Target direction finding method based on single channel instantaneous phase difference weighting of vector hydrophone

Abstract

The present invention discloses a target direction finding method based on the weighted instantaneous phase difference of a single channel of a vector hydrophone. This method utilizes the characteristics that the line spectrum intensity of an underwater target is 10 to 25 dB higher than that of a nearby continuous spectrum, and that the instantaneous phase difference corresponding to its frequency unit is relatively stable, while the instantaneous phase difference corresponding to the background noise frequency unit changes randomly. The variance of the instantaneous phase difference of each channel frequency unit is weighted to further suppress the energy interference of the background noise, enhance the line spectrum signal-to-noise ratio gain, and achieve high-precision direction finding of the underwater target position. Through the above-mentioned method, the present application utilizes the weighted processing principle of the stability of the instantaneous phase difference of the line spectrum to enhance the line spectrum energy in the sound pressure velocity channel signal of the vector hydrophone, and further improves the target position accuracy by combining the complex sound intensity. The simulation analysis and experimental data processing results prove the effectiveness and good engineering practicality of the processing method.

Description

Technical Field

[0001] The invention belongs to the technical field of signal processing, and relates to a target direction finding method under long-range low signal-to-noise ratio conditions, and in particular to a target direction finding method based on weighted instantaneous phase difference variance of a single channel of a vector hydrophone. Background Art

[0002] Target direction finding has always been a hot research topic in the field of hydroacoustics. In recent years, the use of vector hydrophones for detection and azimuth estimation has become the main means of direction finding. Vector hydrophones are composed of sound pressure sensors and vibration velocity sensors, which can synchronously pick up sound pressure and vibration velocity information in the sound field. Compared with traditional sound pressure hydrophones, vector hydrophones can detect sound vector information containing spatial azimuth information. Hui Junying et al. explored the principle of sound pressure and vibration velocity joint information processing technology, and achieved effective suppression of coherent interference in sound pressure and vibration velocity joint processing. Yao Zhixiang et al. proposed a histogram method based on complex sound intensity, and pointed out that this method is suitable for line spectra and broadband target sound source signals that are closer to the actual ships. Therefore, the performance of passive sonar direction finding can be significantly improved by joint processing of sound pressure and vibration velocity. Zhang Xiaoyong et al. used the sliding window method to find the interval where the sound source is located under low signal-to-noise ratio, and used the centroid method to obtain the energy center in the interval. However, these methods do not make full use of the line spectrum characteristics in the radiated noise of underwater targets, resulting in poor azimuth estimation.

[0003] In fact, the composition of noise in the marine environment is extremely complex, such as ocean dynamic noise generated by natural factors such as waves, ocean currents and wind blowing across the sea surface, biological noise generated by various organisms living in the ocean, such as fish, shrimp, mammals, microorganisms, etc., and technical noise caused by human activities, such as the movement and collision of mechanical parts of ships and technical equipment in ports. These factors will cause very serious interference to the received sound signals. At the same time, with the continuous development of stealth technology, the signal-to-noise ratio of vector hydrophone receiving data continues to decrease, and the azimuth estimation of long-range low-signal-to-noise ratio underwater targets still has large errors and estimated value jumps. Summary of the invention

[0004] In view of the fact that the signal-to-noise ratio of passive sonar receiving data is constantly decreasing, the present invention aims to provide a target direction finding method based on vector hydrophone single-channel instantaneous phase difference weighting, weighting the single-channel line spectrum signal to improve the signal-to-noise ratio of the received signal, thereby achieving the purpose of high-precision direction finding. The method comprises the following steps:

[0005] Step 1: Extract low-frequency information from the sound pressure information and two-dimensional vector velocity information collected by the vector hydrophone by channel, and then perform discretization processing.

[0006] Preferably, the frequency of the low-frequency information is 15 Hz to 200 Hz.

[0007] Step 2: Perform FFT analysis on the three-channel discrete signals obtained in step 1, and then extract the instantaneous phase of each frequency unit.

[0008] Step 3: Repeat steps 1 and 2 to update M received signals, and calculate the instantaneous phase difference of each frequency unit according to the signal update sequence.

[0009] Step 4: According to the calculation result of step 3, the variance of the instantaneous phase difference values ​​of all frequency units is calculated, and then the variance of the instantaneous phase difference values ​​of the three channels is averaged and filtered.

