Line spectrum feature enhancement method and system based on phase gradient error estimation technology

Abstract

The present invention discloses a line spectrum feature enhancement method and system based on phase gradient error estimation technology, including: 1. Acquire observation array signals; 2. Roughly estimate the target azimuth based on ideal array beamforming; 3. Perform conventional beamforming processing based on the estimated target azimuth to obtain tracking beam data in the target direction, estimate a wideband continuous spectrum based on the data, and perform background equalization processing on the tracking beam data; 4. Detect L strong line spectra of the target signal based on threshold technology; 5. Eliminate linear delay. 6. Estimate the secondary and higher-order delay gradients between adjacent array elements based on the strong line spectrum after cyclic shift. 7. Integrate the estimated delay gradient to obtain the delay estimate of each array element relative to the reference array element. 8. Compensate for the array signal delay difference based on the estimated delay, obtain the target tracking beam through conventional beamforming, and obtain the line spectrum feature with enhanced fidelity through Fourier transform. The present invention realizes line spectrum feature enhancement under arbitrary array conditions.

Description

Technical Field

[0001] The present invention belongs to the technical field of sonar signal processing, and in particular relates to a line spectrum feature enhancement method based on phase gradient error estimation technology and a system thereof. Background Art

[0002] With the continuous improvement of ship noise reduction technology, the power of ship radiated noise is getting lower and lower, and target detection based on a single hydrophone is becoming more and more difficult. Compared with a single hydrophone, the hydrophone array can effectively improve the signal-to-noise ratio of the received signal and can estimate the direction of the target's wave accordingly. It has been widely used in underwater acoustic target detection and identification. The towed linear array is currently the most widely used hydrophone array form for underwater target detection. By increasing the aperture of the hydrophone array, it can effectively detect long-distance weak underwater targets and has the ability to receive low-frequency and even very low-frequency sound waves. On the other hand. The towed array is far away from the working mother ship, and the noise of the working platform is less affected, which improves the signal-to-noise ratio of the received signal. However, the linear towed array is susceptible to external influences such as ocean currents, wind and waves, and the movement of the working platform, resulting in array distortion and deviation of each array element from the ideal position, thereby reducing the detection performance of the towed array.

[0003] For array distortion, the existing processing methods are mostly array shape estimation. Towed array shape estimation methods are divided into two categories: one is the non-acoustic method, which estimates the movement of array elements by combining additional sensors installed on the array elements with dynamic equations and state space models. The other is the acoustic method, which simply infers the array shape through the array receiving signals, and generally converts the array shape estimation problem into a parameter estimation problem. Among the non-acoustic methods, the specific towed array shape estimation can be divided into two categories: (1) interpolation fitting method. The formation is estimated using the sensor measurement values, that is, the attitude information at different positions such as heading, pitch and roll angles are used to estimate through polynomial fitting or interpolation fitting methods. (2) Fluid dynamics method. The underwater state of the towed array is modeled by fluid dynamics, and the boundary conditions are obtained from the sensor values. The equations are solved to obtain the formation. The non-acoustic correction method can usually obtain a more accurate formation, but the disadvantage is that it requires additional sensors for assistance, which increases the hardware overhead compared to the acoustic method.

[0004] Among acoustic methods, they can be divided into two categories according to whether or not an auxiliary signal source with an accurately known position is required: one is active correction; the other is self-correction. The active correction algorithm estimates the array parameters offline by setting an auxiliary signal source with an accurately known position in space. There is no need to estimate the signal source position, and the amount of computation is relatively small. However, the position information of the auxiliary signal source must be very accurate. When the actual position of the auxiliary signal source deviates, the formation estimation of this type of algorithm will have a large error accordingly. Since the towed array is in a maneuvering state, the applicability of the auxiliary signal source method is limited. The self-correction formation estimation algorithm usually jointly estimates the source target direction and the formation error, and constructs a joint estimation function for optimization. The self-correction algorithm can estimate the source position by itself, without the need for an auxiliary signal source, and has higher accuracy than the active correction. However, the corresponding self-correction algorithm generally has a large amount of computation and a relatively high time complexity. Summary of the invention

[0005] The purpose of the present invention is to provide a line spectrum feature enhancement method and system based on phase gradient error estimation technology to solve the technical problem that line spectrum feature enhancement cannot be achieved under arbitrary array conditions.

