Ultra-short baseline positioning method and system based on CZT cross-correlation delay difference spectrum

Abstract

The present application provides an ultra-short baseline positioning method and system based on CZT cross-correlation delay difference spectrum, which measures the propagation time and phase change of the sound wave between the target cooperation end and the USBL array composed of multiple hydrophones to obtain the propagation distance, and then estimates the three-dimensional coordinates of the target; the method includes: the target cooperation end amplifies and automatically gains the received data; performs fast Fourier transform and matched filtering on the data, and then performs inverse fast Fourier transform and peak detection to obtain the arrival time estimation of multiple channel target signals, and then estimates the distance between the target cooperation end and the center of the USBL array; calculates the frequency domain output of the matched filtering of multiple channels to obtain the secondary correlation output, and then performs band filtering, performs CZT transformation on the secondary correlation function, uses the peak to find the delay difference, and calculates the coordinates of the target cooperation end according to the delay difference. The advantages of the present application are: accurate target position solution and small amount of calculation.

Description

Technical Field

[0001] The present application belongs to the field of ultra-short baseline positioning, and specifically relates to an ultra-short baseline positioning method and system based on CZT cross-correlation delay difference spectrum. Background Art

[0002] Underwater navigation and positioning is of great significance to ensure the safety of submarines and underwater vehicles in deep seas and oceans and to improve their operational efficiency. Therefore, it has attracted widespread attention from the industrial and academic circles of various countries in recent years. Active underwater positioning methods usually obtain the propagation distance by measuring the propagation time and phase change of the sound waves between the target and multiple beacons, and then estimate the three-dimensional coordinates of the target. The line between the beacons (i.e., primitives) is called the baseline, which is divided into long baseline positioning (LBL, 100-6000m), short baseline positioning (SBL, 1-50m) and ultra-short baseline positioning (USBL, <1m) according to the length of the baseline. Among them, the ultra-short baseline has the advantages of flexible installation and use, and has been widely used in platforms such as UUV.

[0003] The USBL system obtains the three-dimensional coordinates by locating the three-dimensional coordinates of the node. The three-dimensional coordinates of the node to be located are obtained from the receiver array. The most common technology is based on ranging estimation, that is, the system obtains the distance between the node and the hydrophone and the arrival time difference between the channels of the hydrophone by measuring the propagation time or phase change of the sound wave. Accurate measurement of the arrival time difference (TDOA) is a key factor to ensure the performance of the USBL system. The classic TDOA method uses the arrival time (TOA) obtained by the matched filter (MF) to estimate the time difference between the echoes extracted from the multi-channel data, and then calculates the difference between the arrival time difference and the echo time difference.

[0004] To improve the estimation accuracy, two solutions are usually adopted: waveform optimization and high-resolution methods. Waveform optimization involves using coded broadband signals to obtain sharp peaks in the output of the matched filter, while high-resolution methods such as MUSIC and compressed sensing (CS) can achieve better estimation accuracy. However, the accuracy of these methods is limited by the sampling frequency. For USBL systems, the ratio of target distance to array element spacing is usually greater than 10,000, the TDOA between array element channels is usually on the order of 10 milliseconds or even 1 millisecond, and the sampling frequency is usually 100kHz, which will produce unacceptable errors for the time difference calculation required for positioning calculation. Although increasing the sampling frequency can reduce this error, it will also lead to a significant increase in computing and data storage requirements.

[0005] In order to avoid the error caused by limited sampling frequency, Omologo proposed the cross-correlation phase (CSP) method in 1994. This method avoids direct calculation in the time domain, but obtains the time difference by inverse Fourier transforming the cross-correlation spectrum between the two channels. This method is widely used in USBL positioning systems, and some scholars have further proposed increasing the number of IFFT points to refine the cross-correlation spectrum, thereby improving the estimation accuracy. However, the CSP method has a large amount of calculation, inaccurate settlement, and requires sufficient computing resources. The computing efficiency cannot meet the requirements of underwater navigation and positioning. Summary of the invention

[0006] The purpose of this application is to overcome the defects of the prior art, such as large amount of calculation and inaccurate settlement.

