Towed array beam forming passive direction finding method and storage medium

Abstract

The invention discloses a towed array beam forming passive direction finding method and a storage medium, and the method comprises the steps: carrying out the frequency mixing processing of a received signal, combining with discrete Fourier transform, fusing the frequency domain information of a higher frequency on the basis of not changing the algorithm complexity, decomposing a broadband signal into a plurality of frequency point signals, and carrying out the fusion of the frequency domain information of the higher frequency; only each frequency point in a signal frequency band range is processed, and compared with a minimum variance distortionless response (MVDR) beam forming method, the method is high in processing speed, high in real-time performance and high in algorithm robustness. Compared with a traditional beam forming method, the method has the advantages that the azimuth resolution is remarkably improved, and meanwhile, the method is superior to an MVDR method in the aspects of operation speed and robustness, so that the method is more suitable for real-time processing application.

Description

Technical Field

[0001] The present invention relates to the field of underwater acoustic direction finding, and relates to a beamforming passive direction finding method, specifically to a towed array beamforming passive direction finding method based on mixing filtering. Background Art

[0002] Compared with traditional hull-mounted sonars, a towed line array, on the one hand, is not restricted by the geometric size of the hull and can deploy more hydrophones, thereby increasing the sonar aperture and having the ability to receive low-frequency or even very low-frequency sound waves; on the other hand, since the towed line array is far from the mother ship, the influence of platform noise is reduced, and the received signal-to-noise ratio is significantly improved. These characteristics greatly enhance the detection ability of towed sonars.

[0003] In conventional beamforming, the output signal-to-noise ratio, azimuth resolution, and sonar operating range increase with the increase of the array length. However, in practice, due to the limitations of platform mobility and other various factors, the designed length of the linear array cannot be too long, which to a certain extent limits the azimuth resolution and operating range of the sonar.

[0004] CN102226837A discloses a vector circular array acoustic pressure and particle velocity combined direction finding method applicable to cylindrical baffle conditions, which generates a cross-covariance matrix using the acoustic pressure array element domain signal and the particle velocity array element domain signal, and performs azimuth estimation based on the cross-covariance matrix to finally obtain the azimuth result. However, in the process of solving the cross-covariance matrix using the acoustic pressure phase mode domain signal and the particle velocity phase mode domain signal, matrix singularity is likely to occur, resulting in algorithm instability and the appearance of non-numeric results.

[0005] In the minimum variance distortionless response (MVDR) beamforming technology, due to the influence of matrix singularity problems, the calculation speed is reduced and the result is unstable, so real-time target azimuth estimation cannot be achieved. Generally speaking, the MVDR algorithm has certain performance advantages compared with traditional beamforming algorithms, but its calculation speed is slow and the feasibility of practical applications is low. Therefore, there is an urgent need for a real-time processing method that can perform beamforming on broadband target signals and has high resolution. Summary of the Invention

[0006] The purpose of the present invention is to overcome the deficiencies of the existing MVDR algorithm such as slow calculation speed and unstable calculation results, and provide a towed array beamforming passive direction finding method based on mixing filtering. This method forms a new passive direction finding method by performing mixing processing on the received signal and combining discrete Fourier transform. Compared with traditional beamforming methods, this method has a significant improvement in azimuth resolution, and is also superior to the MVDR beamforming method in terms of operation speed and robustness, so it is more suitable for real-time processing applications.

[0007] The technical solution of the present invention is as follows:

[0008] A passive direction finding method for a towed array beamforming based on mixing filtering, which combines conventional beamforming with frequency shifting (mixing), can be used for azimuth estimation of underwater targets. The method includes the following steps:

[0009] (1) Obtain the received array signal x m (t), m = 1, 2,..., M, where M is the number of array elements in the towed array;

[0010] (2) Perform a fast Fourier transform on x m (t) in the time domain to determine the frequency range [FMIN, FMAX] of the received signal;

[0011] (3) Construct the mixing factor A = e (1j*2*π*F1*t)

[0012] where t is the time series, and F1 is the mixing frequency, FMIN < F1 < FMAX;

[0013] (4) Multiply x m (t) by the mixing factor A to obtain the mixed signal X1 m (t)

[0014] X1 m (t) = x m (t) * A

[0015] (5) Perform a discrete Fourier transform on the mixed signal X1 m (t) to obtain the frequency domain signal F m (f), and select the signal component fk of each array element in the frequency domain signal X1 m (t) within [FMIN + F1, F1 + FMAX];

[0016] (6) Construct the spatial discrete Fourier transform rotation factor wc1:

[0017] v0 = [cos(theta); sin(theta)]

[0018] p0 = [0:M - 1; zeros(1, M)]′

[0019] wc1 = 1 / N * e (-1i*2*pi*d / c*p0*v0*fk)