[0006] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows:

[0007] A method for enhancing line spectrum features based on phase gradient error estimation technology specifically comprises the following steps:

[0008] Step 1: Get the observation array signal x i (t), i = 1, 2, ..., M, M is the number of array elements in the tow array;

[0009] Step 2: Estimate target direction based on ideal array beamforming is the guidance angle when the target signal beam energy is maximum;

[0010] Step 3: Perform beamforming processing based on the estimated target orientation to obtain the tracking beam g(t) and spectrum G(f) in the target direction, estimate the broadband continuous spectrum based on the data, perform background equalization processing on the tracking beam g(t) and spectrum G(f), whiten the non-white broadband continuous spectrum, and remove the continuous spectrum;

[0011] Step 4: equalize the spectrum G line (f) Perform line spectrum detection to detect the L strong line spectrum frequencies of the target signal whose equalized spectrum amplitude is greater than the local 3σ L is the number of estimated line spectra; σ is G line (f) surrounding noise variance;

[0012] Step 5: Perform Fourier transform on the time domain waveform of each array element to obtain X i(f), and for the L detected Perform beamforming processing to obtain L spatial spectra Y l (θ), determine Y l The peak point direction angle θ of (θ) l , the spatial spectrum center angle is θ0, the cyclic shift is Y l (θ), so that its peak point points to θ0, eliminating the linear delay component in the single-element signal; is the strong line spectrum of the array signal;

[0013] Step 6: Normalize the strong line spectrum to estimate the time delay gradient, using Estimate the secondary and higher-order delay gradients Δτ of adjacent array elements i ; It is the spectrum of the array signal strong line spectrum after removing the linear time delay;

[0014] Step 7: Δτ i Integrate to get the delay estimate τ of each array element relative to the reference array element m , then for the array signal x i (t) Perform estimated delay compensation and repeat steps 5 to 7 until the estimation converges;

[0015] Step 8: After compensating for the secondary and higher-order delays of each array signal, conventional beamforming is used to obtain the target tracking beam. The line spectrum features with enhanced fidelity are obtained through Fourier transform.

[0016] Furthermore, the step 2 specifically includes the following steps:

[0017] Step 2.1: Calculate the ideal uniform linear array at the guiding angle φ j Next adjacent element delay τ j :

[0018]

[0019] Where j = 1,…,J+1, J+1 is the total number of guiding angles, d is the distance between adjacent array elements, v is the propagation speed of sound in water; φ j ∈[0,2π];

[0020] Step 2.2, delay and add the array data to obtain the target signal beam energy map B[b(1),…,b(J+1)], where b(j) is the beamforming signal energy of the jth steering angle;

[0021]

[0022] Where i = 1, 2, ..., M, M is the number of array elements in the towed array, and t is the signal time;

[0023] Step 2.3: Find the guidance angle at the maximum beam energy position through energy detection as a rough estimate of the target direction.

[0024] Furthermore, the step 3 specifically includes the following steps:

[0025] Step 3.1: Based on the rough estimate of the target position Calculate the delay estimate of each array element

[0026]

[0027] Where v is the speed of sound in water;

[0028] Step 3.2: Estimate the delay of each array data Align with the reference array element and coherently add the aligned array data to obtain the target tracking beam g(t):

[0029]

[0030] Step 3.3, perform Fourier transform on g(t) to obtain the target signal spectrum G(f);

[0031] Step 3.4: Perform background equalization on G(f) to whiten the noise of G(f); add windows to the discrete sampling sequence G(k) of G(f) and sort it in ascending order to form a new sequence Ω(k) = {Y(1), Y(2), ... Y(N)}, and determine the median of the sampling points of Ω(k) Remove all items greater than Points, set there are P less than Calculate the noise mean Y(1)…Y(N) are the windowed G(k) arranged from small to large; N is the window length; r is between 1 and 1.1;

[0032] Step 3.5: The spectrum after equalization is G line (k) = G(k) / μ(k), then G line Only strong line spectrum signals and Gaussian white noise are retained in (k); μ(k) is the estimated noise mean.

