Noise suppression method based on UUV motion under complex noise background

Abstract

The invention belongs to the field of signal processing, and particularly relates to a noise suppression method based on UUV motion under a complex noise background. The invention aims to solve the problem that the existing noise processing method is low in efficiency. According to the noise suppression method based on UUV motion under the complex noise background, the sound source positions and arrays of propeller noise and vibration noise are relatively unchanged in the steering period of the UUV, and the incident angle of a signal changes along with the course angle of a platform. If the environmental noise is kept stable, the noise shows the characteristic of being stable and unchanged during steering, and the signal shows the characteristic of changing along with the course angle. According to the method, the difference between the signal covariance and the noise covariance is utilized, changing signals are filtered out in the frequency domain, and the noise covariance is estimated. And then noise suppression is carried out by using the estimated noise covariance, so that detection in a complex noise environment is further realized, and the problem of low efficiency of an existing noise processing method is solved.

Description

Technical Field

[0001] This invention belongs to the field of signal processing, and in particular relates to a noise suppression method based on UUV motion under complex noise background. Background Technology

[0002] A UUV in motion is a high-noise platform. The array elements receive noise including flow noise, vibration noise, propeller noise, and environmental noise. When the array moves relative to the seawater, the outer surface of the hydrophone sheath generates end-flow pulsations in the seawater, thus exciting flow noise. Vibration noise is generated by the operation of the platform's main and auxiliary systems and is transmitted through the platform's mechanical structure. Propeller noise is caused by the periodic cutting of the medium by the propeller and is concentrated at the tail of the platform. Environmental noise is relatively simple and is generally considered to be isotropic Gaussian noise.

[0003] Different types of noise exhibit different characteristics in the target detection process. In terms of correlation, flow noise and environmental noise show very low correlation between different array elements, while propeller noise and vibration noise show strong correlation between elements and can be considered as a combination of complex near-field point sources. Spectrally, propeller noise and vibration noise are superpositions of line and continuous spectra, while flow noise and environmental noise are predominantly continuous spectra, with low and medium frequencies dominating in the environmental noise.

[0004] To address platform interference, a common approach is to suppress the interference before target detection. Currently, mainstream suppression methods fall into three categories: one is noise suppression based on the beam domain. For example, adaptive filtering techniques are used to minimize the error between the output signal and the desired signal. The difficulty of this type of algorithm lies in extracting the reference signal. For propeller noise and vibration noise, the former manifests as a near-field source in the end-fire direction and can be extracted using beam focusing methods. However, the spatial distribution of vibration noise is unknown, making it difficult to extract using the same method. Alternatively, beamforming can be used to generate nulls, preserving the array manifold of the signal direction while suppressing the array manifold of the noise. However, this creates a detection blind zone, resulting in vulnerabilities in target detection.

[0005] One approach is based on spatial sparsity and subspace projection to preserve signal features and suppress interference features. However, in constructing a sparse dictionary, near-field sources located in the Fresnel zone are generally considered. Furthermore, subspace-based methods have certain requirements regarding prior information such as the number of sources and noise power.

[0006] The last category is methods based on the statistical properties of the signal. These methods rely on the differences in statistical properties between the signal and interference, such as correlation, independence, and probability distribution, to establish statistical models and achieve interference suppression. For example, blind source separation methods, without knowing the mixing model or prior knowledge of the source signal, separate the source signal solely using the statistical independence of the observed signal. Alternatively, higher-order statistical methods utilize higher-order statistics of the signal, such as third- or fourth-order cumulants, to distinguish between the target and interference. However, both methods suffer from enormous computational complexity when dealing with arrays with a large number of elements, resulting in low efficiency in noise processing. Summary of the Invention

[0007] The purpose of this invention is to address the problem of low efficiency in existing noise processing methods. It provides a noise suppression method based on UUV motion in complex noise backgrounds, comprising:

[0008] Step 1: Allow the UUV to perform a constant speed turning motion for T seconds, and collect the UUV's heading angle data and array reception data during the constant speed turning motion; T is a positive integer.