[0010] Step 5: Use the inverse of the result calculated in step 4 as the weight, and weight the variance of the instantaneous phase difference of all frequency units under the three channels to obtain the final spectrum and signal line spectrum estimation value.

[0011] Step 6: After the signals collected by the vector hydrophone are processed in steps 1 to 5, the sound pressure is conjugated with the X-axis and Y-axis velocity channels to obtain the real part, and the frequency point sound energy flow is statistically analyzed to obtain the target direction.

[0012] The present invention has the following beneficial effects:

[0013] 1. This method is based on the complex sound intensity direction finding technology of vector hydrophone, which makes full use of the sound field vector information picked up by the sound vector sensor, and uses the joint information processing technology of sound pressure and particle vibration velocity to have a strong line spectrum direction finding capability, and can estimate the directions of multiple targets at the same time.

[0014] 2. Based on the complex sound intensity, this method uses the characteristics that the instantaneous phase difference corresponding to the target line spectrum frequency unit is relatively stable, while the instantaneous phase difference corresponding to the background noise frequency unit changes randomly. The variance of the instantaneous phase difference of each channel frequency unit is weighted, which can further suppress the energy interference of the background noise, enhance the line spectrum signal-to-noise ratio gain, and achieve high-precision direction finding for underwater target azimuth estimation.

[0015] 3. This method has no requirements on the bandwidth and frequency of the received signal and has good robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a schematic diagram of the composition of ocean background noise and target radiation noise.

[0017] Figure 2 It is a schematic diagram of the noise spectrum radiated by underwater targets.

[0018] Figure 3 It is a schematic diagram of underwater target direction finding based on instantaneous phase difference weighting.

[0019] Figure 4 Schematic diagram of simulation results in the embodiment.

[0020] Figure 5 Schematic diagram of the deployment of the signal collection system in the embodiment.

[0021] FIG6(a) and FIG6(b) are respectively a conventional LOFAR diagram and a LOFAR diagram based on a weighted method of single-channel instantaneous phase difference in the embodiment.

[0022] FIG. 7( a ) and FIG. 7( b ) are respectively a conventional time-azimuth history diagram and a time-azimuth history diagram based on a single-channel instantaneous phase difference weighted method in the embodiment. DETAILED DESCRIPTION

[0023] The present invention will be described in further detail below with reference to the accompanying drawings.

[0024] The present invention is a target direction finding method based on the weighted instantaneous phase difference of a single channel of a vector hydrophone. The method is based on the complex intensifier multi-target direction of arrival (DOA) estimation method, and makes full use of the sound field vector information picked up by the acoustic vector sensor. At the same time, by utilizing the characteristics that the instantaneous phase difference corresponding to the target line spectrum frequency unit is relatively stable, and the instantaneous phase difference corresponding to the background noise frequency unit changes randomly, the variance of the instantaneous phase difference of each channel frequency unit is weighted, which can further suppress the energy interference of the background noise, enhance the line spectrum signal-to-noise ratio gain, and realize high-precision direction finding for underwater target azimuth estimation.

[0025] 1. Data Model

[0026] When a ship, submarine or torpedo is sailing or working, the vibration of various equipment such as propellers will radiate sound waves through the hull to the underwater, becoming the main noise source of ship target radiation. The sources of underwater target radiation noise field are mainly divided into three categories: mechanical noise, propeller noise and hydrodynamic noise. The basic composition and spectral characteristics are as follows: Figure 1 shown.

[0027] The noise radiated by underwater targets can be broken down into the noise of each component, and the noise of each component can be represented by a continuous spectrum and a line spectrum. Therefore, the total power spectrum of the noise radiated by underwater targets can be represented by the superposition of the power spectrum of the broadband continuous noise spectrum and the power spectrum of the narrowband line spectrum. Figure 2 Schematic diagram of the noise spectrum radiated by underwater targets.

[0028] The continuous spectrum of underwater target radiation noise has a wide frequency band, ranging from a few Hz to tens of thousands of Hz. The low frequency band has a positive slope of 6 to 12dB / oct, and the high frequency band has a negative slope of about -6dB / oct. The peak value generally appears between 100 and 1000Hz. The line spectrum of underwater target radiation noise is mainly distributed in the low frequency band, and in a few cases it extends to around 1000Hz. The line spectrum frequency is relatively stable and very regular. It is the main feature for ship type identification. The simplified form of the underwater target radiation signal x(t) can be expressed as:

[0029]

[0030] In the formula, A n (t) is the amplitude of the line spectrum signal, which is a random variable uniformly distributed on [0, 1]; f n is the line spectrum signal frequency, is the initial phase of the line spectrum signal, t is the target radiation signal moment, bs(t) is the bandwidth signal; N is the number of assumed independent components, and bs(t) are independent of each other, and It obeys the uniform distribution of [0~2π].