[0033] Furthermore, in step 4, after being equalized, the entire noise is Gaussian white noise, and the strong line spectrum is its singular value. The noise amplitude range is concentrated within [-3σ, 3σ], and the strong line spectrum amplitude exceeds 3σ, which specifically includes the following steps:

[0034] Step 4.1: Take 2σ of the mean amplitude μ of the retained white noise as the initial threshold, and calculate the amplitude of each frequency f|G line(f)|Test, if|G line (f)|>u*2σ, then keep f; |G line (f)| is the spectrum amplitude of the signal after equalization;

[0035] Step 4.2: For the retained line spectrum frequency f, take the noise points around the Q point to estimate its local noise mean μ', and take μ'*3σ as the second threshold. If the retained frequency line spectrum amplitude |G line (f)| is greater than μ'*3σ, then the line spectrum is judged to be a strong line spectrum;

[0036] Step 4.3: Use the global low threshold for primary screening and the local high threshold for secondary screening to detect and estimate G line The frequencies of the L strong line spectra in (f)

[0037] Furthermore, the cyclic shift in step 5 specifically includes the following steps:

[0038] Each element in the towed array has a frequency A linear phase shift is applied to The peak of the spatial spectrum moves to the center of the spatial spectrum θ0;

[0039]

[0040] Eliminate the linear delay in the signal, so that each element in the towed array is at the same frequency. There are only secondary and higher-order delay components relative to the reference array element.

[0041] Further, the step 6 comprises:

[0042] Step 6.1 The spectrum signals of two adjacent array elements are conjugate multiplied and normalized, retaining only the secondary and higher-order phase difference characteristics of adjacent array elements. Where conj(·) means the complex number is conjugated; To include The unit complex exponential signal of ;

[0043] Step 6.2 The phase differences of adjacent array elements estimated by L strong line spectra are summed and normalized to obtain the average phase difference of adjacent array elements. To include Unit complex exponential signal;

[0044] Step 6.3 The average phase difference of adjacent array elements Calculate the delay difference Δτ between adjacent array elements i , where phase(·) represents complex phase processing.

[0045] Furthermore, the step 7 specifically includes the following steps:

[0046] The estimated time delay gradient Δτ i Integrate to obtain the secondary and higher-order delay differences of each array element relative to the reference array element Then, delay compensation x is performed on each array element signal. i '(t) = x i (t+τ m ); use x i '(t) replaces x i (t), repeat iterative steps 5-7 until the estimation reaches convergence.

[0047] Further, the step 8 comprises:

[0048] Compensate each element signal x based on the estimated delay difference i '(t), and then the target direction estimation is obtained through beamforming technology Get target tracking beam

[0049]

[0050]

[0051] Tracking beam on target Perform Fourier transform to obtain enhanced line spectrum features.

[0052] A line spectrum feature enhancement system based on phase gradient error estimation technology, characterized in that it includes an observation array signal acquisition module, a target azimuth rough estimation module, a spectrum equalization module, a line spectrum detection module, a cyclic shift module, a secondary and higher-order delay gradient estimation module, a delay compensation module, and a line spectrum feature enhancement module in sequence;

[0053] The observation array signal acquisition module is used to acquire the observation array signal x i (t), i = 1, 2, ..., M, M is the number of array elements in the tow array;

[0054] Target position rough estimation module, used to roughly estimate the target position is the guidance angle when the target signal beam energy is maximum;

[0055] The spectrum equalization module is used to whiten the colored noise of the input tracking beam g(t) spectrum G(f) and remove the continuous spectrum trend in the spectrum to obtain G line (f);

[0056] Line spectrum detection module, used to detect the spectrum G after whitening noiseline (f) L strong line spectra are accurately detected L is the number of estimated line spectra;

[0057] The circular shift module is used to The peak of the spatial spectrum is moved to the center position θ0 of the spatial spectrum, eliminating the linear delay of the array;

[0058] The secondary and higher-order delay gradient estimation module is used to estimate the delay gradient Δτ between two adjacent array elements. i ;

[0059] The delay compensation module is used to integrate the secondary and higher delay gradients to obtain the delay estimate τ of each array element relative to the reference array element caused by array distortion. m , and iteratively compensate;

[0060] Line spectrum feature enhancement module, used to obtain target tracking beam after compensating for secondary and higher-order delays The line spectrum features with enhanced fidelity are obtained through Fourier transform.