[0009] The array receives data including signal reception data and noise reception data;

[0010] Step 2: Calculate the covariance matrix of the array-received data of the UUV during the uniform turning motion, and obtain T data covariance matrices; each second of array-received data of the UUV corresponds to one covariance matrix.

[0011] Sort the values ​​of the T elements at each position in the T data covariance matrices by time to obtain an array of the elements at all positions of the data covariance matrices as a function of time.

[0012] Furthermore, assuming the data covariance matrix has B elements, where B is a positive integer, then B arrays will be generated;

[0013] Each element at a given position corresponds to an array;

[0014] An array contains the values ​​of the element at T time points;

[0015] Step 3: Based on the heading angle data of the UUV during the uniform turning motion, calculate the frequency band distribution of all elements in the signal covariance matrix during the uniform turning motion.

[0016] Step 4: Perform a Fourier transform on the array of elements at all positions in the data covariance matrix obtained in Step 2, showing the changes over time, to obtain the frequency band distribution of elements at all positions in the data covariance matrix during the uniform turning motion.

[0017] Based on the frequency band distribution of all elements in the data covariance matrix during uniform speed turning motion and the frequency band distribution of all elements in the signal covariance matrix obtained in step three, the frequency band distribution of all elements in the noise covariance matrix during uniform speed turning motion is obtained.

[0018] The frequency band distribution of all elements in the noise covariance matrix during uniform turning motion is subjected to inverse Fourier transform to obtain an array of the elements of all positions in the noise covariance matrix as a function of time.

[0019] Step 5: Combine the arrays of elements at all positions of the noise covariance matrix that change over time into T co-noise covariance matrices;

[0020] Furthermore, since the elements of the noise covariance matrix and the data covariance matrix are the same, and the noise covariance matrix also has B elements, where B is a positive integer, then B arrays will be generated to show how the elements at all positions of the noise covariance matrix change over time.

[0021] An array contains the values ​​of an element at T time points; by filling in the values ​​of all (B) elements at the same time point according to their respective positions, T co-noise covariance matrices are obtained.

[0022] Each co-noise covariance matrix corresponds to a time period;

[0023] Step 6: Subtract the noise covariance matrix from the data covariance matrix to obtain the noise-suppressed data covariance matrix.

[0024] The beneficial effects of this invention are as follows:

[0025] This invention is based on the fact that during UUV turning, the source positions and arrays of propeller noise and vibration noise remain relatively constant, while the signal incident angle varies with the platform's heading angle. If we consider that environmental noise remains stable, then during turning, the noise exhibits a stable characteristic, while the signal exhibits a characteristic that varies with the heading angle. This invention utilizes the difference between the signal covariance and the noise covariance to filter out changing signals in the frequency domain and estimate the noise covariance. Then, it uses the estimated noise covariance for noise suppression, enabling detection in complex noise environments and solving the problem of low efficiency in existing noise processing methods. Attached Figure Description

[0026] Figure 1 This is a time history graph of the traditional CBF method;

[0027] Figure 2 This is the time history graph calculated by the present invention;

[0028] Figure 3 This is a comparison diagram of the azimuth spectrum under different preset power levels according to the present invention;

[0029] Figure 4 This is a comparison chart of azimuth estimation errors when SNR is between -15dB and 0dB;

[0030] Figure 5 This is a comparison chart of detection probabilities when SNR is between -15dB and 0dB. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will now be clearly and completely described with reference to the method flowcharts in the examples. Obviously, the described examples are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0032] Specific implementation method one: Combining Figures 1-5 This invention is described.

[0033] Step 1: Allow the UUV to perform a constant speed turning motion for T seconds, and collect the UUV's heading angle data and array reception data during the constant speed turning motion; T is a positive integer.

[0034] The array receives data including signal reception data and noise reception data;

[0035] Step 2: Calculate the covariance matrix of the array-received data of the UUV during the uniform turning motion, and obtain T data covariance matrices; each second of array-received data of the UUV corresponds to one covariance matrix.

