[0032] In some embodiments, the optimal processing frequency band of the array is predetermined through array performance testing, and line spectrum trajectory identification is achieved by screening the continuity features of multiple frames of power spectrum.

[0033] Specifically, in a time-frequency graph composed of multiple power spectra, a stable narrowband signal will form a continuous, bright line, i.e., a line spectrum trajectory. This is distinct from randomly fluctuating broadband noise and intermittent interference. These line spectrum trajectories can be automatically extracted using image recognition algorithms or threshold detection methods. Combined with the optimal processing frequency band of the receiving array... arrive Extract the line spectrum frequencies present in this frequency band. .

[0034] S103 performs a discrete Fourier transform on the discrete time-domain signal received by the array, divides the broadband frequency-domain data into multiple sub-bands, and extracts the narrowband signal centered on each line spectrum frequency. For example, performing a Discrete Fourier Transform on the array-received data divides the wideband frequency domain data into... The sub-bands, with center frequencies in the following order: The frequency domain array received signal is obtained. To fully utilize the line spectrum characteristics of underwater acoustic target radiated noise, only the line spectrum is extracted here. The covariance matrix of the narrowband signal with the center frequency is calculated and expressed as: ; in, For the first The array receives data at each line spectrum frequency point.

[0035] S104, based on the estimated incident angle and preset reference frequency of the broadband signal, constructs the focusing transformation matrix corresponding to each line spectrum frequency, so that the error between the focused array manifold and the array manifold at the reference frequency is minimized. In some embodiments, the construction of the focusing transformation matrix is ​​based on the Rotational Signal Subspace (RSS) algorithm. The basic idea of ​​the RSS algorithm is to apply different frequency points... Construct different focusing transformation matrices from the received data This minimizes the error between the focused array manifold and the array manifold at the reference frequency, satisfying the following relationship: ; in, To focus the transformation matrix, Line spectrum frequency The array manifold matrix at that location, Reference frequency The array manifold matrix at that location, For the estimated incident angle of broadband signals, The reference frequency, also called the focusing frequency, For line spectrum frequencies; Mapping the signal subspace of each line spectrum frequency onto the reference frequency, and then solving using the following formula: ; in, Denotes the Frobenius norm. Let be the identity matrix; the solution to the above formula is: ; in, They are The left and right singular vector matrices of the singular value decomposition.

[0036] S105, according to the focusing transformation matrix, the narrowband signal corresponding to each line spectrum frequency is focused and transformed, and the signal subspace of each line spectrum frequency is mapped to the signal subspace of the reference frequency. The covariance matrix of the transformed signal is calculated and the geometric mean is taken to obtain the covariance matrix at the reference frequency. Specifically, for each extracted target line spectrum frequency, the array received data at that frequency point is transformed according to the focusing transformation matrix. The purpose is to map the signal subspace at each line spectrum frequency to a pre-selected common reference frequency, thereby solving the problem of inconsistent spatial phase response of broadband signals due to different frequencies.

[0037] In some embodiments, the covariance matrix at the reference frequency is calculated using the following formula: ; in, The covariance matrix at the reference frequency, Let be the covariance matrix at each line spectrum frequency. This is the conjugate transpose of the focusing transformation matrix at each line spectrum frequency. This is the symbol for the conjugate transpose.

[0038] S106. Eigenvalue decomposition is performed on the covariance matrix at the reference frequency. After arranging the eigenvalues ​​in descending order, the signal subspace and noise subspace are divided according to the number of signal sources and array elements. The MUSIC algorithm is used to search for spectral peaks to obtain the azimuth estimation result of the underwater acoustic target.