[0061] The line spectrum feature enhancement method and system based on the phase gradient error estimation technology of the present invention have the following advantages:

[0062] The present invention utilizes the strong line spectrum components in the target radiation noise signal to estimate the secondary and higher-order delays of the array signal. The method estimates the secondary and higher-order delays through iterative calculations until they converge. The application is simple and direct, with low economic cost and obvious effect, small amount of calculation and high accuracy. The method only needs to utilize the strong line spectrum components in the target radiation noise signal, and does not need to make any assumptions about the distortion shape of the towed array, so it can achieve line spectrum feature enhancement under any array conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 A schematic diagram of array element arrangement of a distorted towed array and an ideal linear array according to an embodiment of the present invention;

[0064] Figure 2 It is a flow chart of a method for enhancing line spectrum features based on phase gradient error estimation technology according to an embodiment of the present invention;

[0065] Figure 3 It is the beam energy diagram based on the ideal array of the present invention;

[0066] Figure 4 This is the part before the strong line spectrum phase gradient error estimation of the present invention. Spatial spectrum;

[0067] Figure 5 This is the part after the strong line spectrum phase gradient error estimation of the present invention Spatial spectrum;

[0068] Figure 6 It is an ideal array beam energy diagram after strong line spectrum phase gradient error estimation of the present invention;

[0069] Figure 7 A comparison diagram of the tracking beam spectrum based on the ideal array of the present invention and the tracking target spectrum of the method of the present invention;

[0070] Figure 8 It is a composition diagram of a line spectrum feature enhancement system based on phase gradient error estimation technology according to an embodiment of the present invention;

[0071] Fig. 9 A comparison diagram of relative errors of line spectrum amplitudes according to an embodiment of the present invention;

[0072] Explanation of the marks in the figure: 1. Observation array signal acquisition module; 2. Target azimuth coarse estimation module; 3. Spectrum equalization module; 4. Line spectrum detection module; 5. Cyclic shift module; 6. Secondary and higher-order delay gradient estimation module; 7. Delay compensation module; 8. Line spectrum feature enhancement module. DETAILED DESCRIPTION

[0073] In order to better understand the purpose, structure and function of the present invention, the following is a further detailed description of a line spectrum feature enhancement method and system based on phase gradient error estimation technology of the present invention in conjunction with the accompanying drawings.

[0074] Embodiment 1:

[0075] The present invention discloses a line spectrum feature enhancement method and system based on phase gradient error estimation technology. The method can correct the distortion of a towed linear array and the influence of the delay vector field value on beamforming under low signal-to-noise ratio conditions, and realize line spectrum feature enhancement of the distorted towed array in an underwater acoustic environment. In order to verify the effectiveness of the method, this embodiment is based on Figure 1 Take the distorted drag array shown in the figure as an example to obtain the line spectrum characteristics of the tracked target. In an ideal array, the array elements are arranged in straight lines with equal spacing. Figure 1 As shown, the array element arrangement of the distorted array in this embodiment is shown as the hollow circle in the figure, including 64 array elements, that is, M = 64, and the array element spacing has a slight difference. In this embodiment, the spacing difference is ignored and it is assumed that the array element spacing is approximately equal, which is: d = 1.5 meters. The first array element on the left is used as the reference array element, and a coordinate system is established with the position of the reference array element as the origin. It can be seen from the figure that the arrangement of each array element is quite different from the ideal array.

[0076] The specific steps of the line spectrum feature enhancement method based on the phase gradient error estimation technology disclosed in this embodiment are as follows: Figure 2 As shown, including:

[0077] Step 1: Get the observation array signal x i(t), i = 1, 2, ..., M, M is the number of array elements in the tow array;

[0078] Step 2: Roughly estimate the target direction based on ideal beamforming is the guidance angle when the target signal beam energy is maximum;

[0079] Step 2.1: Calculate the ideal uniform linear array at the guiding angle φ j Next adjacent element delay τ j :

[0080]

[0081] Where j = 1,…,J+1, J+1 is the total number of guiding angles, d is the distance between adjacent array elements, v is the propagation speed of sound in water; φ j ∈[0,2π], J=360;

[0082] Step 2.2: Delay the addition of each array data to obtain the target signal beam energy diagram B[b(1),…,b(J+1)], where b(j) is the signal energy of the jth steering angle.

[0083]

[0084] Where i = 1, 2, ..., M, M is the number of array elements in the towed array, and t is the signal time.

[0085] Step 2.3: Find the guidance angle at the maximum beam energy position through energy detection as a rough estimate of the target direction. Conventional beamforming based on ideal arrays is as follows Figure 3 As shown in the figure, the maximum beam energy position is found through energy detection, and a rough estimate of the target direction is obtained.