[0036] Sort the values ​​of the T elements at each position in the T data covariance matrices by time to obtain an array of the elements at all positions of the data covariance matrices as a function of time.

[0037] Furthermore, assuming the data covariance matrix has B elements, where B is a positive integer, then B arrays will be generated;

[0038] Each element at a given position corresponds to an array;

[0039] An array contains the values ​​of the element at T time points;

[0040] Step 3: Based on the heading angle data of the UUV during the uniform turning motion, calculate the frequency band distribution of all elements in the signal covariance matrix during the uniform turning motion.

[0041] Step 4: Perform a Fourier transform on the array of elements at all positions in the data covariance matrix obtained in Step 2, showing the changes over time, to obtain the frequency band distribution of elements at all positions in the data covariance matrix during the uniform turning motion.

[0042] Based on the frequency band distribution of all elements in the data covariance matrix during uniform speed turning motion and the frequency band distribution of all elements in the signal covariance matrix obtained in step three, the frequency band distribution of all elements in the noise covariance matrix during uniform speed turning motion is obtained.

[0043] The frequency band distribution of all elements in the noise covariance matrix during uniform turning motion is subjected to inverse Fourier transform to obtain an array of the elements of all positions in the noise covariance matrix as a function of time.

[0044] Step 5: Combine the arrays of elements at all positions of the noise covariance matrix that change over time into T co-noise covariance matrices;

[0045] Furthermore, since the elements of the noise covariance matrix and the data covariance matrix are the same, and the noise covariance matrix also has B elements, where B is a positive integer, then B arrays will be generated to show how the elements at all positions of the noise covariance matrix change over time.

[0046] An array contains the values ​​of an element at T time points; by filling in the values ​​of all (B) elements at the same time point according to their respective positions, T co-noise covariance matrices are obtained.

[0047] Each co-noise covariance matrix corresponds to a time period;

[0048] Step 6: Subtract the noise covariance matrix from the data covariance matrix to obtain the noise-suppressed data covariance matrix.

[0049] Specific Implementation Method Two: The difference between this implementation method and Specific Implementation Method One is that...

[0050] The heading angle data of the UUV during the uniform turning motion in step one is represented by U, and can be expressed by the formula:

[0051]

[0052] In the formula, This represents the heading angle of the UUV at time t during its uniform turning motion. The heading angle of the UUV at time T during its uniform turning motion;

[0053] The array-received data of the UUV during the uniform turning motion is represented by X, and can be expressed by the formula:

[0054]

[0055] In the formula, This represents the array-received data of the UUV at time t during its uniform turning motion. The array received data of the UUV at time T during uniform turning motion;

[0056] The array received data of the UUV at time t during the uniform turning motion. Expressed by the formula:

[0057]

[0058] In the formula, A represents the array manifold. This represents the signal data received by the array of the UUV at time t during its uniform turning motion. This represents the noise data in the array-received data of the UUV at time t during uniform turning motion;

[0059] An array manifold is an M×L matrix (M is the number of array elements, and L is the number of signal sources). Its core function is to establish the mapping relationship between the direction of signal arrival and the data received by the array. Specifically, it includes: the phase delay of each array element with respect to the signal; array geometric structure information (such as uniform linear array / circular array, etc.); and array element radiation pattern characteristics.

[0060] The signal is the echo (L×1 vector) reflected from the target; there are a total of L signal sources; the number of signal sources is arbitrary and unknown, and it is equal to the number of signals in the actual sound field.

[0061] The noise received by the array elements includes flow noise, vibration noise, propeller noise, and environmental noise;

[0062] The other steps and parameters are the same as in Specific Implementation Method 1.

[0063] Specific Implementation Method Three: The difference between this implementation method and Specific Implementation Methods One and Two is that...