[0031] Among them, the ratio of the nth line spectrum signal component to the average spectrum level of the continuous spectrum is for:

[0032]

[0033] Where B is the bandwidth of the bandwidth signal bs(t), σ bs 2 is the variance of the bandwidth signal bs(t).

[0034] From formula (2), we can see that the line spectrum signal level of the target radiation signal is much higher than the average spectrum level of the broadband signal. If we want to detect the target direction, we can do so by detecting the line spectrum signal.

[0035] 2. Principle of Direction Estimation

[0036] The two-dimensional vector hydrophone cannot directly output the target radiation signal vector information. It is composed of a sound pressure sensor and two particle velocity sensors located in the X-axis direction and the Y-axis direction respectively. Through these three channels, the two-dimensional vector velocity information and sound pressure information of the sound field can be obtained synchronously at the same point. It should be noted that the X-axis direction and the Y-axis direction here do not represent the actual orientation, but only represent that the two-channel velocity signals in the two-dimensional vector hydrophone are orthogonal. The information model picked up by the vector hydrophone can be expressed as follows:

[0037]

[0038] Where p(t) is the sound pressure signal picked up by the vector hydrophone, vx (t) is the velocity signal in the X-axis direction picked up by the vector hydrophone, v y (t) is the Y-axis velocity signal picked up by the vector hydrophone, α is the azimuth angle of the target in the horizontal direction, n p (t), n vx (t), n vy (t) are the interference noise of sound pressure and vibration velocity in the X-axis and Y-axis directions respectively.

[0039] The complex sound intensity method first performs Fourier transform on the output of the vector hydrophone to obtain P(f), V x (f) V y (f) Then the sound pressure is conjugate-multiplied with the two velocity components to obtain two cross spectra:

[0040]

[0041] In the ocean channel, the sound pressure and the vibration velocity are in phase. According to the basic characteristics of Fourier transform, the energy of the two in-phase inputs is concentrated in the real part of the cross spectrum. Therefore, taking the real part of the result of equation (4) will give the sound energy flow component I of each frequency. x (f) and I y (f), thus obtaining the modulus R(f) of the acoustic energy flow at each frequency point:

[0042]

[0043] The estimated value of the sound source direction at each frequency point is:

[0044]

[0045] The azimuth angle corresponding to each frequency point is calculated by formula (6). When the frequency range of the target radiation noise is unknown, the target azimuth angle cannot be directly obtained by formula (6). Therefore, it is necessary to use the histogram statistics method to count the azimuth angles corresponding to each frequency point. When there is a broadband target signal, the azimuth angles corresponding to each frequency point of the signal fall within a certain interval, while the azimuth angles corresponding to each frequency point of the noise band are randomly distributed. Divide [0°, 360°] into several intervals △θ, then the total number of intervals N allowed by the spatial angle is:

[0046]

[0047] From the above formula, we can get the interval range of the horizontal azimuth angle: 0~△θ, △θ~2△θ, 2△θ~3△θ, ..., (N-1)△θ~N△θ. Scan the entire frequency band to get the corresponding azimuth angle. Assume that the number of estimated azimuths of each frequency point falling into each statistical angle interval is m1, m2, m3, ..., m N, then calculate the following statistic S(θ):

[0048]

[0049] In the formula, S s (f) is the modulus of the acoustic energy flow that conforms to the angular range of the statistical interval. The distribution of the acoustic energy flow in each statistical interval, i.e., the weighted histogram, is obtained by equation (8).

[0050] 3. Principle of weighted processing of line spectrum instantaneous phase difference stability

[0051] First, perform FFT analysis on the collected data in blocks, and the resulting spectrum is recorded as X(m,f k ), m = 1, 2, ..., M, k = 1, 2, ..., K; where M is the number of blocks of collected data, and K is the number of frequency units obtained by FFT analysis of each block. The conventional FFT analysis method is used for line spectrum detection, and the final spectrum can be expressed as:

[0052]

[0053] In the filter band f1~f K When the target radiation signal occupies only one or several frequency units f l , l=1,2,…,L, the rest are noise frequency units f s , s = 1, 2, ..., S, S + L = K. If the final spectrum is obtained by using formula (9), the corresponding spectrum values ​​of all frequency units will be equivalently weighted into the final spectrum. Since the background noise spectrum value fluctuates greatly, the final spectrum is greatly affected by the noise, which is not convenient for the final complex sound intensity direction finding.