[0007] In order to achieve the above purpose, this application proposes an ultra-short baseline positioning method based on CZT cross-correlation delay difference spectrum, which measures the propagation time and phase change of the sound wave between the target cooperation end and the USBL array composed of multiple hydrophones to obtain the propagation distance, and then estimates the three-dimensional coordinates of the target;

[0008] The method comprises:

[0009] Step 1: The target cooperation end amplifies and performs automatic gain control processing on the received data;

[0010] Step 2: Perform fast Fourier transform and matched filtering on the data processed by the target cooperation end, and then perform inverse fast Fourier transform and peak detection to obtain the arrival time estimation of multiple channel target signals, and then estimate the distance between the target cooperation end and the center of the USBL array;

[0011] Step 3: Calculate the frequency domain output of the matched filter of multiple channels to obtain the quadratic correlation output, then perform band filtering, perform CZT transformation on the quadratic correlation function, use the peak value to find the delay difference, and calculate the coordinates of the target cooperation end based on the delay difference.

[0012] As an improvement of the above method, the USBL array is a four-element cross array on a plane, and the four hydrophones are numbered from the 1st to the 4th in a counterclockwise direction, with the center of the positions of the four hydrophones as the origin, the line connecting the 1st and the 3rd hydrophones as the x-axis, the line connecting the 2nd and the 4th hydrophones as the y-axis, and the direction perpendicular to the xy plane as the z-axis.

[0013] As an improvement of the above method, step 2 comprises:

[0014] Perform a fast Fourier transform on the waveform transmitted by the target partner to generate a frequency domain replica signal X rep :

[0015] X rep=DFT(s(t),N)

[0016] Wherein, N represents the number of fast Fourier transform points; s(t) represents the waveform transmitted by the target cooperation end; DFT() represents fast Fourier transform;

[0017] Each frame of data of each channel is obtained with a set multiple of redundancy, and a fast Fourier transform is performed to convert the broadband time domain data into the frequency domain. The frequency domain data of the nth frame is represented as X n , One channel corresponds to one hydrophone;

[0018] X n With X rep Multiply to obtain the frequency domain result of matched filtering;

[0019] For each frame of data, the frequency domain output of the matched filter is subjected to inverse fast Fourier transform to obtain the time domain result of the matched filter, and peak detection is performed respectively;

[0020] Compare the time domain output with the set threshold. If both the positive and negative frequency modulation outputs are greater than the set threshold, it is considered that there is a target echo. Otherwise, it is considered that there is no target echo in this frame, so it is skipped and the next frame is processed.

[0021] For the case where a target echo is considered to exist, the arrival time of the target signal of the four channels is estimated from the peak position.

[0022] Estimate the distance between the target partner and the center of the USBL array for:

[0023]

[0024] Among them, C USBL Indicates the speed of sound near the USBL system.

[0025] As an improvement of the above method, N = 4Tf s , where f s is the sampling frequency; T is the pulse width of the transmitted waveform.

[0026] As an improvement of the above method, the setting multiple is 0.5.

[0027] As an improvement of the above method, step 3 comprises:

[0028] For the frame with the target signal, the frequency domain outputs of the matched filter of the 1st and 3rd channels are Z1 and Z3, and Z3 is conjugated and multiplied to obtain the quadratic correlation output Y1=Z1*Z'3; the frequency domain outputs of the matched filter of the 2nd and 4th channels are Z2 and Z4, and Z4 is conjugated and multiplied to obtain the quadratic correlation output Y2=Z2*Z'4;