[0064] The specific process for calculating the covariance matrix of the array received data per second of UUV in step two, and obtaining T covariance matrices, is as follows:

[0065] The covariance matrix of the array received data at time t during uniform turning motion is expressed as follows: ,calculate The formula is expressed as:

[0066]

[0067] In the formula, Indicates the heading angle is The array manifold at that time; express The conjugate; express The conjugate; express conjugate, The signal data at time t during uniform turning motion; The noise data is for time t during uniform turning motion.

[0068] The other steps and parameters are the same as in one of the specific implementation methods one or two.

[0069] Specific Implementation Method Four: This implementation method differs from Specific Implementation Methods One to Three in that...

[0070] In step two, the values ​​of the T elements at each position in the T data covariance matrices are sorted by time to obtain an array showing the changes of elements at all positions over time; the specific process is as follows:

[0071] Sort the values ​​of the T elements at position (p, r) in the covariance matrix of T data points by time to obtain an array showing the change of the element at position (p, r) over time. This ultimately yields an array showing how the elements at all positions change over time.

[0072] The other steps and parameters are the same as those in one of the specific implementation methods one to three.

[0073] Specific Implementation Method Five: The difference between this implementation method and Specific Implementation Methods One to Four is that...

[0074] The process involves sorting the values ​​of the T elements at position (p, r) in the T data covariance matrices by time to obtain an array showing the changes of the elements at position (p, r) over time. This can be expressed as a formula:

[0075]

[0076] In the formula, p represents the row position and r represents the column position. express The element at position (p, r) in the middle, express The element at position (p, r) in the middle, express The element at position (p, r); The covariance matrix of the array received data at time 1 during uniform turning motion; The covariance matrix of the array received data at time 2 during uniform turning motion; The covariance matrix of the array received data at time 3 during uniform turning motion; The covariance matrix of the array received data at time t during uniform turning motion; The covariance matrix of the array received data at time T during uniform turning motion;

[0077] Expressed as a formula:

[0078]

[0079] in, The first time at time t of uniform turning motion l Position in the covariance matrix of each signal Element; The noise covariance matrix at time t during uniform turning motion The element at position, where L represents the total number of signals;

[0080] The other steps and parameters are the same as those in one of the specific implementation methods one to four.

[0081] Specific Implementation Method Six: The difference between this implementation method and Specific Implementation Methods One to Five is that...

[0082] The time t of the uniform turning motion l Position in the covariance matrix of each signal elements Expressed as a formula:

[0083]

[0084] in, For the first l The power of the signal; j represents the imaginary number; f is the signal frequency; The spacing between array elements; For the first l The azimuth angle of the signal; c represents the speed of sound.

[0085] The array receives data including L signal data and noise data. Therefore, the covariance matrix of the array receives data also includes L signal covariance matrices and noise covariance matrices.

[0086] The other steps and parameters are the same as those in any of the specific implementation methods one to five.

[0087] Specific Implementation Method Seven: The difference between this implementation method and Specific Implementation Methods One through Six is ​​that...

[0088] In step three, the frequency band distribution of each element in the signal covariance matrix during the uniform turning motion is calculated based on the heading angle data of the UUV collected during the uniform turning motion; the specific process is as follows:

[0089] Based on the heading angle data of the UUV collected during the uniform turning motion, the first yaw angle at time t during the uniform turning motion is calculated. l Position in the covariance matrix of each signal elements The frequency band distribution is expressed by the formula:

[0090]

[0091]

[0092] In the formula, Indicates intermediate variables. Indicates the first l The location of the signal;

[0093] The signal covariance matrix is l The result is obtained by adding the covariance matrices of the signals;

[0094] From the above formula, it can be concluded that, neglecting minute errors, Consists of a finite number of single-frequency components Combinatorial representation.

[0095] In this invention, the signal orientation is considered... constant, The spectrum is only related to the heading angle Related, manifested as octave distribution;

[0096] By analyzing the spectrum of a signal, we can determine its frequency band distribution. For example, calculating the spectrum of a Bluetooth signal and finding that its energy is mainly concentrated between 2.400 GHz and 2.480 GHz, we can then determine the frequency band distribution of this Bluetooth signal.