[0020] where theta takes values from -90° to 90° at intervals of θ°, d is the element spacing, c is the sound speed in water; zeros represents generating a matrix with all elements being 0, zeros(1, M) represents generating a 1×M - dimensional matrix with all elements in the matrix being 0;

[0021] (7) Spatial domain signal F1 m (f) = sum(F m (f) * wc1), where sum represents the summation operation; where sum represents the summation operation, F1 m (f) is equal to F m (f) and the sum of each corresponding frequency after multiplying by wc1;

[0022] (8) Perform zero-padding operation on the spatial domain signal F1 m (f) to obtain the zero-padded spatial domain signal F2 m (f);

[0023] (9) Perform inverse Fourier transform on the F2 m (f) signal to obtain the beam output result, and the maximum value of the beam output corresponds to the angle of azimuth estimation.

[0024] Furthermore, the specific operation of performing zero-padding on the spatial domain signal F1 m (f) is as follows: Zero-pad the data of F1 m (f) before the frequency F1 + FMIN, and zero-pad the data after F1 + FMAX.

[0025] A computer-readable storage medium stores a computer program thereon, and the computer program can be executed by a processor to implement the steps of a passive direction finding method for a towed array beamforming according to the present invention.

[0026] Advantages of the present invention:

[0027] (1) By performing frequency domain processing on the received signal, higher frequency domain information is fused on the basis of not changing the algorithm complexity, which can be used for azimuth estimation of underwater targets, has higher azimuth resolution than the conventional beamforming method, and also has better detection ability for weak targets;

[0028] (2) Decompose the broadband signal into multiple frequency point signals, and only process each frequency point within the signal frequency band, which is faster in processing speed, stronger in real-time performance, and higher in algorithm robustness than the MVDR beamforming;

[0029] (3) Measured data prove that the method of the present invention has higher azimuth resolution compared to CBF, faster operation speed and better robustness compared to MVDR. Description of the Drawings

[0030] Figure 1 is the algorithm flow chart of the present invention.

[0031] Figure 2 is the spectrogram of the received signal.

[0032] Figure 3 It is the spectrogram of the target frequency band of the received signal.

[0033] Figure 4 They are the beam patterns of conventional beamforming, MVDR, mixed-frequency conventional beamforming, and mixed-frequency MVDR.

[0034] Figure 5 They are the time history diagrams before and after mixing of the measured data. Detailed implementation manners

[0035] The present invention will be further described below with reference to the accompanying drawings.

[0036] Embodiment 1

[0037] A passive direction finding method for a towed array beamforming based on mixing filtering, which forms a new passive direction finding technology by performing mixing processing on the received signal and combining with the discrete Fourier transform. Compared with the traditional beamforming method, this method has a significant improvement in azimuth resolution, and is also superior to the minimum variance distortionless response (MVDR) beamforming method in terms of operation speed and robustness, and thus is more suitable for real-time processing applications.

[0038] In this example, the towed array has 72 array elements (i.e., M = 72), the element spacing is 0.15 meters, and the sampling rate is 50 kHz.

[0039] As Figure 1 shown, the specific steps of the method for enhancing the spectral line characteristics of the distorted towed array in this embodiment include:

[0040] Step 1: Obtain the received array signal x m (t), m = 1, 2,..., M, where M is the number of array elements in the towed array;

[0041] Step 2: Perform a fast Fourier transform on the received signal:

[0042]

[0043] As Figure 2 , Figure 3 shown, determine that the approximate frequency band range of the target is: 0 - 5 kHz;

[0044] Step 3: Determine the value of the mixing factor according to the frequency range. In this example, the mixing frequency is 2 kHz, and its value is within the selected frequency band range. Construct the mixing factor from the frequency:

[0045] A = e (1j*2*π*F1*t) ;

[0046] Step 4: Multiply the received signal x m(t) is multiplied by the mixing factor to obtain the mixed signal X1 of the received signal m (t);

[0047] Step 5: Perform a discrete Fourier transform on the mixed signal to obtain the frequency-domain signal F m (f), and select the frequency-domain signals of each array element according to the frequency band: select the signal components with frequencies in the range of [2000 Hz, 7000 Hz];

[0048] Step 6: Construct the spatial discrete Fourier transform rotation factor wc1:

[0049] v0 = [cos(theta); sin(theta)]

[0050] p0 = [0:M - 1; zeros(1, M)]′

[0051] wc1 = 1 / N * e (-1i*2*pi*d / c*p0*v0*fk)