[0086] Step 3: Perform background equalization on the target tracking beam g(t) spectrum G(f), whiten the broadband non-white noise, and remove the continuous spectrum.

[0087] Step 3.1: Based on the rough estimate of the target position Calculate the delay estimate of each array element

[0088]

[0089] Where v is the speed of sound in water, and d is the distance between adjacent array elements;

[0090] Step 3.2: Estimate the delay of each array data Align with the reference array element and coherently add the aligned array data to obtain the target tracking beam g(t):

[0091]

[0092] Step 3.3, perform Fourier transform on g(t) to obtain the target signal spectrum G(f);

[0093] Step 3.4: Background equalization G(f) is performed to whiten the noise of G(f). The discrete sampling sequence G(k) of G(f) is windowed and sorted in ascending order to form a new sequence Ω(k) = {Y(1), Y(2), ... Y(N)}, and the median of the sampling points of Ω(k) is determined. Remove all items greater than Assume that there are P points smaller than Calculate the noise mean Y(1)...Y(N) are the windowed G(k) arranged from small to large; N is the window length; r is between 1 and 1.1. In this embodiment, N is 33 and r is 1.1.

[0094] Step 3.5: The spectrum after equalization is G line (k) = G(k) / μ(k), at this time G line (k) Only strong line spectrum signals and Gaussian white noise are retained.

[0095] Step 4: equalize the spectrum G line (f) Perform line spectrum detection to detect the L strong line spectrum frequencies of the target signal whose equalized spectrum amplitude is greater than the local 3σ L is the number of estimated line spectra; σ is G line (f) surrounding noise variance;

[0096] Step 4.1: Take 2σ of the mean amplitude μ of the retained white noise as the initial threshold, and calculate the amplitude of each frequency f|G line (f)|Test, if|G line (f)|>u*2σ, then f is retained.

[0097] Step 4.2: For the retained frequency line spectrum f, take the noise points around the Q point to estimate its local noise mean μ', and take μ'*3σ as the second threshold. If the amplitude of the retained frequency line spectrum |G line (f)| is greater than μ'*3σ, the line spectrum is determined to be a strong line spectrum. In this embodiment, Q is 20.

[0098] Step 4.3: Use the global low threshold for primary screening and the local high threshold for secondary screening to detect and estimate G line The frequencies of the L strong line spectra in (f) In this embodiment, L is 7.

[0099] Step 5: Perform Fourier transform on the time domain waveform of each array element to obtain X i (f), and for the L detected Perform beamforming processing to obtain L spatial spectra Y l (θ), determine Y l The peak point direction angle θ of (θ) l , the spatial spectrum center angle is θ0, the cyclic shift is Y l (θ), so that its peak point points to θ0, eliminating the linear delay component in the single-element signal; is the strong line spectrum of the array signal.

[0100] The cyclic shift specifically includes the following steps:

[0101] Each element in the towed array has a frequency A linear phase shift is applied to The peak of the spatial spectrum moves to the center of the spatial spectrum θ0.

[0102]

[0103] This can eliminate the linear delay of the array signal and make each element in the towed array There are only secondary and higher order delays relative to the reference array element. θ0 of this embodiment = 90°.

[0104] Step 6: Normalize the strong line spectrum to estimate the time delay gradient, using Estimate the secondary and higher-order delay gradients Δτ of adjacent array elements i ; It is the spectrum of the strong line spectrum of the array signal after removing the linear delay.

[0105] Step 6.1 The spectrum signals of two adjacent array elements are conjugate multiplied and normalized, retaining only the secondary and higher-order phase difference characteristics of adjacent array elements. Where conj(·) represents the conjugate treatment of complex numbers. To include A unit complex exponential signal.

[0106] Step 6.2 The phase differences of adjacent array elements estimated by L strong line spectra are summed and normalized to obtain the average phase difference of adjacent array elements. To include A unit complex exponential signal.

[0107] Step 6.3 Where phase(·) represents complex phase processing, which is the average phase difference of adjacent array elements. Calculate the delay difference Δτ between adjacent array elements i .

[0108] Step 7: Iterate delay compensation

[0109] The estimated time delay gradient Δτ i Integrate to obtain the secondary and higher-order delay differences of each array element relative to the reference array element Then, the delay compensation is performed on each array element signal.

[0110] x i '(t) = x i (t+τ m )

[0111] Use x i '(t) replaces x i (t), repeat iterative steps 5, 6, and 7 until the estimation converges. The convergence condition of this embodiment is the mean square error RMS (τ m )<0.0001.