[0029] After band filtering, we get the quadratic correlation function Keep the values ​​of the frequency range f1 to f2 where the target signal may exist, and set the other regions to 0; f1 = f c -B / 2, f2=f c +B / 2;f c represents the center frequency of the waveform transmitted by the target cooperation end; B represents the bandwidth of the waveform transmitted by the target cooperation end;

[0030] Perform CZT on the quadratic correlation function;

[0031] The CZT transform of the first group of channel correlation functions obtains the cross-correlation delay difference spectrum The peak delay difference is The CZT transform of the second group of channel correlation functions obtains the cross-correlation delay difference spectrum The peak delay difference is Where, m = 0, ..., M-1, M is the number of points of CZT transformation; ω1 = exp(j2π(Δt end1 -Δt begin1 ) / (NM)), j represents the imaginary part;

[0032] ω2=exp(j2π(Δt end2 -Δt begin2 ) / (NM)), [Δt begin1 ,Δt end1 ] is the range of the delay difference between channel 1 and channel 3, [Δt begin2 ,Δt end2 ] is the range of the delay difference between channel 2 and channel 4;

[0033] a1=exp(-j*2πΔt begin1 / N), a2=exp(-j*2πΔt begin2 / N); the first group of channels are the 1st and 3rd channels; the second group of channels are the 2nd and 4th channels;

[0034] Calculate the coordinates of the target partner based on the delay difference:

[0035]

[0036] Where d is the distance between each group of hydrophones.

[0037] The present application also provides an ultra-short baseline positioning system based on CZT cross-correlation delay difference spectrum, which is implemented based on any of the above methods, and the system includes:

[0038] The front-end processing module is used for the target cooperation end to amplify and automatically gain control the received data;

[0039] The module for estimating the distance between the target cooperation end and the center of the array is used to perform fast Fourier transform and matched filtering on the data processed by the target cooperation end, and then perform inverse fast Fourier transform and peak detection to obtain the arrival time estimation of multiple channel target signals, and then estimate the distance between the target cooperation end and the center of the USBL array;

[0040] The target cooperation end coordinate calculation module is used to calculate the frequency domain output of the matched filter of multiple channels to obtain the secondary correlation output, then perform frequency band filtering, perform CZT transformation on the secondary correlation function, use the peak value to find the delay difference, and calculate the coordinates of the target cooperation end according to the delay difference.

[0041] Compared with the prior art, the advantages of this application are:

[0042] The present invention innovatively uses CZT transformation in USBL system positioning, which can obtain accurate solution of target position, and the amount of calculation is less than that of traditional CSP method and CSP refinement method. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Shown is a schematic diagram of USBL array reception;

[0044] Figure 2 The USBL system positioning flow chart of the CZT correlation function delay difference spectrum is shown;

[0045] Figure 3 The figure shows the estimation errors of different methods under different SNRs; ​​SNR represents signal-to-noise ratio; and 3D RMSE represents root mean square error. DETAILED DESCRIPTION

[0046] The technical solution of the present application is described in detail below with reference to the accompanying drawings.

[0047] The ultra-short baseline positioning method and system based on CZT cross-correlation delay difference spectrum provided in this application has the following system settings of USBL: Figure 1 As shown: The USBL array is a four-element cross array on a plane. The four hydrophones are numbered from 1 to 4 in a counterclockwise direction. The center of the four hydrophones is the origin, the line between the 1st and 3rd hydrophones is the x-axis, the line between the 2nd and 4th hydrophones is the y-axis, and the direction perpendicular to the xy-axis plane is the z-axis.

[0048] The target cooperation end transmit waveform is set as follows:

[0049] The distance between the target cooperation end and the center of the USBL array is R. In order to improve the estimation accuracy of R and the estimation accuracy of the subsequent delay difference, the bandwidth of the transmission waveform can be appropriately increased. The center frequency f is set according to the propagation distance and the array element spacing. c , set the wavelength corresponding to the center frequency to be roughly equivalent to the spacing; generally set to a linear frequency modulation (LFM) signal, and the bandwidth B is set by requirements such as the effective distance and ranging accuracy; T is the pulse width of the transmitted waveform.

