A method for estimating the depth of a moving target based on a single vector hydrophone

Abstract

The present invention provides a method for estimating the depth of a moving target based on a single vector hydrophone, which uses the vector hydrophone to receive sound pressure and particle velocity signals; constructs a sound pressure-particle velocity cross-spectrum; performs time-domain equally-spaced discretization processing on the received sound pressure and particle velocity signals to construct a discrete signal matrix; obtains the preliminary noise-reduced sound pressure and particle velocity signals; decomposes the signals through EWT, estimates the signal frequency according to the LOFAR spectrum analysis, determines the mode function component containing the frequency, and reconstructs the signals; constructs a sound pressure-particle velocity sound field spatial interference spectrum; calculates the horizontal complex sound intensity and the vertical complex sound intensity to estimate the vertical arrival angle of the target direct wave; performs a discrete Fourier transform on the sound pressure-particle velocity sound field spatial interference spectrum along the curve of the vertical arrival angle of the target direct wave, and the sub-maximum value after removing the maximum value at the vertical symmetry axis corresponds to the target depth estimation value. The present invention realizes the estimation of the depth of a moving target by using the received signals under the influence of noise, and has good practical engineering application capabilities.

Description

Technical Field

[0001] The present invention relates to a method for estimating the depth of a moving target based on a single vector hydrophone, and belongs to the field of underwater acoustic vector signal processing. Background Art

[0002] In the field of underwater acoustics, the passive target detection technology has the advantage of concealment compared with the active detection technology. Compared with the traditional pressure hydrophone, a single vector hydrophone can simultaneously obtain the sound pressure and particle velocity components, which contain the azimuth information of the sound source; the linear combination of different physical quantities can bring a directivity independent of frequency, and to a certain extent, it can better suppress the isotropic noise and obtain additional spatial gain. In addition, compared with the array, the single vector hydrophone has the advantages of simpler system composition and convenient deployment, and has always attracted the attention of scholars all over the world.

[0003] Reference 1 (“Performance metrics for depth-based signal separation using deep vertical line arrays”, The Journal of the Acoustical Society of America 139, 418-425 (2016)) proposed a method for estimating the depth of a moving target based on the interference characteristics of the sound field with the vertical arrival angle. This method needs to rely on the vertical array placed near the seabed to detect the periodic interference structure of the output energy of the narrowband beam in the incoming wave direction to achieve the depth estimation of the target. In addition, this method ignores the factor of the change of the sound speed in water.

[0004] Reference 2 (“Matched beam-intensity processing for a deep vertical line array”, The Journal of the Acoustical Society of America 148, 347-358 (2020)) proposed a method for estimating the depth of a moving sound source by copying the beam energy target. This method uses a vertical array to perform conventional beamforming on the received sound pressure signal to obtain the variation of the beam energy with time and the steering angle, and estimates the depth of the sound source by copying the beam energy and defining a depth ambiguity estimation function. This method requires marine environmental parameters to obtain accurate modeling of the model calculation.

[0005] A vector hydrophone can simultaneously receive the sound pressure and particle velocity components in the sound field, and can obtain multi-dimensional sound field information compared with a scalar hydrophone. In the deep-sea environment, the sound wave excited by a target near the sea surface can propagate to the seabed through the direct wave zone or the reliable acoustic path, and has the characteristics of small propagation loss and stable propagation characteristics. Reference 3 ("Passive broadband source depth estimation in the deep ocean using a single vector sensor", The Journal of the Acoustical Society of America 148, EL88-EL92 (2020)) proposed a method for estimating the depth of a broadband sound source using a single vector hydrophone, and estimated the target depth by using the relationship between the sound source depth, the periodicity of frequency interference, and the target arrival angle. This method ignores the applicability of the method when the received signal is affected by noise interference. Summary of the Invention

[0006] The object of the present invention is to utilize the relationship between the received signals of a single vector hydrophone, construct a sound pressure-velocity spatial interference spectrum by using the relationship between the direct wave and the sea surface reflection wave, and combine the direct wave arrival angle trajectory to estimate the depth of a moving target. At the same time, the influence of environmental noise on this method is considered, and a denoising method is proposed to realize the estimation of the depth of a moving target by using the received signal under the influence of noise, and it has good practical engineering application ability.