[0112] Figure 4 and Figure 5 Strong line spectrum Comparison of conventional frequency beamforming and frequency beamforming after processing by the method of the present invention. By comparing the two figures, it can be clearly seen that after adopting the method of the present invention, the spatial spectra of each chaotic strong line spectrum are ordered, the main lobe broadening is eliminated, and the peak points point to the same spatial angle, indicating that they come from the same target. Figure 6 The new conventional beamforming spatial spectrum after self-focusing is completed is given. Figure 6 and Figure 3 By comparison, it can be seen Figure 6 The main lobe does not widen, and the peak energy is more concentrated. The above shows that the method proposed by the present invention can effectively compensate for the high-order delay of the array signal and eliminate the influence of the towed array shape distortion on the beam focusing.

[0113] Step 8: The array element signal x that compensates for the secondary and higher-order delay differences i '(t), beamforming again to get the target direction Get target tracking beam

[0114]

[0115]

[0116] Tracking beam on target Perform Fourier transform to obtain enhanced line spectrum features.

[0117] Figure 7The comparison diagrams of the original data spectrum after FFT using conventional beam forming (CBF) and the tracking target spectrum obtained using the method disclosed in the present invention are given. It can be seen from the figure that compared with the conventional beam forming method, the line spectrum feature enhancement method disclosed in the present invention forms a beam target line spectrum gain that is higher, and the beam forming effect is enhanced by the line spectrum feature.

[0118] This embodiment also discloses a system for implementing the above-mentioned line spectrum feature enhancement system based on phase gradient error estimation technology, such as Figure 8 As shown, it includes an observation array signal acquisition module 1, a target azimuth rough estimation module 2, a spectrum equalization module 3, a line spectrum detection module 4, a cyclic shift module 5, a secondary and higher-order delay gradient estimation module 6, a delay compensation module 7 and a line spectrum feature enhancement module 8:

[0119] The observation array signal acquisition module 1 is used to acquire the observation array signal x i (t), i = 1, 2, ..., M, M is the number of array elements in the tow array;

[0120] The target position rough estimation module 2 is used to roughly estimate the target position is the guidance angle when the target signal beam energy is maximum;

[0121] The spectrum equalization module 3 is used to whiten the colored noise of the input tracking beam g(t) spectrum G(f) to remove the continuous spectrum trend in the spectrum and obtain G line (f).

[0122] The line spectrum detection module 4 is used to detect the spectrum G after whitening noise. line (f) L strong line spectra are accurately detected L is the number of estimated line spectra;

[0123] The cyclic shift module 5 is used to The peak of the spatial spectrum is moved to the center position θ0 of the spatial spectrum, eliminating the linear delay of the array;

[0124] The secondary and higher-order delay gradient estimation module 6 is used to estimate the delay gradient Δτ between two adjacent array elements. i ;

[0125] The delay compensation module 7 is used to integrate the secondary and higher delay gradients to obtain the delay estimate τ of each array element relative to the reference array element caused by array distortion. m , and iteratively compensate;

[0126] The line spectrum feature enhancement module 8 is used to obtain the target tracking beam after compensating the secondary and higher delays. The line spectrum features with enhanced fidelity are obtained through Fourier transform.

[0127] To verify the effect of the method disclosed in the present invention under low signal-to-noise ratio conditions, the signal-to-noise ratio (SNR) was changed to -24dB. <SNR<-10dB。

[0128] The relative error of the estimated beamforming line spectrum amplitude is E, A l Represents the amplitude of the original data spectrum at the lth line spectrum position, PA l It represents the amplitude of the estimated beamforming spectrum at the position of the lth line spectrum, and L represents the number of selected line spectrums. The relative error of the tracking beam line spectrum amplitude is used as the performance evaluation index.

[0129] The relative errors of the line spectrum amplitudes of the conventional beamforming and phase array correction methods are compared under different signal-to-noise ratios (SNRs). The results are as follows: Fig. 9 As shown in the figure:

[0130] 1. The method of the present invention is highly robust and still has relatively superior performance in a low signal-to-noise ratio environment.

[0131] 2. As the signal-to-noise ratio increases, the error of the method disclosed in the present invention gradually decreases and the line spectrum characteristics are enhanced; however, the conventional beamforming method has no effective formation correction capability, and as the signal-to-noise ratio increases, its reconstruction error changes very little, and the line spectrum characteristics are not enhanced.