[0097] The other steps and parameters are the same as those in any of the specific implementation methods one to six.

[0098] Specific Implementation Method Eight: The difference between this implementation method and Specific Implementation Methods One to Seven is that...

[0099] In step four, a Fourier transform is performed on the array of elements at all positions in the data covariance matrix obtained in step two, which represents the changes over time, to obtain the frequency band distribution of elements at all positions in the data covariance matrix during the uniform turning motion.

[0100] Based on the frequency band distribution of all elements in the data covariance matrix during uniform speed turning motion and the frequency band distribution of all elements in the signal covariance matrix obtained in step three, the frequency band distribution of all elements in the noise covariance matrix during uniform speed turning motion is obtained.

[0101] The frequency band distribution of all elements in the noise covariance matrix during uniform turning motion is subjected to inverse Fourier transform to obtain an array of the elements of all positions in the noise covariance matrix as a function of time.

[0102] The specific process is as follows: The positions in the data covariance matrix obtained in step two... An array whose elements change over time Perform a Fourier transform to obtain the position in the covariance matrix of the data during the uniform turning motion. The frequency band distribution of the elements,

[0103] Based on the position in the covariance matrix of the data during uniform turning motion elements The position in the frequency band distribution and the signal covariance matrix obtained in step three The frequency band distribution of the elements is used to obtain the positions in the noise covariance matrix during uniform turning motion. The frequency band distribution of the elements;

[0104] Position in the noise covariance matrix during uniform turning motion The frequency band distribution of the elements is subjected to inverse Fourier transform to obtain the position of the noise covariance matrix. An array whose elements change over time The formula is as follows:

[0105]

[0106] In the formula, express The element at position (p, r) in the middle, express The element at position (p, r) in the middle, express The element at position (p, r); The noise covariance matrix represents the noise data at time 1 of the uniform turning motion. The noise covariance matrix represents the noise data at time 2 of uniform turning motion. The noise covariance matrix represents the noise data at time 3 during uniform turning motion. The noise covariance matrix represents the noise data at time t during uniform turning motion. The noise covariance matrix represents the noise data at time T during uniform turning motion.

[0107] The other steps and parameters are the same as those in any of the specific implementation methods one to seven.

[0108] Specific Implementation Method Nine: The difference between this implementation method and Specific Implementation Methods One through Eight is that...

[0109] The noise covariance matrix of the noise data at time t during the uniform turning motion The element at position (p, r) Expressed as a formula:

[0110] when hour, Expressed as a formula:

[0111]

[0112] in This is a vector for frequency filtering.

[0113] Frequency filtering vector The element at the corresponding frequency The value of is 0, and the value of the rest is 1;

[0114] For a dimension Vie Perform a Fourier transform to obtain the spectrum. .

[0115] Too It is dimensional, but it represents different frequencies in different positions.

[0116] The frequency filtering method involves multiplying the unwanted frequencies by 0 and the desired frequencies by 1. This distribution corresponds to the frequency filtering coefficients g. Dimensional, at a fixed frequency The value is 0 at one frequency and 1 at other frequencies.

[0117] so After multiplying it, the energy at a specific frequency becomes 0. After inverse Fourier transform, the effect of frequency filtering is achieved.

[0118] Fourier transform; Inverse Fourier transform;

[0119] This represents the DC component of the unit energy signal; This is the preset total power of the signal;

[0120] Here, "signal" refers to "Unit energy" means that the power of the signal is 1. Its own power is 1, for example This indicates a signal with a power of 2.

[0121] The DC component refers to the part of the Fourier transform where the frequency is 0.

[0122] It can be calculated from the formula, in relation to The decomposition formula includes constant terms, and the sum of all constant terms is the DC component.

[0123] when hour, Expressed as a formula:

[0124]

[0125] The other steps and parameters are the same as those in one of the specific implementation methods one to eight.