[0007] The object of the present invention is achieved as follows: The steps are as follows:

[0008] Step 1: Through a single vector hydrophone, the noisy sound pressure p(r, ω, t) and particle velocity signals v r (r, ω, t), v z (r, ω, t) propagated by a target near the deep-sea surface through the direct wave zone or the reliable acoustic path are received;

[0009] Step 2: Construct a sound pressure-velocity cross-spectrum, make full use of the non-correlation between the signals and noise in the sound pressure-velocity joint processing, and estimate the signal frequency by combining LOFAR spectrum analysis;

[0010] Step 3: Perform time-domain equally spaced discretization processing on the sound pressure and particle velocity received signals to construct a discrete received signal matrix;

[0011] Step 4: Perform random linear coherent accumulation denoising on the discrete sound pressure and particle velocity signals respectively to obtain preliminary denoised sound pressure and particle velocity signals;

[0012] Step 5: Decompose the signal by EWT, estimate the signal frequency according to the LOFAR spectrum analysis, and reconstruct the target signal from the mode function component containing the frequency;

[0013] Step 6: Calculate the spectra of the denoised discrete sound pressure and vibration velocity signals by using the fast Fourier transform, and construct the sound pressure-vibration velocity sound field spatial interference spectrum;

[0014] Step 7: Calculate the horizontal complex sound intensity and the vertical complex sound intensity by using the arrival characteristics of the direct wave of the received signal, and estimate the vertical arrival angle of the target direct wave;

[0015] Step 8: Perform a discrete Fourier transform on the sound pressure-vibration velocity sound field spatial interference spectrum along the curve of the vertical arrival angle of the target direct wave. The depth corresponding to the sub-maximum value after removing the maximum value at the vertical symmetry axis is the estimated value of the target depth.

[0016] The present invention also includes the following structural features:

[0017] 1. Step 3 is specifically as follows:

[0018] For the signal sampling time T, divide the received signal at time intervals of Δt, where T >> Δt. The divided signals are rearranged according to a matrix. The sound pressure and vibration velocity signals in the i-th column can be expressed as:

[0019] p i (sinθ i , ω, t) = p s (sinθ i , ω, t) + n pi

[0020] v ri (sinθ i , ω, t) = p s (sinθ i , ω, t)cosθ i + n vri

[0021] v zi (sinθ i , ω, t) = p s (sinθ i , ω, t)cosθ i + n vri

[0022]

[0023]

[0024]

[0025] where \(t\) represents the sampling time series within \(\Delta t\), and \(t\) i represents the time corresponding to the discrete segmentation point, and \(r\) i represents the horizontal distance between the target and the vector hydrophone at the discrete time, and \(R\) i represents the distance between the target and the vector hydrophone at the discrete time, \(H\) represents the depth of the vector hydrophone, and \(\theta\) i represents the vertical arrival angle.

[0026] According to the virtual source theory, the acoustic pressure signal received by the above single vector hydrophone from the target can be further expressed as:

[0027]

[0028] where \(e\) jωt represents the time factor.

[0029] 2. The specific content of Step Four is as follows:

[0030] The actually acquired signal segment after division can be expressed as \(x\) i (t), where \(i\) represents the serial number of the signal segment after division; considering that the interval \(\Delta t\) is much smaller than the sampling time \(T\), the target signal is approximately unchanged, and the signal segment \(x\) i (t) can be expressed by the following formula:

[0031] \(x\) i (t)= \(x\) is (t)+ \(n\) ix (t)

[0032] where \(x\) is (t) represents the target signal contained in the signal segment, and \(n\) ix (t) represents the noise within this signal segment. \(x\) i (t) contains \(L\) signal cycles, and the \(k\)-th signal cycle can be expressed by the following formula:

[0033] \(x\) ik (t)= \(x\) iks (t)+ \(n\) iks (t), \(k = 1, 2, \cdots, L\)

[0034] Performing \(L\) times of random cumulative averaging on the \(k\)-th cycle signal can be expressed as:

[0035]

[0036] 3. The specific content of Step Five is as follows:

[0037] The signal output by Step Four is decomposed by EWT to obtain several intrinsic mode components, and the corresponding components of the underwater acoustic signal are determined according to the estimated frequency and reconstructed;

[0038] 4. The specific content of Step Six is as follows:

[0039]

[0040] Among them, <> represents calculating the average, and the superscript * represents the complex conjugate. represents the Fourier transform of the reconstructed and denoised discrete sound pressure. represents the Fourier transform of the reconstructed and denoised discrete horizontal particle velocity.

[0041] 5. Step seven is specifically as follows:

[0042] Calculate the direct wave horizontal sound energy flux and vertical sound energy flux:

[0043]

[0044]

[0045]

[0046] Among them, the superscript * represents the complex conjugate, and the symbol represents taking the real part, and <> represents calculating the average. represents the Fourier transform of the reconstructed and denoised discrete direct wave sound pressure. represents the Fourier transform of the denoised discrete direct wave horizontal particle velocity. represents the Fourier transform of the denoised discrete direct wave horizontal particle velocity. represents the calculated discrete direct wave arrival angle.

[0047] 6. Step eight is specifically as follows:

[0048] The sound pressure - particle velocity sound field spatial interference spectrum output depth function along the target direct wave arrival angle is specifically as follows:

[0049]

[0050] Among them, I represents the number of discrete signals.

[0051] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention realizes the acquisition of the sound pressure - particle velocity interference structure generated by the target by using a single vector hydrophone without prior information; the vertical arrival angle of the target can be estimated through the sound pressure and particle velocity components received by the vector hydrophone. Compared with the traditional vertical array, the system structure composition and the vertical arrival angle estimation method are greatly simplified; the present invention makes full use of the characteristic of the vector hydrophone to suppress isotropic noise and can suppress the influence of noise to a certain extent; aiming at the influence of noise on the received signal, the present invention also proposes a signal denoising method, which can better realize signal denoising; the present invention only uses a single vector hydrophone to obtain sound pressure and particle velocity information, and combines signal processing methods to realize depth estimation, and has great advantages in terms of system scale, etc., and is applicable to fields such as underwater target early warning and detection. Brief Description of the Drawings

[0052] Figure 1 is the flowchart of the method of the present invention;

[0053] Figure 2 is the schematic diagram of the single vector hydrophone and the target spatial position of the present invention;

[0054] Figure 3 is the received signal diagram of the sound pressure and particle velocity of the present invention;

[0055] Figure 4 is the sound pressure-particle velocity cross-spectrum LOAFR spectrogram of the present invention;

[0056] Figure 5 is the denoising processing and theoretical sound pressure-particle velocity spatial interference structure diagram of the present invention;

[0057] Figure 6 is the denoising processing and theoretical direct wave vertical arrival angle estimation diagram of the present invention;

[0058] Figure 7 is the estimated depth output result of the present invention. Detailed Embodiment

[0059] The present invention will be further described in detail below in conjunction with the drawings and the detailed embodiment.

[0060] In conjunction with Figures 1 to 7 , the present invention proposes a method for estimating the depth of a moving target through the vector sound field information received by a single vector hydrophone in a deep-sea environment.

[0061] In the first step, the single vector hydrophone needs to be placed near the seabed in the deep-sea waveguide environment to receive the sound pressure signal and the particle velocity signal propagated from the target through the direct wave area or the reliable sound path. The received signal is interfered by noise.

[0062] In the second step, the cross-spectrum is constructed using the received sound pressure signal and particle velocity signal, and the frequency of the signal is estimated through the LOFAR spectrum.