[0050] Example 1

[0051] like Figure 2 As shown, the ultra-short baseline positioning method based on CZT cross-correlation delay difference spectrum provided by the present application includes:

[0052] Step 1: Receive front-end processing;

[0053] Consistent with the front-end processing of other existing sonar / radar platforms, this application first amplifies the received data, performs automatic gain control AGC, etc. The received signal obtained by the i-th hydrophone can be expressed as: i (t) = a i s(t-τ i )+n(t). Among them, s(·) represents the waveform transmitted by the target cooperation end, and n(t) represents the noise, including interference and background noise. Considering the propagation attenuation, we introduce the attenuation factor a i and time delay τ i τ i is the time when the signal arrives at the i-th hydrophone. To simplify the problem, we assume that n(t) is Gaussian white noise and is uncorrelated with s(·).

[0054] Step 2: Estimation of the distance between the target partner and the center of the USBL array;

[0055] Step 2.1: Set the number of points N of the window width;

[0056] Set the number of points of the window width N = 4Tf s , where f s is the sampling frequency.

[0057] Step 2.2: Generate a replica signal;

[0058] Perform a fast Fourier transform (FFT) on the transmitted waveform to generate a frequency domain replica signal X rep =DFT(s(t),N), where the number of FFT points is N.

[0059] Step 2.3: Perform FFT with divided windows;

[0060] Each frame of data of the first channel is obtained with 0.5 times redundancy, that is, the length of each frame of data is the window width, half of which is the second half of the previous frame, and the other half is the new data. Perform FFT to convert the broadband time domain data to the frequency domain. The frequency domain data of the nth frame is X n ,

[0061] Step 2.4: Matched filtering;

[0062] X n With X rep Multiply them to obtain the frequency domain result of matched filtering.

[0063] Step 2.5: Threshold detection to select peaks

[0064] Under each window, the frequency domain output of the matched filter is subjected to an inverse fast Fourier transform (IFFT) to obtain the time domain result of the matched filter, and peak detection is performed respectively.

[0065] Compare the time domain output with the threshold. If both the positive and negative frequency modulation outputs are greater than a certain threshold, it is considered that there is a target echo. Otherwise, it is considered that there is no target echo in this frame, so it is skipped and the next frame is processed.

[0066] Step 2.6: Determine the delay;

[0067] For the case where a target signal is considered to exist, the arrival time of the target signal of the first channel is estimated from the peak position. Process them sequentially to obtain the target signal arrival time estimates for the other three channels.

[0068] Step 2.7: Estimate distance;

[0069] The estimated distance between the target partner and the center of the USBL array is:

[0070]

[0071] Among them, C USBL Indicates the speed of sound near the USBL system.

[0072] Step 3: Obtain the delay difference;

[0073] Figure 1 In , the hydrophone groups consist of the 1st and 3rd hydrophones (group 1) and the 2nd and 4th hydrophones (group 2), respectively. Under the far-field plane wave assumption, θ x1 ≈θ x2 ≈θ x .based on Figure 1 The triangular relationship shown in , we get:

[0074]

[0075] Where d is the spacing between each group of array elements. The range difference of the first group of hydrophones is R x1 -R x2 , the corresponding TDOA measurement value is Δt x , the real time difference Δt x =τ1-τ3. So we have R x1 -R x2 =Δt x C USBL .

[0076] Similarly, Δt y =τ2-τ4,

[0077] Therefore, measuring the time difference between the two groups of hydrophones is a key step in the USBL system positioning: the time difference of the first group of channels is used to estimate the x-coordinate of the target partner, and the time difference of the second group of channels is used to estimate the y-coordinate of the target partner. Calculating the x-coordinate will result in large errors due to insufficient sampling frequency.