[0054] Therefore, FFT analysis is performed on each block of data, and the instantaneous phase of all frequency units in the result is extracted, which is recorded as Since the sampled data is processed in blocks, the number of frequency units obtained by FFT analysis of each block is a fixed value, and it can be obtained that the phase difference of the frequency unit is relatively stable at different times. That is, it satisfies the following expression:

[0055]

[0056] Among them, due to the strong stability of the line spectrum, its f n It is relatively fixed. Stable and with relatively small variance. The noise f s is a random change. It is not stable and its variance is much larger than the line spectrum.

[0057] In this regard, by using the difference in the instantaneous phase difference between the line spectrum frequency unit and the noise frequency unit, the variance of the instantaneous phase difference of each channel frequency unit is weighted and summed, which can weaken the influence of noise on the final spectrum, enhance the signal-to-noise ratio gain of the line spectrum detection, and suppress the interference of background noise energy. The above content theoretically discusses the effectiveness of this method.

[0058] 4. Direction finding method based on single-channel instantaneous phase difference weighting

[0059] The energy fluctuation of the noise signal received by the vector hydrophone is relatively large, and the distribution is uneven and random, which seriously interferes with the direction finding accuracy. Based on the stability of the instantaneous phase difference of the line spectrum, the target line spectrum received by each channel of the vector hydrophone is enhanced, and the interference noise and continuous spectrum are suppressed at the same time, so that high-precision direction finding of underwater targets can be achieved. The steps of this method are introduced as follows:

[0060] Step 1: Figure 3 As shown in the figure, taking the sound pressure signal received by the two-dimensional vector hydrophone as an example, the received signal is first filtered and amplified to extract its 15Hz~200Hz low-frequency information, and then sampled through the A / D circuit with a sampling rate of f s , and obtain the discrete signal p(n'T s ), where T s =1 / f s , n'=1,2,…,N', N' is the data length;

[0061] Step 2: According to formula (11), p(n'T s ) to perform FFT analysis and extract the analysis result P(kω s ) The instantaneous phase of each frequency unit

[0062]

[0063] In formula (11), ω s =2π / KT s is the frequency sampling interval in FFT analysis.

[0064] Step 3: Update the received signal, repeat steps 1 and 2 M times, then each frequency unit will get M instantaneous phase values, and the mth instantaneous phase value is recorded as The instantaneous phase difference of all frequency units is calculated according to formula (12), which is expressed as

[0065]

[0066] In formula (12), 1≤n≤M-1, 1≤k≤K.

[0067] Step 4: Calculate the variance of the instantaneous phase difference of all frequency units according to formula (13), which is expressed as

[0068]

[0069] In formula (13), Represents the average value of the instantaneous phase difference of all frequency units.

[0070] Step 5: Process the vibration velocity information in the X-axis and Y-axis directions received by the two-dimensional vector hydrophone through steps 1 to 4 to obtain the variance of the instantaneous phase difference under the corresponding channel, and then perform an average fitting on the variance of the three-channel data according to formula (14), and perform mean filtering on it to obtain

[0071]

[0072] Step 6: According to the calculation results of step 5, the variance weighted statistics of the instantaneous phase difference of the frequency units under the three channels are performed to obtain the final spectrum and signal line spectrum estimation values, as shown in formula (15):

[0073]

[0074] When the fitted value When it is small, its corresponding frequency unit spectrum value is The larger the value, the enhanced line spectrum is, which is convenient for complex sound intensity direction finding.

[0075] Step 7: Substitute p(t) and v x (t), v y (t) After the instantaneous phase difference weighted processing, the conjugate cross-spectrum is performed respectively, the real part is taken, and the azimuth of each frequency point is calculated using the inverse trigonometric function. The acoustic energy flow of the frequency point is statistically analyzed to obtain the final target azimuth.

[0076] V. Performance Analysis

[0077] In order to verify the performance of the weighted processing of the instantaneous phase difference stability of the line spectrum, the broadband noise obtained by exciting the FIR linear filter with Gaussian white noise is used as the continuous spectrum signal, and the line spectrum is superimposed on the target signal. The line spectrum intensity is 10 to 25 dB higher than the nearby continuous spectrum.