[0063] In the third step, the time-domain equidistant discretization processing is performed on the received sound pressure and particle velocity signals to construct a discrete received signal matrix. The specific implementation is as follows:

[0064] For the signal sampling time T, the received signal is divided according to the time interval Δt, where T >> Δt. The divided signals are rearranged according to the matrix. The sound pressure and particle velocity signals of the i-th column can be expressed as:

[0065] p i (sinθ i , ω, t) = p s (sinθi , ω, t) + n pi

[0066] v ri (sinθ i , ω, t) = p s (sinθ i , ω, t)cosθ i + n vri

[0067] v zi (sinθ i , ω, t) = p s (sinθ i , ω, t)cosθ i + n vri

[0068]

[0069]

[0070]

[0071] Among them, t represents the sampling time series within Δt, and t i represents the time corresponding to the discrete segmentation point, r i represents the horizontal distance between the target and the vector hydrophone at discrete time points, R i represents the distance between the target and the vector hydrophone at discrete time points, H represents the depth of the vector hydrophone, and θ i represents the vertical arrival angle.

[0072] According to the virtual source theory, the acoustic pressure signal received by the above single vector hydrophone from the target can be further expressed as:

[0073]

[0074] Among them, e jωt represents the time factor.

[0075] Fourthly, perform random linear coherent accumulation denoising on the discrete acoustic pressure and vibration velocity signals respectively to obtain the preliminary denoised acoustic pressure and vibration velocity signals. The specific implementation is as follows:

[0076] The actually acquired signal segments after division can be expressed as x i (t), where i represents the serial number of the signal segments after division; considering that the interval Δt is much smaller than the sampling time T and the target signal is approximately unchanged, the signal segment x i (t) can be expressed by the following formula:

[0077] x i (t) = x is(t) + n ix (t)

[0078] where x is (t) represents the target signal contained in the signal segment, and n ix (t) represents the noise within this signal segment. x i (t) contains L signal cycles, and the k-th signal cycle can be expressed by the following formula:

[0079] x ik (t) = x iks (t) + n iks (t), k = 1, 2, …, L

[0080] Performing L times of cumulative averaging on the k-th cycle signal can be expressed as:

[0081]

[0082] Step 5: Decompose the signal through EWT, estimate the signal frequency according to the LOFAR spectrum analysis, and reconstruct the target signal by determining the mode function component containing this frequency;

[0083] Step 6: Use the fast Fourier transform to calculate the spectra of the denoised discrete sound pressure and vibration velocity signals, and construct the sound pressure - vibration velocity sound field spatial interference spectrum. The specific implementation method is as follows:

[0084]

[0085] where < > represents calculating the average, and the superscript * represents the complex conjugate, represents the Fourier transform of the reconstructed denoised discrete sound pressure, represents the Fourier transform of the reconstructed denoised discrete horizontal vibration velocity. The sound pressure - vibration velocity spatial interference spectrum can also be approximately expanded and can be expressed as:

[0086]

[0087] where the influence of the noise term on the sound pressure - vibration velocity spatial interference spectrum is ignored in the above formula, and it only represents the periodic interference structure of the spatial interference spectrum.

[0088] Step 7: Utilize the arrival characteristics of the direct wave acoustic ray of the received signal to calculate the horizontal complex sound intensity and vertical complex sound intensity to estimate the vertical arrival angle of the target direct wave. The specific implementation method is as follows:

[0089] Calculate the direct wave horizontal sound energy flux and vertical sound energy flux:

[0090]

[0091]

[0092]

[0093] Among them, the superscript * represents the complex conjugate, and the symbol represents taking the real part, and < > represents calculating the average. represents the reconstructed denoised discrete direct wave sound pressure Fourier transform. represents the denoised discrete direct wave horizontal particle velocity Fourier transform. represents the denoised discrete direct wave horizontal particle velocity Fourier transform. represents the calculated discrete direct wave arrival angle.