[0052] where theta takes values at intervals of 1° from -90° to 90°, M is the number of array elements, which is 72, fk takes values within the frequency range of [2000 Hz, 7000 Hz], d is the array element spacing of 0.15 meters, c is the sound speed in water of 1500 m / s; zeros represents generating a matrix with all elements being 0, zeros(1, 72) represents generating a 1×72-dimensional matrix with all elements in the matrix being 0;

[0053] Step 7: Multiply and sum F m (f) with *wc1 to obtain the spatial F1 m (f) signal,

[0054] F1 m (f) = sum(F m (f) * wc1);

[0055] Step 8: Perform zero-padding on F1 m (f), specifically, pad zeros to the data of F1 m (f) before the frequency of 2000 Hz and after the frequency of 7000 Hz to obtain the zero-padded signal F2 m (f);

[0056] Step 9: Perform an inverse Fourier transform on the F2 m (f) signal and sum the inverse transforms at this angle. Finally, obtain the beam output result at each angle, and the maximum value of the beam output is the azimuth estimation angle.

[0057] According to Figure 4From the azimuth estimation results of the target by different algorithms shown, the mixed-frequency MVDR algorithm has higher azimuth estimation resolution, specifically manifested as a narrower main lobe width, and the mixed-frequency MVDR algorithm has better sidelobe suppression ability, specifically manifested as the normalized amplitude of the sidelobe being smaller than that of other algorithms.

[0058] As Figure 4 shown, even when using the same azimuth estimation algorithm, the resolution of azimuth estimation is significantly improved after mixed-frequency processing. After mixed-frequency processing, the main lobe width (under -3dB condition) of the CBF beamforming algorithm is narrowed by 6°, and the sidelobe amplitude is reduced by nearly 5dB; after mixed-frequency processing, the main lobe width (under -3dB condition) of the MVDR beamforming algorithm is narrowed by 1°, and the sidelobe amplitude is reduced by nearly 3dB.

[0059] According to Figure 5 it can more clearly show the advantages of the algorithm of the present invention. Comparing the time-angle history diagrams, the trajectory of the target of the algorithm before mixed-frequency is wider, with low resolution, and there is interference from clutter around the target trajectory. The trajectory of the target of the algorithm after mixed-frequency is clear, with high resolution, and the surrounding clutter signals are suppressed, obtaining a relatively clear and clean time-angle history diagram.

Claims

1. A passive direction finding method for a towed array beamforming, characterized in that, It includes the following steps: (1) Obtain the received array signal x m (t), where m = 1, 2,..., M and M is the number of array elements in the towed array; (2) For x m (t) Perform a fast Fourier transform in the time domain to determine the frequency range [FMIN, FMAX] of the received signal; (3) Construct the mixing factor A = e (1j*2*π*F1*t) , where t is the time series, F1 is the mixing frequency, and FMIN < F1 < FMAX; (4) Calculate the mixed-frequency signal X1 m (t) = x m (t) * A; (5) Discretely Fourier transform X1 m (t) to obtain the frequency-domain signal F m (f), and select the signal component fk of X1 m (t) of each array element within [FMIN + F1, F1 + FMAX]; (6) Construct the spatial domain discrete Fourier transform rotation factor wc1 v0 = [cos(theta); sin(theta)] p0 = [0:M-1; zeros(1,M)]' wc1 = 1 / N * e (-1i*2*pi*d / c*p0*v0*fk) where theta takes values from -90° to 90° at intervals of θ°, d is the element spacing, and c is the sound speed in water; zeros represents generating a matrix with all elements being 0, and zeros(1,M) means generating a 1×M-dimensional matrix with all elements in the matrix being 0; (7) Airspace signal F1 m f) = sum(F m (f) * wc1), where sum represents the summation operation; (8)Perform zero-padding operation on the airspace signal F1 m (f) to obtain the zero-padded airspace signal F2 m (f); (9) Apply F2 m (f) Perform an inverse Fourier transform on the signal to obtain the beam output result. The angle corresponding to the azimuth of the maximum value of the beam output.

2. The method according to claim 1, wherein The airspace signal F1 m (f) The specific zero-padding operation is as follows: Shift F1 m (f) Pad zeros to the data before the frequency FMIN + F1 and to the data after the frequency FMAX + F1.

3. The method according to claim 1, characterized in that, wherein θ = 1°.

4. The method according to any one of claims 1 to 3, characterized in that, wherein M = 72.

5. The method according to claim 4, wherein wherein d = 0.15 m.

6. The method according to any one of claims 1-3, characterized in that, wherein F1 = 2 kHz.

7. The method according to claim 6, wherein wherein [FMIN,FMAX] is [0 Hz, 5000 Hz].[[]END] 8. A computer-readable storage medium having a computer program stored thereon, characterized in that, This computer program can be executed by a processor to implement the steps of a towed array beamforming passive direction finding method according to any one of claims 1-7.

Images