[0132] It is to be understood that the present invention is described by some embodiments, and it is known to those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the scope of protection of the present invention.

Claims

1. A line spectrum feature enhancement method based on phase gradient error estimation technology, characterized in that: The specific steps include: Step 1: Get the observation array signal x i (t), i = 1, 2, ..., M, M is the number of array elements in the tow array; Step 2: Estimate target direction based on ideal array beamforming is the guidance angle when the target signal beam energy is maximum; Step 3: Perform beamforming processing based on the estimated target orientation to obtain the tracking beam g(t) and spectrum G(f) in the target direction, estimate the broadband continuous spectrum based on the data, perform background equalization processing on the tracking beam g(t) and spectrum G(f), whiten the non-white broadband continuous spectrum, and remove the continuous spectrum; Step 4: equalize the spectrum G line (f) Perform line spectrum detection to detect the L strong line spectrum frequencies of the target signal whose equalized spectrum amplitude is greater than the local 3σ L is the number of estimated line spectra; σ is G line (f) surrounding noise variance; Step 5: Perform Fourier transform on the time domain waveform of each array element to obtain X i (f), and for the L detected Perform beamforming processing to obtain L spatial spectra Y l (θ), determine Y l The peak point direction angle θ of (θ) l , the spatial spectrum center angle is θ0, the cyclic shift is Y l (θ), so that its peak point points to θ0, eliminating the linear delay component in the single-element signal; is the strong line spectrum of the array signal; Step 6: Normalize the strong line spectrum to estimate the time delay gradient, using Estimate the secondary and higher-order delay gradients Δτ of adjacent array elements i ; It is the spectrum of the array signal strong line spectrum after removing the linear time delay; Step 7: Δτ i Integrate to get the delay estimate τ of each array element relative to the reference array element m , then for the array signal x i (t) Perform estimated delay compensation and repeat steps 5 to 7 until the estimation converges; Step 8: After compensating for the secondary and higher-order delays of each array signal, conventional beamforming is used to obtain the target tracking beam. The line spectrum features with enhanced fidelity are obtained through Fourier transform.

2. The line spectrum feature enhancement method based on phase gradient error estimation technology according to claim 1 is characterized in that: The step 2 specifically includes the following steps: Step 2.1: Calculate the ideal uniform linear array at the guiding angle φ j Next adjacent element delay τ j : Where j = 1,…,J+1, J+1 is the total number of guiding angles, d is the distance between adjacent array elements, v is the propagation speed of sound in water; φ j ∈[0,2π]; Step 2.2, delay and add the array data to obtain the target signal beam energy map B[b(1),…,b(J+1)], where b(j) is the beamforming signal energy of the jth steering angle; Where i = 1, 2, ..., M, M is the number of array elements in the towed array, and t is the signal time; Step 2.3: Find the guidance angle at the maximum beam energy position through energy detection as a rough estimate of the target direction.

3. The line spectrum feature enhancement method based on phase gradient error estimation technology according to claim 2 is characterized in that: The step 3 specifically comprises the following steps: Step 3.1: Based on the rough estimate of the target position Calculate the delay estimate of each array element Where v is the speed of sound in water; Step 3.2: Estimate the delay of each array data Align with the reference array element, and coherently add the aligned array data to obtain the target tracking beam g(t): Step 3.3, perform Fourier transform on g(t) to obtain the target signal spectrum G(f); Step 3.4: Perform background equalization on G(f) to whiten the noise of G(f); perform windowing on the discrete sampling sequence G(k) of G(f) and sort it in ascending order to form a new sequence Ω(k) = {Y(1), Y(2), ... Y(N)}, and determine the median of the sampling points of Ω(k) Remove all items greater than Points, set there are P less than Calculate the noise mean Y(1)…Y(N) are the windowed G(k) arranged from small to large; N is the window length; r is between 1 and 1.1; Step 3.5: The spectrum after equalization is G line (k) = G(k) / μ(k), then G line Only strong line spectrum signals and Gaussian white noise are retained in (k); μ(k) is the estimated noise mean.