[0094] In the eighth step, perform a discrete Fourier transform on the sound pressure-particle velocity sound field spatial interference spectrum along the vertical arrival angle curve of the target direct wave. The depth corresponding to the secondary maximum value after removing the maximum value at the vertical symmetry axis is the target depth estimation value. The specific implementation method is as follows:

[0095] The depth function of the sound pressure-particle velocity sound field spatial interference spectrum output along the arrival angle of the target direct wave is specifically:

[0096]

[0097] Among them, I represents the number of discrete signals.

[0098] The above has described in detail the specific implementation manners of the invention content. Through the above steps, the method of the present invention realizes the construction of the sound field-particle velocity spatial interference spectrum for the noisy received signal by using a single vector hydrophone, and realizes the estimation of the depth of a moving target in combination with the vertical arrival angle of the direct wave; in addition, it also has a strong ability to suppress isotropic noise; for the received noisy signal, the signal is denoised by the random linear summation and EWT methods, making the construction of the sound field-particle velocity spatial interference spectrum and the estimation of the direct wave arrival angle feasible. The present invention will be further described below through simulation experiments.

[0099] The parameter settings are as follows: The depth of the single vector hydrophone is 4600 m, a target with a frequency of 100 Hz propagates from the far field through the direct wave area to the single vector hydrophone, the target depth is 200 m, the total sampling time is 600 s, the discrete time interval is 1 s, and Gaussian white noise with a signal-to-noise ratio of 10 dB is added in the frequency band of 50 Hz to 150 Hz.

[0100] It can be seen from the above simulation examples that the present invention can obtain the sound pressure-particle velocity spatial interference spectrum under strong noise interference. The comparison diagram of the sound pressure-particle velocity spatial interference spectrum is as Figure 5 shown, the comparison diagram of the vertical arrival angle estimation is as Figure 6 shown, and the estimated depth of the moving target is as Figure 7 shown. The simulation experiment shows that the method of the present invention can better realize the estimation of the depth of a moving target.

[0101] In summary, the present invention discloses a method for estimating the depth of a moving target based on a single vector hydrophone. The present invention includes: using a vector hydrophone deployed near the seabed to receive the sound pressure and particle velocity signals excited by a target near the deep-sea surface and propagated through the direct wave zone or a reliable acoustic path; constructing a cross-spectrum of sound pressure and particle velocity, making full use of the non-correlation between signals and noise in the joint processing of sound pressure and particle velocity, and combining LOFAR spectrum analysis to estimate the signal frequency; performing time-domain equally spaced discretization processing on the received sound pressure and particle velocity signals to construct a discrete signal matrix; performing random linear coherent accumulation denoising on the discrete sound pressure and particle velocity signals respectively to obtain preliminary denoised sound pressure and particle velocity signals; decomposing the signal by EWT, and reconstructing the signal according to the LOFAR spectrum analysis to estimate the signal frequency and determining the mode function component containing the frequency; calculating the spectra of the denoised discrete sound pressure and particle velocity signals by fast Fourier transform to construct a sound pressure-particle velocity sound field spatial interference spectrum; using the arrival characteristics of the direct wave acoustic ray of the received signal to calculate the horizontal complex acoustic intensity and the vertical complex acoustic intensity to estimate the vertical arrival angle of the target direct wave; performing discrete Fourier transform on the sound pressure-particle velocity sound field spatial interference spectrum along the curve of the vertical arrival angle of the target direct wave, and the depth corresponding to the sub-maximum value after removing the maximum value at the vertical symmetry axis is the estimated value of the target depth.