[0039] In some embodiments, eigenvalue decomposition of the covariance matrix at the reference frequency refers to performing algebraic decomposition on the single-frequency equivalent covariance matrix obtained after the array received signal undergoes line spectrum focusing transformation, in order to reveal its inherent signal and noise subspace structure. This covariance matrix has already been processed in the preceding steps by fusing information from multiple line spectrum frequencies using geometric averaging, resulting in a higher signal-to-noise ratio and purer target energy concentration characteristics. The specific eigenvalue decomposition process can be expressed as follows: ; in, The number of array elements For eigenvalues, For feature vectors, This is the conjugate transpose of the eigenvectors. The eigenvectors corresponding to the signal subspace. This is the diagonal matrix corresponding to the signal subspace. This is the conjugate transpose of the eigenvectors corresponding to the signal subspace. The feature vector corresponding to the noise subspace. This is the diagonal matrix corresponding to the noise subspace. This is the conjugate transpose of the eigenvectors corresponding to the noise subspace; Arrange the eigenvalues ​​in descending order, and then construct the signal subspaces of the corresponding eigenvectors according to the magnitude of their eigenvalues. and noise subspace , Let be the number of signal sources. Then, the spectral estimation formula for the MUSIC algorithm can be expressed as: ; in, The azimuth spectrum is calculated using the MUSIC method. This is the conjugate transpose of the array manifold vector. For the noise subspace, This is the conjugate transpose of the noise subspace. For array manifold vectors; Through the The location estimation result of the underwater acoustic target can be obtained by performing a spectral peak search.

[0040] In this embodiment of the invention, the line spectrum frequency components generated by the target itself are precisely selected. By reconstructing the covariance matrix and projecting the line spectrum signal subspace onto the reference frequency subspace, noise and interference are effectively eliminated at the frequency domain level. This method significantly improves the accuracy and spatial resolution of azimuth estimation while avoiding the computational burden of matrix operations over a wide frequency range. Ultimately, it achieves high-resolution, highly interference-resistant, and computationally inefficient azimuth estimation performance.

[0041] The process of steps S101 to S106 described above will be specifically explained using a specific embodiment below. The 12-element circular array has a diameter of 250 mm, a sampling rate of 32 kHz, and a deployment depth of 1.5 m. The relationship between the positional changes of the receiving array and the detection target is as follows: Figures 2a to 2b As shown, Figure 2a This is a diagram showing the movement of the speedboat. Figure 2b This diagram illustrates the change in distance between the speedboat target and the receiving array. As the speedboat target gradually approaches the receiving array, to verify the improved resolution and anti-interference capabilities of the proposed method, an LFM signal with a bandwidth of 2k-4k Hz is periodically transmitted as an interference source in a 288° direction. The signal duration is 2 seconds, with intervals of 5 seconds. Figure 2a The yellow dots in the middle.

[0042] The time-domain waveform and time-frequency diagram of the data received by the array reference element are as follows: Figures 3a to 3b As shown, Figure 3a This is a time-domain waveform diagram. Figure 3b This is a time-frequency diagram. From... Figure 3b As can be seen from the time-frequency diagram, even with periodic strong interference, the line spectrum component in the target radiated noise still exists and the signal-to-noise ratio is high. The optimal processing performance frequency band of this 12-ring array is located in the 2k-4k Hz band. Within this range, the line spectrum frequencies of the target motion are extracted as [1880 2080 3130 4200] Hz. The broadband processing frequency band is selected as 1800 Hz-4200 Hz. The ISM method and the method proposed in this embodiment are used to estimate the target azimuth of the speedboat. The BTR diagram obtained by stitching the azimuth spectrum along the time dimension is shown below. Figures 4a to 4b As shown, Figure 4a BTR diagram for the broadband MUSIC method. Figure 4b This is a BTR diagram of the method provided in the embodiments of the present invention.

[0043] In the presence of interference, many bright spots that do not belong to the target will appear on the BTR map, which will affect the trajectory tracking of the target's orientation change. When the spatial interval between the target and the interference is too small and the algorithm's resolution is too low, there will be a situation where the target and the interference are confused. Figure 4b This is the BTR map estimated by the ISM method in the presence of interference. The solid black line represents the target's true azimuth change, and the dashed black line represents the azimuth of the interference source. It can be seen that due to the presence of the interference source, the target's azimuth change trajectory exhibits periodic expansion, the main lobe is broadened, and the target's trajectory change becomes less clear. And... Figure 4bOn the BTR image, only the target's own line spectrum energy is used to calculate the covariance matrix. This eliminates interference sources as much as possible in the frequency domain. Combined with the data accumulation effect of the CSM method, noise immunity is enhanced. Therefore, there are no bright spots in the direction of the dashed line on the BTR image, indicating that interference sources are essentially suppressed, leaving only the trajectory of the detected target's azimuth change. The target's own trajectory is also clearer, with more prominent peaks within the same color scale range and a significantly reduced main lobe width. The target's azimuth can be estimated relatively accurately at each time point. Therefore, the method proposed in this invention has better anti-interference capability and azimuth resolution performance.