[0126] Specific Implementation Method Ten: The difference between this implementation method and Specific Implementation Methods One through Nine is that...

[0127] In step six, the noise covariance matrix is ​​subtracted from the data covariance matrix to obtain the noise-suppressed data covariance matrix, which can be expressed by the formula:

[0128]

[0129] In the formula, This represents the data covariance matrix after noise suppression. Represents the data covariance matrix. Represents the noise covariance matrix;

[0130] Furthermore, the noise-suppressed data covariance matrix obtained in this invention can be used in target detection methods, which are methods for azimuth estimation using data covariance. Examples include CBF, music, and MVDR methods.

[0131] An embodiment of the orientation estimation method of the present invention takes CBF as an example, and the orientation spectrum in the CBF algorithm... The following calculations are obtained:

[0132] (10)

[0133] in For angle The guide vector

[0134] The other steps and parameters are the same as those in any of the specific implementation methods one to nine.

[0135] Simulation experiments will be conducted to illustrate specific implementation methods one through ten:

[0136] An 8-element uniform linear array with an element spacing of 0.75 meters and a sound speed of 1500 m / s was used. The array performed a complete circular motion within the sound field, acquiring data from 0° to 360° heading angles in 0.1° increments. The number of snapshots per heading angle was 1024.

[0137] Assuming all noise follows a Gaussian distribution, the noise composition is as follows:

[0138] 1. Propeller noise: Single near-field source, located 5 meters from the array end-fire direction, with a signal-to-noise ratio of 5 dB;

[0139] 2. Vibration noise: Near-field sources with close proximity are randomly distributed within a 0.5m*0.5m rectangular area 1 meter away from the array normal, with a signal-to-interference ratio of 5dB;

[0140] 3. Flow noise and ambient noise: uncorrelated Gaussian noise between array elements;

[0141] Simulations were performed using the above noise parameters. Two 1kHz sound sources were present in the sound field, located at 30° and 60° at time 0. The signal-to-noise ratio remained at 0dB, and the total signal power was 1, following a Gaussian distribution. The CBF method was selected to verify the effectiveness of the proposed method, yielding the following results. Figure 1 , 2 The timeline shown is a graph.

[0142] from Figure 1 , Figure 2 As can be seen from the data, there is strong propeller noise at the 0° and 180° positions, which generates high sidelobes in the CBF method, making the signal peak insignificant. The method of this invention can effectively remove propeller noise and vibration noise while preserving the signal, and still has good detection capability even when the signal is close to the end-firing direction.

[0143] Figure 3 The azimuth spectrum under different preset power is given, which shows that changing the preset value of signal power within a certain range will not have a significant impact on the position of the signal peak.

[0144] Figure 4 , Figure 5 The traditional CBF method is presented, along with the detection probability and azimuth estimation error under different preset power levels. From... Figure 4 As can be seen, under complex noise interference, even with a high signal-to-noise ratio, the CBF method without denoising still struggles to accurately detect the signal, and the estimated signal location also has significant errors. However, different preset power levels have minimal impact on the method of this invention, all exhibiting good detection performance.

[0145] The main technical features of this invention include:

[0146] During UUV turning, the source positions and arrays of propeller noise and vibration noise remain relatively constant, while the signal incident angle varies with the platform's heading angle. If we consider that environmental noise remains stable, then during turning, the noise exhibits a stable characteristic, while the signal exhibits a characteristic that varies with the heading angle. This invention utilizes the difference between signal covariance and noise covariance to filter out changing signals in the frequency domain and estimate the noise covariance. Then, the estimated noise covariance is used for noise suppression, enabling detection in complex noise environments.

[0147] The above description is merely of preferred embodiments of the present invention. It should be understood that the present invention is not limited to the specific embodiments described above. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent substitutions, and improvements made to the above embodiments without departing from the scope of the present invention, based on the technical essence of the present invention, and within the spirit and principles of the present invention, shall still fall within the protection scope of the present invention.