[0078] Step 3.1: Perform secondary correlation calculation;

[0079] For the frame with the target signal, the frequency domain outputs of the matched filters of the first and third channels are Z1 and Z3. The conjugate of Z3 is taken and then multiplied to obtain the quadratic correlation output Y=Z1*Z'3.

[0080] Step 3.2: frequency band filtering;

[0081] Secondary correlation output after band filtering Keep the values ​​of the frequency range f1 to f2 where the target signal may exist, and set the other regions to 0. f1 and f2 should be f c -B / 2 and f c +B / 2 and taking into account the Doppler shift. This processing can improve the signal-to-noise ratio.

[0082] Step 3.3: Perform linear chirp transform (CZT) on the quadratic correlation function;

[0083] The traditional CSP method performs IFFT on Y, and the peak position is the time delay Δt corresponding to the Doppler x. This method uses quadratic correlation to skip the estimation of τ1 and τ2 and directly estimates the difference of the double delay, which can reduce the estimation error. Increasing the number of IFFT points to several times N can improve the estimation accuracy. However, the IFFT processing method of Y is equivalent to integrating on the entire delay difference plane. In fact, the possible delay difference of the target only exists in a certain range. It is meaningless to perform inverse Fourier transform or integration in the entire delay difference domain, and it wastes the amount of calculation.

[0084] To this end, the present application uses chirp-z transform, namely CZT transform, to perform local processing only within the range of interest, and can further refine within this range, thereby greatly reducing the amount of calculation.

[0085] Step 3.4: Set CZT parameters;

[0086] [Δt begin ,Δt end ] is the possible range of the delay difference between channel 1 and channel 3. Generally speaking, the delay difference between the two is in the order of tens of sampling points. M is the number of points of CZT transformation. begin ,Δt end ] is divided into M-1 intervals. This process is equivalent to changing the sampling frequency from f s Increase to Mf s 2 / (Δt end -Δt begin ).

[0087] z is the sampling point on the Z plane, and z=aω -m , where m = 0, ..., M-1, and a = exp(-j*2πΔt begin / N), a is the starting point of the complex, ω is the extension rate of the spiral ω=exp(j2π(Δt end -Δt begin ) / (NM)).

[0088] Step 3.5: Perform CZT transformation;

[0089] CZT cross-correlation delay difference spectrum Each element of the P vector is the refined [Δt begin ,Δt end ] every point.

[0090] The CZT transform is equivalent to a local Fourier transform, which avoids spectral estimation on the complete delay difference domain and only calculates and searches within the range of interest. Compared with the original CSP method and the CSP method with refined points, the CZT cross-correlation delay difference spectrum method is locally refined, which not only reduces the amount of calculation, but also refines the delay difference spectrum and obtains a more precise estimation result.

[0091] Step 3.6: Peak search to find the delay difference;

[0092] The CZT transform of the first group of channel correlation functions gives P, whose peak delay difference is The delay difference obtained by the second group is

[0093] Calculate the coordinates of the target partner based on the delay difference:

[0094]

[0095] The z coordinate is estimated from the x coordinate, y coordinate, and R.

[0096] The results of the simulation test using this application method are as follows:

[0097] like Figure 1 The array shown has an array element spacing of d = 0.134m and an array element center as the origin. The target position is set to [1492, -2505, 1452], the transmitted signal is an LFM signal, B = 6kHz, f c =8kHz, pulse width T is 20ms, f s =150kHz, the signal-to-noise ratio is set to 5dB.

[0098] For the CSP method of refining the number of IFFT points, the refinement factor is set to 64, that is, the original 1 sampling point is further refined into 64 parts, and the number of IFFT points increases from 16384 to 1048576. For the CZT cross-correlation delay difference spectrum method we proposed, the possible delay difference range is set to plus or minus 50 sampling points, and the number of CZT points is set to 6400.