[0044] In this embodiment of the invention, when strong broadband interference exists in the environment and the spatial spacing of sound sources is small, existing broadband target azimuth estimation methods often struggle to effectively distinguish between the target and interference signals when processing actual underwater acoustic targets. Furthermore, they also face the problem of high computational burden in frequency domain processing. To address these issues, this embodiment proposes an underwater acoustic target azimuth estimation method based on line spectrum focusing. By utilizing the specificity of the line spectrum in the frequency domain of different target radiated noise, interference suppression and extraction of the specified target trajectory are achieved. Unlike traditional methods that simply perform incoherent accumulation and averaging of frequency domain energy, the method of this embodiment fully considers the inherent frequency domain characteristics of the target, focusing the energy near the target's own line spectrum frequency, thereby effectively eliminating interference and noise components in the frequency domain. This method not only significantly improves the accuracy of azimuth estimation and avoids the phenomenon of multiple targets being confused in the BTR diagram, but also greatly reduces computational complexity and saves computation time by performing focusing transformation only within a narrow band near the line spectrum.

[0045] Based on the same inventive concept, this application also provides an underwater acoustic target orientation estimation device based on line spectrum focusing. Figure 5 This is a schematic diagram of a hydroacoustic target location estimation device based on line spectrum focusing, as described in an embodiment of the present invention. Figure 5 As shown, the underwater acoustic target location estimation device 500 based on line spectrum focusing may include: The signal acquisition module 501 is used to receive the broadband radiated noise signal of the underwater acoustic target through the sensor array, convert the received underwater acoustic signal into an electrical signal, process it to obtain a discrete time domain signal, and select an array element as a reference array element. The line spectrum frequency extraction module 502 is used to perform short-time Fourier transform on the discrete time-domain signal of the reference array element and extract the line spectrum frequency generated when the target moves through the time-frequency diagram. The narrowband signal extraction module 503 is used to perform discrete Fourier transform on the discrete time domain signal received by the array, divide the broadband frequency domain data into multiple sub-bands, and extract the narrowband signal centered on each line spectrum frequency. The focusing transformation matrix construction module 504 is used to construct the focusing transformation matrix corresponding to each line spectrum frequency based on the estimated incident angle and preset reference frequency of the broadband signal, so as to minimize the error between the focused array manifold and the array manifold at the reference frequency. The focusing transformation module 505 is used to perform focusing transformation on the narrowband signal corresponding to each line spectrum frequency according to the focusing transformation matrix, map the signal subspace of each line spectrum frequency to the signal subspace of the reference frequency, calculate the covariance matrix of the transformed signal and take the geometric mean to obtain the covariance matrix at the reference frequency. The azimuth estimation module 506 is used to perform eigenvalue decomposition on the covariance matrix at the reference frequency. After arranging the eigenvalues ​​in descending order, the signal subspace and noise subspace are divided according to the number of signal sources and array elements. The MUSIC algorithm is used to search for spectral peaks to obtain the azimuth estimation result of the underwater acoustic target.

[0046] Based on the same inventive concept, this application provides an electronic device that can be consistent with the offshore wind power system reliability optimization method in one or more of the above embodiments. Figure 6 This is a schematic diagram of an electronic device structure according to an embodiment of the present invention. See also... Figure 6 As shown, the electronic device 600 can use general computer hardware, including a processor 601 and a memory 602.