Claims

1. A method for estimating the depth of a moving target based on a single vector hydrophone, characterized in that, the steps are as follows: Step 1: Use a single vector hydrophone to obtain the noise-containing sound pressure p(r, ω, t) and vibration velocity signals v r (r, ω, t), v z (r, ω, t) that are excited by a target near the deep-sea surface and propagated through the direct wave zone or a reliable acoustic path Step two: Construct the cross-spectrum of sound pressure and particle velocity, make full use of the non-correlation between the signal and the noise in the joint processing of sound pressure and particle velocity, and estimate the signal frequency by combining LOFAR spectrum analysis; Step three: Perform time-domain equally spaced discretization processing on the sound pressure and particle velocity received signals, and construct a discrete received signal matrix; Step four: Perform random linear coherent accumulation denoising on the discrete sound pressure and particle velocity signals respectively to obtain the preliminary denoised sound pressure and particle velocity signals; Step five: Decompose the signal by EWT, and reconstruct the target signal according to the modal function component containing the frequency determined by estimating the signal frequency through LOFAR spectrum analysis; Step six: Use the fast Fourier transform to calculate the spectra of the denoised discrete sound pressure and particle velocity signals, and construct the sound pressure-particle velocity sound field space interference spectrum; Step seven: Use the arrival characteristics of the direct wave sound ray of the received signal to calculate the horizontal complex sound intensity and the vertical complex sound intensity to estimate the vertical arrival angle of the target direct wave; Step eight: Perform a discrete Fourier transform on the sound pressure-particle velocity sound field space interference spectrum along the curve of the vertical arrival angle of the target direct wave, and the sub-maximum value corresponding to the maximum value removed at the vertical symmetry axis is the estimated value of the target depth.

2. The method for estimating the depth of a moving target based on a single vector hydrophone according to claim 1, characterized in that: Step three is specifically: For the signal sampling time T, divide the received signal according to the time interval Δt, where T >> Δt, and the divided signals are rearranged according to the matrix. The sound pressure and particle velocity signals in the i-th column are expressed as: p i = p(sinθ i , ω, t) + n pi v ri = p(sinθ i , ω, t)cosθ i + n vri v zi = p(sinθ i , ω, t)cosθ i + n vri where t represents the sampling time series within Δt, and t i represents the time corresponding to the discrete segmentation, and r i represents the horizontal distance between the target and the vector hydrophone at discrete times, and R i represents the distance between the target and the vector hydrophone at discrete times, H represents the depth of the vector hydrophone, and θ i represents the vertical arrival angle; According to the virtual source theory, the above-mentioned sound pressure signal received by the single vector hydrophone from the target is further expressed as: Among them, e jωt represents the time factor.

3. The method for estimating the depth of a moving target based on a single vector hydrophone according to claim 1, characterized in that: Step 4 is specifically as follows: The actual acquired signal segment after division is x i (t), where i represents the serial number of the signal segment after division; Considering that the interval Δt is much smaller than the sampling time T, the target signal is approximately unchanged, and the signal segment x i (t) is: x i y(t) = x iS y(t) + n ix y(t) where x is (t) represents the target signal included in the signal segment, and n ix (t) represents the noise within this signal segment; x i (t) contains L signal cycles, and the k-th signal cycle is: x ik y(t) = x ikS y(t) + n ikS y(t), k = 1, 2, …, L Perform L times of random accumulation averaging on the k-th cycle signal, which is expressed as:

4. The method for estimating the depth of a moving target based on a single vector hydrophone according to claim 1, characterized in that: Step seven is specifically: Calculate the horizontal sound energy flux and the vertical sound energy flux of the direct wave: where the superscript * represents the complex conjugate, and the symbol represents taking the real part, and <> represents calculating the average, represents the reconstructed denoised discrete direct wave sound pressure Fourier transform, represents the denoised discrete direct wave horizontal particle velocity Fourier transform, represents the denoised discrete direct wave horizontal particle velocity Fourier transform, represents the calculated discrete direct wave arrival angle.

5. The method for estimating the depth of a moving target based on a single vector hydrophone according to claim 1, characterized in that: Step eight is specifically: The depth function output by the sound pressure-particle velocity sound field space interference spectrum along the arrival angle of the target direct wave is specifically: where I represents the number of discrete signals.

Images