[0099] Different signal-to-noise ratios (SNRs) are set. Under each SNR, 200 simulations are performed to obtain the root mean square error (3D RMSE) of the positioning error statistics of different positioning methods. The results are shown in Tables 1 and Figure 3 shown.

[0100] Table 1 Positioning errors of various methods under different SNRs (unit: m)

[0101] CSP CSP-interp CSP-Zoom CZT CZT-interp -10dB 380.51 320.64 321.04 321.04 321.06 -5dB 249.03 170.02 167.44 167.44 167.36 0dB 192.05 112.40 102.23 102.23 102.15 5dB 177.85 69.23 58.22 58.22 57.98 10dB 172.72 58.36 29.59 29.59 29.26 15dB 172.72 55.99 17.29 17.29 17.44 20dB 172.72 56.67 9.89 9.89 9.18 25dB 172.72 57.67 6.12 6.12 5.7 30dB 172.72 57.90 4.65 4.65 3.07

[0102] It can be seen from the experimental results that when the refinement multiple is the same, the CZT and CSP-Zoom used in this application achieve the same accuracy, but the calculation amount of the CZT method is significantly reduced.

[0103] Example 2

[0104] The present application also provides an ultra-short baseline positioning system based on CZT cross-correlation delay difference spectrum, which is implemented based on the above method, and the system includes:

[0105] The front-end processing module is used for the target cooperation end to amplify and automatically gain control the received data;

[0106] The module for estimating the distance between the target cooperation end and the center of the array is used to perform fast Fourier transform and matched filtering on the data processed by the target cooperation end, and then perform inverse fast Fourier transform and peak detection to obtain the arrival time estimation of multiple channel target signals, and then estimate the distance between the target cooperation end and the center of the USBL array;

[0107] The target cooperation end coordinate calculation module is used to calculate the frequency domain output of the matched filter of multiple channels to obtain the secondary correlation output, then perform frequency band filtering, perform CZT transformation on the secondary correlation function, use the peak value to find the delay difference, and calculate the coordinates of the target cooperation end according to the delay difference.

[0108] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application and are not intended to limit it. Although the present application is described in detail with reference to the embodiments, a person skilled in the art should understand that any modification or equivalent replacement of the technical solution of the present application does not depart from the spirit and scope of the technical solution of the present application and should be included in the scope of the claims of the present application.