[0047] In some possible implementations, at least one processor can constitute any physical device having circuitry that performs logical operations on one or more inputs. For example, at least one processor may include one or more integrated circuits, including application-specific integrated circuits, microchips, microcontrollers, microprocessors, all or part of a central processing unit, graphics processing unit, digital signal processor, field-programmable gate array, or other circuitry suitable for executing instructions or performing logical operations. Instructions executed by at least one processor may, for example, be preloaded into memory integrated with or embedded in the controller, or may be stored in separate memory. Memory may include random access memory, read-only memory, hard disk, optical disk, magnetic media, flash memory, other permanent, fixed, or volatile memory, or any other mechanism capable of storing instructions. In some embodiments, at least one processor may include more than one processor. Each processor may have a similar architecture, or processors may have different configurations that are electrically connected or disconnected from each other. For example, processors may be separate circuits or integrated into a single circuit. When more than one processor is used, the processors may be configured to operate independently or collaboratively. Processors may be coupled electrically, magnetically, optically, acoustically, mechanically, or by other means that allow them to interact.

[0048] Based on the same inventive concept, this application provides a computer storage medium storing computer-executable instructions. After being executed by a processor, the computer-executable instructions can realize the underwater acoustic target orientation estimation method based on line spectrum focusing as described in one or more of the above embodiments.

[0049] The various embodiments in this specification are described in a progressive manner. For the same or similar parts between the various embodiments, please refer to each other. Each embodiment focuses on describing the differences from other embodiments.

[0050] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of this application.

Claims

1. A method for underwater acoustic target direction estimation based on line spectrum focusing, characterized in that, The method comprises the following steps: Receiving a broadband radiation noise signal of an underwater acoustic target through a sensor array, converting the received underwater acoustic signal into an electric signal, obtaining a discrete time domain signal after processing, and selecting one array element as a reference array element; Performing short-time Fourier transform on the discrete time domain signal of the reference array element, and extracting a line spectrum frequency generated by target movement through a time-frequency diagram; Performing discrete Fourier transform on the discrete time domain signal received by the array, dividing broadband frequency domain data into multiple sub-bands, and extracting a narrowband signal centered on each line spectrum frequency; Based on a pre-estimated incident angle of the broadband signal and a pre-set reference frequency, constructing a focusing transform matrix corresponding to each line spectrum frequency, so that the array manifold after focusing has the minimum error with the array manifold of the reference frequency; According to the focusing transform matrix, performing focusing transform on the narrowband signal corresponding to each line spectrum frequency, mapping the signal subspace of each line spectrum frequency to the signal subspace of the reference frequency, calculating the covariance matrix of the transformed signal and taking the geometric mean, and obtaining the covariance matrix at the reference frequency; Performing eigenvalue decomposition on the covariance matrix at the reference frequency, arranging the eigenvalues in descending order, dividing the signal subspace and the noise subspace according to the number of signal sources and the number of array elements, performing spectral peak search by using the MUSIC algorithm, and obtaining the bearing estimation result of the underwater acoustic target.

2. The method of claim 1, wherein, The short-time Fourier transform on the discrete time domain signal of the reference array element and the extraction of the line spectrum frequency through the time-frequency diagram comprise the following steps: discrete short-time Fourier transform is performed on the time-domain signal of the reference array element to obtain a time-frequency graph of the signal; wherein the time-domain signal of the reference array element whose discrete short-time Fourier transform is represented as: ； wherein, is a discrete time instant, is an abscissa in the time-frequency diagram, is a corresponding frequency coordinate, denotes a window function, is a window length, the signal is analyzed in segments by sliding the window function over the time axis, resulting in a time-frequency diagram of the signal; Identifying a line spectrum track formed by a stable narrowband signal from the time-frequency diagram; Extracting one or more line spectrum frequencies within the array optimal processing frequency band.

3. The method of claim 2, wherein, The array optimal processing frequency band is determined in advance through array performance test, and the line spectrum track identification is realized through the continuity feature screening of multiple power spectrums.