[0078] The root mean square error (RMSE) of the proposed method is compared with the conventional complex sound intensity filter method when the signal-to-noise ratio is -20 to 30 dB, and each spectrum level is independently counted 10,000 times. The sampling length of each time is T = 3 minutes, and the sampled data is divided into 36 segments, and FFT analysis is performed on each segment for 5 seconds. The frequency of the target line spectrum of each sampling is uniformly distributed within 15Hz to 200Hz, the horizontal incident angle is uniformly distributed within 0° to 360°, and it is in a background noise environment composed of Gaussian white noise and noise of different directions and intensities. The simulation results are shown in Figure 2. Figure 4 shown.

[0079] In order to verify the effectiveness of this method, relevant lake test data were processed. This data was collected and stored in the Fuchun River in October 2018. The water depth was about 20m and the river bottom was flat. The target ship and receiving equipment used in the test are shown in the figure below. Figure 5 As shown in the figure, the vector hydrophone connected to the buoy is deployed at a depth of about 18m. The target low-frequency sound source moves about 1km behind the buoy. The sound source is 5m away from the water surface, with a frequency of 62Hz and an intensity about 10dB higher than that of the nearby waters. The sampling rate of the acquisition equipment connected to the hydrophone is 512Hz. The acquisition length is about 3min. The acquired data is divided into 36 segments for FFT analysis. Then, the final LOFAR map and time-azimuth history map are obtained by conventional methods and instantaneous phase difference weighting methods. The LOFAR comparison diagram is shown in Figure 6(a) and Figure 6(b), and the time-azimuth comparison diagram is shown in Figure 7(a) and Figure 7(b).

[0080] Comparing Figure 6 and Figure 7, we can see that under low signal-to-noise ratio, the target azimuth calculated by the conventional complex sound intensity method has a large deviation due to the interference of noise, and the target is difficult to identify. Relatively speaking, the spectrum calculated based on this method is enhanced, thereby suppressing noise interference, the target azimuth accuracy is improved to a certain extent, and the time azimuth history diagram is clearer.

[0081] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the scope of the present application should be based on the protection scope of the claims.

Claims

1. A target direction finding method based on weighted instantaneous phase difference of a single channel of a vector hydrophone, characterized in that: The method comprises the following steps: Step 1: Extract low-frequency information from the sound pressure information and two-dimensional vector velocity information collected by the vector hydrophone according to the channel; s The sound pressure information collected by the vector hydrophone is sampled to obtain a discrete signal p(n'T s ); Step 2: Perform FFT analysis on the three-channel discrete signals obtained in step 1 according to formula (1), and then extract the instantaneous phase of each frequency unit. Among them, T s =1 / f s , n'=1,2,…,N', N' is the data length; ω s =2π / KT s is the frequency sampling interval in FFT analysis; Step 3: Repeat steps 1 to 2 to update M received signals. Each frequency unit obtains M instantaneous phase values. The mth instantaneous phase value is recorded as According to the signal update order, calculate the instantaneous phase difference of each frequency unit Among them, 1≤n≤M-1, 1≤k≤K; Step 4: Based on the calculation result of step 3, calculate the variance of the instantaneous phase difference value of all frequency units according to formula (3): in, Represents the average value of the instantaneous phase difference of all frequency units; the two-dimensional vector velocity information is processed by the same method to obtain the variance value δ of the instantaneous phase difference of all frequency units. vx (kω s ),δ vy (kω s ); Then the variance of the instantaneous phase difference values ​​of the three channels is averaged and filtered; Step 5: Use the inverse of the result calculated in step 4 as the weight, and weight the variance of the instantaneous phase difference of all frequency units under the three channels respectively, and obtain the final spectrum and signal line spectrum estimation values ​​as follows: in, The variance values ​​of the instantaneous phase difference values ​​of all frequency units under three channels are averaged and filtered to obtain: Step 6: After the signals collected by the vector hydrophone are processed in steps 1 to 5, the sound pressure channel is conjugated with the two-dimensional vector velocity channel to obtain the real part, and the frequency point sound energy flow is counted to obtain the target direction.

2. The target direction finding method based on weighted instantaneous phase difference of a single channel of a vector hydrophone as claimed in claim 1, characterized in that: The frequency of the low-frequency information extracted in step 1 is 15 Hz to 200 Hz.

Images