4. The line spectrum feature enhancement method based on phase gradient error estimation technology according to claim 3 is characterized in that: In step 4, after being equalized, the entire noise is Gaussian white noise, and the strong line spectrum is its singular value. The noise amplitude range is concentrated within [-3σ, 3σ], and the strong line spectrum amplitude exceeds 3σ. Specifically, the following steps are included: Step 4.1: Take 2σ of the mean amplitude μ of the retained white noise as the initial threshold, and calculate the amplitude of each frequency f|G line (f)|Test, if|G line (f)|>u*2σ, then keep f; |G line (f)| is the spectrum amplitude of the signal after equalization; Step 4.2: For the retained line spectrum frequency f, take the noise points around the Q point to estimate its local noise mean μ', and take μ'*3σ as the second threshold. If the retained frequency line spectrum amplitude |G line (f)| is greater than μ'*3σ, then the line spectrum is judged to be a strong line spectrum; Step 4.3: Use the global low threshold for primary screening and the local high threshold for secondary screening to detect and estimate G line The frequencies of the L strong line spectra in (f) 5. The line spectrum feature enhancement method based on phase gradient error estimation technology according to claim 4 is characterized in that: The cyclic shift in step 5 specifically includes the following steps: Each element in the towed array has a frequency A linear phase shift is applied to The peak of the spatial spectrum moves to the center of the spatial spectrum θ0; Eliminate the linear delay in the signal, so that each element in the towed array is at the same frequency. There are only secondary and higher-order delay components relative to the reference array element.

6. The line spectrum feature enhancement method based on phase gradient error estimation technology according to claim 5 is characterized in that: The step 6 comprises: Step 6.1 The spectrum signals of two adjacent array elements are conjugate multiplied and normalized, retaining only the secondary and higher-order phase difference characteristics of adjacent array elements. Where conj(·) means the complex number is conjugated; To include The unit complex exponential signal of ; Step 6.2 The phase differences of adjacent array elements estimated by L strong line spectra are summed and normalized to obtain the average phase difference of adjacent array elements. To include Unit complex exponential signal; Step 6.3 The average phase difference of adjacent array elements Calculate the delay difference Δτ between adjacent array elements i , where phase(·) represents complex phase processing.

7. The line spectrum feature enhancement method based on phase gradient error estimation technology according to claim 6 is characterized in that: The step 7 specifically includes the following steps: The estimated time delay gradient Δτ i Integrate to obtain the secondary and higher-order delay differences of each array element relative to the reference array element Then, delay compensation x is performed on each array element signal. i '(t) = x i (t+τ m ); use x i '(t) replaces x i (t), repeat iterative steps 5-7 until the estimation reaches convergence.

8. The line spectrum feature enhancement method based on phase gradient error estimation technology according to claim 7 is characterized in that: The step 8 comprises: Compensate each element signal x based on the estimated delay difference i '(t), and then the target direction estimation is obtained through beamforming technology Get target tracking beam Tracking beam on target Perform Fourier transform to obtain enhanced line spectrum features.

9. The line spectrum feature enhancement system of the line spectrum feature enhancement method based on phase gradient error estimation technology according to any one of claims 1 to 8, characterized in that: It includes an observation array signal acquisition module, a target azimuth rough estimation module, a spectrum equalization module, a line spectrum detection module, a cyclic shift module, a secondary and higher-order delay gradient estimation module, a delay compensation module, and a line spectrum feature enhancement module in sequence; The observation array signal acquisition module is used to acquire the observation array signal x i (t), i = 1, 2, ..., M, M is the number of array elements in the tow array; Target position rough estimation module, used to roughly estimate the target position is the guidance angle when the target signal beam energy is maximum; The spectrum equalization module is used to whiten the colored noise of the input tracking beam g(t) spectrum G(f) and remove the continuous spectrum trend in the spectrum to obtain G line (f); Line spectrum detection module, used to detect the spectrum G after whitening noise line (f) L strong line spectra are accurately detected L is the number of estimated line spectra; The circular shift module is used to The peak of the spatial spectrum is moved to the center position θ0 of the spatial spectrum, eliminating the linear delay of the array; The secondary and higher-order delay gradient estimation module is used to estimate the delay gradient Δτ between two adjacent array elements. i ; The delay compensation module is used to integrate the secondary and higher delay gradients to obtain the delay estimate τ of each array element relative to the reference array element caused by array distortion. m , and iteratively compensate; Line spectrum feature enhancement module, used to obtain target tracking beam after compensating for secondary and higher-order delays The line spectrum features with enhanced fidelity are obtained through Fourier transform.

Images