4. The method of claim 1, wherein, The construction of the focusing transform matrix is based on the rotation subspace algorithm, which satisfies the following relationship: ； where, is the focusing matrix, is the line spectrum frequency is the array manifold at is the reference frequency is the array manifold at is the estimated angle of arrival of the wideband signal, is the reference frequency, also called the focusing frequency, is the line spectrum frequency; mapping the signal subspaces of each line spectrum frequency to the reference frequency is solved by the following equation: ； wherein denotes the Frobenius norm, is the identity matrix; the solution of the above formula is: ； wherein are respectively Left and right singular vector matrices of singular value decomposition.

5. The method of claim 1, wherein, The covariance matrix at the reference frequency is calculated by the following formula: ； wherein is the covariance matrix over the reference frequency, is the covariance matrix over the individual line spectrum frequencies, is the conjugate transpose of the focusing transform matrix over the individual line spectrum frequencies, is the conjugate transpose symbol.

6. The method of claim 5, wherein, The eigenvalue decomposition on the covariance matrix at the reference frequency is represented as: ； wherein, is the number of array elements, is the eigenvalue, is the eigenvector, is the conjugate transpose of the eigenvector, is the eigenvector corresponding to the signal subspace, is the diagonal matrix corresponding to the signal subspace, is the conjugate transpose of the eigenvector corresponding to the signal subspace, is the eigenvector corresponding to the noise subspace, is the diagonal matrix corresponding to the noise subspace, is the conjugate transpose of the eigenvector corresponding to the noise subspace; The characteristic values are arranged in descending order, and the corresponding characteristic vectors are respectively arranged in signal subspace according to the characteristic value size and noise subspace , M is the number of signal sources, and a spectrum estimation formula of the MUSIC algorithm is: ； wherein the direction spectrum computed for the MUSIC method, is the conjugate transpose of the array manifold vector, is the noise subspace, is the conjugate transpose of the noise subspace, is the array manifold vector; By performing a spectral peak search on the The bearing estimation result of the underwater acoustic target of the target signal can be obtained.

7. An apparatus for underwater acoustic target direction estimation based on line spectrum focusing, characterized in that, The method comprises the following steps: A signal acquisition module is configured to receive a broadband radiation noise signal of an underwater acoustic target through a sensor array, convert the received underwater acoustic signal into an electric signal, obtain a discrete time domain signal after processing, and select one array element as a reference array element; A line spectrum frequency extraction module is configured to perform short-time Fourier transform on the discrete time domain signal of the reference array element, and extract a line spectrum frequency generated by target movement through a time-frequency diagram; A narrowband signal extraction module is configured to perform discrete Fourier transform on the discrete time domain signal received by the array, divide broadband frequency domain data into multiple sub-bands, and extract a narrowband signal centered on each line spectrum frequency; A focusing transform matrix construction module is configured to construct a focusing transform matrix corresponding to each line spectrum frequency based on a pre-estimated incident angle of the broadband signal and a pre-set reference frequency, so that the array manifold after focusing has the minimum error with the array manifold of the reference frequency; A focusing transform matrix construction module is configured to construct a focusing transform matrix corresponding to each line spectrum frequency based on a pre-estimated incident angle of the broadband signal and a pre-set reference frequency, so that the array manifold after focusing has the minimum error with the array manifold of the reference frequency; a focusing transformation module, configured to perform focusing transformation on the narrowband signals corresponding to each line spectrum frequency according to the focusing transformation matrix, map the signal subspace of each line spectrum frequency to the signal subspace of the reference frequency, calculate the covariance matrix of the transformed signals and take the geometric mean to obtain the covariance matrix at the reference frequency; an orientation estimation module, configured to perform eigenvalue decomposition on the covariance matrix at the reference frequency, arrange the eigenvalues in descending order, divide the signal subspace and the noise subspace according to the number of signal sources and the number of array elements, perform spectral peak search by using the MUSIC algorithm, and obtain the orientation estimation result of the underwater acoustic target.

8. An electronic device, comprising: The electronic device comprises: a memory, configured to store executable instructions; a processor, configured to execute the executable instructions or computer programs stored in the memory, and realize the method in any one of claims 1 to 6.

9. A computer-readable storage medium storing executable instructions or a computer program, characterized in that, The executable instructions are executed by the processor to realize the method in any one of claims 1 to 6.

Images