Claims

1. A noise suppression method based on UUV motion under complex noise background, characterized in that, include: Step 1: Allow the UUV to perform a constant speed turning motion for T seconds, and collect the UUV's heading angle data and array reception data during the constant speed turning motion. T is a positive integer; The array receives data including signal reception data and noise reception data; Step 2: Calculate the covariance matrix of the array-received data of the UUV during the uniform turning motion, and obtain T data covariance matrices. Sort the values ​​of the T elements at each position in the T data covariance matrices by time to obtain an array of the elements at all positions of the data covariance matrices as a function of time. Step 3: Based on the heading angle data of the UUV during the uniform turning motion, calculate the frequency band distribution of all elements in the signal covariance matrix during the uniform turning motion. Step 4: Perform a Fourier transform on the array of elements at all positions in the data covariance matrix obtained in Step 2, showing the changes over time, to obtain the frequency band distribution of elements at all positions in the data covariance matrix during the uniform turning motion. Based on the frequency band distribution of all elements in the data covariance matrix during uniform speed turning motion and the frequency band distribution of all elements in the signal covariance matrix obtained in step three, the frequency band distribution of all elements in the noise covariance matrix during uniform speed turning motion is obtained. The frequency band distribution of all elements in the noise covariance matrix during uniform turning motion is subjected to inverse Fourier transform to obtain an array of the elements of all positions in the noise covariance matrix as a function of time. Step 5: Combine the arrays of elements at all positions of the noise covariance matrix that change over time into T co-noise covariance matrices; Step 6: Subtract the noise covariance matrix from the data covariance matrix to obtain the noise-suppressed data covariance matrix.

2. The noise suppression method based on UUV motion under complex noise background according to claim 1, characterized in that, The heading angle data of the UUV during the uniform turning motion in step one is represented by U, and can be expressed by the formula: In the formula, This represents the heading angle of the UUV at time t during its uniform turning motion. The heading angle of the UUV at time T during its uniform turning motion; The array-received data of the UUV during the uniform turning motion is represented by X, and can be expressed by the formula: In the formula, This represents the array-received data of the UUV at time t during its uniform turning motion. The array received data of the UUV at time T during uniform turning motion; The array received data of the UUV at time t during the uniform turning motion. Expressed by the formula: In the formula, A represents the array manifold. This represents the signal data received by the array of the UUV at time t during its uniform turning motion. This represents the noise data in the array received data of the UUV at time t during uniform turning motion.

3. The noise suppression method based on UUV motion under complex noise background according to claim 2, characterized in that, The specific process for calculating the covariance matrix of the array received data per second of UUV in step two, and obtaining T covariance matrices, is as follows: The covariance matrix of the array received data at time t during uniform turning motion is expressed as follows: ,calculate The formula is expressed as: In the formula, Indicates the heading angle is The array manifold at that time; express Conjugate; express Conjugate; express conjugate, The signal data at time t during uniform turning motion; The noise data is for time t during uniform turning motion.

4. The noise suppression method based on UUV motion under complex noise background according to claim 3, characterized in that, In step two, the values ​​of the T elements at each position in the T data covariance matrices are sorted by time to obtain an array showing the changes of elements at all positions over time; the specific process is as follows: Sort the values ​​of the T elements at position (p, r) in the covariance matrix of T data points by time to obtain an array showing the change of the element at position (p, r) over time. This ultimately yields an array of all elements at all positions that change over time.

5. A noise suppression method based on UUV motion under complex noise background according to claim 4, characterized in that, The process involves sorting the values ​​of the T elements at position (p, r) in the T data covariance matrices by time to obtain an array showing the changes of the elements at position (p, r) over time. This can be expressed as a formula: In the formula, p represents the row position and r represents the column position. express The element at position (p, r) in the middle, express The element at position (p, r) in the middle, express The element at position (p, r); The data covariance matrix at time 1 of uniform turning motion; This represents the data covariance matrix at time 2 during the uniform turning motion; The data covariance matrix at time 3 during uniform turning motion; This represents the data covariance matrix at time t during uniform turning motion; This represents the data covariance matrix at time T during uniform turning motion; Expressed as a formula: in, The first time at time t of uniform turning motion l Position in the covariance matrix of each signal Element; The noise covariance matrix at time t during uniform turning motion The element at position, where L represents the total number of signals.