Claims

1. An ultra-short baseline positioning method based on CZT cross-correlation delay difference spectrum, which measures the propagation time and phase change of the sound wave between the target cooperation end and the USBL array composed of multiple hydrophones to obtain the propagation distance, and then estimates the three-dimensional coordinates of the target; The method comprises: Step 1: The target cooperation end amplifies and performs automatic gain control processing on the received data; Step 2: Perform fast Fourier transform and matched filtering on the data processed by the target cooperation end, and then perform inverse fast Fourier transform and peak detection to obtain the arrival time estimation of multiple channel target signals, and then estimate the distance between the target cooperation end and the center of the USBL array; Step 3: Calculate the frequency domain outputs of the matched filters of multiple channels to obtain the secondary correlation output, then perform band filtering, perform CZT transformation on the secondary correlation function, use the peak value to find the delay difference, and calculate the coordinates of the target cooperation end according to the delay difference; The USBL array is a four-element cross array on a plane, and the four hydrophones are numbered from the first to the fourth in a counterclockwise direction, with the center of the four hydrophone positions as the origin, the line connecting the first and third hydrophones as the x-axis, the line connecting the second and fourth hydrophones as the y-axis, and the direction perpendicular to the xy plane as the z-axis; The step 2 comprises: Perform a fast Fourier transform on the waveform transmitted by the target partner to generate a frequency domain replica signal X rep : X rep =DFT(s(t),N) Wherein, N represents the number of fast Fourier transform points; s(t) represents the waveform transmitted by the target cooperation end; DFT() represents fast Fourier transform; Each frame of data of each channel is obtained with a set multiple of redundancy, and a fast Fourier transform is performed to convert the broadband time domain data into the frequency domain. The frequency domain data of the nth frame is represented as X n , One channel corresponds to one hydrophone; X n With X rep Multiply to obtain the frequency domain result of matched filtering; For each frame of data, the frequency domain output of the matched filter is subjected to inverse fast Fourier transform to obtain the time domain result of the matched filter, and peak detection is performed respectively; Compare the time domain output with the set threshold. If both the positive and negative frequency modulation outputs are greater than the set threshold, it is considered that there is a target echo. Otherwise, it is considered that there is no target echo in this frame, so it is skipped and the next frame is processed. For the case where a target echo is considered to exist, the arrival time of the target signal of the four channels is estimated from the peak position. Estimate the distance between the target partner and the center of the USBL array for: Among them, C USBL represents the speed of sound near the USBL system; The step 3 comprises: For the frame with the target signal, the frequency domain outputs of the matched filter of the 1st and 3rd channels are Z1 and Z3, and Z3 is conjugated and multiplied to obtain the quadratic correlation output Y1=Z1*Z3'; the frequency domain outputs of the matched filter of the 2nd and 4th channels are Z2 and Z4, and Z4 is conjugated and multiplied to obtain the quadratic correlation output Y2=Z2*Z'4; After band filtering, we get the quadratic correlation function Keep the values ​​of the frequency range f1 to f2 where the target signal may exist, and set the other regions to 0; f1 = f c -B / 2, f2=f c +B / 2;f c represents the center frequency of the waveform transmitted by the target cooperation end; B represents the bandwidth of the waveform transmitted by the target cooperation end; Perform CZT on the quadratic correlation function; The CZT transform of the first group of channel correlation functions obtains the cross-correlation delay difference spectrum The peak delay difference is The CZT transform of the second group of channel correlation functions obtains the cross-correlation delay difference spectrum The peak delay difference is Where, m = 0, ..., M-1, M is the number of points of CZT transformation; ω1 = exp(j2π(Δt end1 -Δt begin1 ) / (NM)), j represents the imaginary part; ω2=exp(j2π(Δt end2 -Δt begin2 ) / (NM)), [Δt begin1 ,Δt end1 ] is the range of the delay difference between channel 1 and channel 3, [Δt begin2 ,Δt end2 ] is the range of the delay difference between channel 2 and channel 4; a1=exp(-j*2πΔt begin1 / N), a2=exp(-j*2πΔt begin2 / N); the first group of channels are the 1st and 3rd channels; the second group of channels are the 2nd and 4th channels; Calculate the coordinates of the target partner based on the delay difference: Where d is the distance between each group of hydrophones.

2. The ultra-short baseline positioning method based on CZT cross-correlation delay difference spectrum according to claim 1 is characterized in that: N=4Tf s , where f s is the sampling frequency; T is the pulse width of the transmitted waveform.

3. The ultra-short baseline positioning method based on CZT cross-correlation delay difference spectrum according to claim 1 is characterized in that: The setting multiple is 0.

5.

4. An ultra-short baseline positioning system based on CZT cross-correlation delay difference spectrum, implemented based on any method described in claims 1-3, characterized in that: The system comprises: The front-end processing module is used for the target cooperation end to amplify and automatically gain control the received data; A module for estimating the distance between the target cooperation end and the center of the array is used to perform fast Fourier transform and matched filtering on the data processed by the target cooperation end, and then perform inverse fast Fourier transform and peak detection to obtain the arrival time estimation of multiple channel target signals, and then estimate the distance between the target cooperation end and the center of the USBL array; and The target cooperation end coordinate calculation module is used to calculate the frequency domain output of the matched filter of multiple channels to obtain the secondary correlation output, then perform frequency band filtering, perform CZT transformation on the secondary correlation function, use the peak value to find the delay difference, and calculate the coordinates of the target cooperation end according to the delay difference.

Images