6. A noise suppression method based on UUV motion under complex noise background according to claim 5, characterized in that, The time t of the uniform turning motion l Position in the covariance matrix of each signal elements Expressed as a formula: in, For the first l The power of the signal; j represents the imaginary number; f is the signal frequency; The spacing between array elements; For the first l The azimuth angle of the signal; c represents the speed of sound.

7. A noise suppression method based on UUV motion under complex noise background as described in claim 6, characterized in that, In step three, the frequency band distribution of each element in the signal covariance matrix during the uniform turning motion is calculated based on the heading angle data of the UUV collected during the uniform turning motion; the specific process is as follows: Based on the heading angle data of the UUV collected during the uniform turning motion, the first yaw angle at time t during the uniform turning motion is calculated. l Position in the covariance matrix of each signal elements The frequency band distribution is expressed by the formula: In the formula, Indicates intermediate variables. Indicates the first l The location of the signal.

8. A noise suppression method based on UUV motion under complex noise background according to claim 7, characterized in that, In step four, a Fourier transform is performed on the array of elements at all positions in the data covariance matrix obtained in step two, which represents the changes over time, to obtain the frequency band distribution of elements at all positions in the data covariance matrix during the uniform turning motion. Based on the frequency band distribution of all elements in the data covariance matrix during uniform speed turning motion and the frequency band distribution of all elements in the signal covariance matrix obtained in step three, the frequency band distribution of all elements in the noise covariance matrix during uniform speed turning motion is obtained. The inverse Fourier transform is performed on the frequency band distribution of all elements in the noise covariance matrix during the uniform turning motion to obtain an array showing the time-varying elements of the noise covariance matrix. The specific process is as follows: The position of the data in the covariance matrix obtained in step two An array whose elements change over time Perform a Fourier transform to obtain the position in the covariance matrix of the data during the uniform turning motion. The frequency band distribution of the elements, Based on the position in the covariance matrix of the data during uniform turning motion elements The position in the frequency band distribution and the signal covariance matrix obtained in step three The frequency band distribution of the elements is used to obtain the positions in the noise covariance matrix during uniform turning motion. The frequency band distribution of the elements; Position in the noise covariance matrix during uniform turning motion The frequency band distribution of the elements is subjected to inverse Fourier transform to obtain the position of the noise covariance matrix. An array whose elements change over time The formula is as follows: In the formula, express The element at position (p, r) in the middle, express The element at position (p, r) in the middle, express The element at position (p, r); The noise covariance matrix represents the time 1 of the uniform turning motion. The noise covariance matrix represents the time 2 of the uniform turning motion. The noise covariance matrix represents the time 3 of uniform turning motion. The noise covariance matrix represents the time t during uniform turning motion. This represents the noise covariance matrix at time T during uniform turning motion.

9. A noise suppression method based on UUV motion under complex noise background according to claim 8, characterized in that, The noise covariance matrix at time t during the uniform turning motion The element at position (p, r) The calculation formula is expressed as follows: when hour, The calculation formula is expressed as follows: in This is a vector for frequency filtering. Fourier transform; Inverse Fourier transform; This represents the DC component of the unit energy signal; This is the preset total power of the signal; when hour, The calculation formula is expressed as follows: 。 10. A noise suppression method based on UUV motion under complex noise background according to claim 8, characterized in that, In step six, the noise covariance matrix is ​​subtracted from the data covariance matrix to obtain the noise-suppressed data covariance matrix, which can be expressed by the formula: In the formula, This represents the data covariance matrix after noise suppression. Represents the data covariance matrix. This represents the noise covariance matrix.

Images