A method and system for locating low-frequency broadband pulse signals under unknown environmental information

Abstract

This application provides a method and system for locating low-frequency broadband pulse signals under unknown environmental information. The method includes: obtaining the broadband wavenumber spectrum of the pulse signal using a hypothetical dedispersion formula; obtaining preliminary estimates of each mode using a variational Bayesian method; searching for optimal waveguide invariants and reference sound velocities; re-estimating each mode; and finally, substituting the wavenumber and phase information of each mode into a distance estimation cost function to estimate the distance to the pulse source. The advantages of this application are: by using the dedispersion formula and the variational Bayesian method, mode separation of broadband pulse signals can be achieved even with a short horizontal array aperture; and the phase difference between different modes can be used to estimate the source distance. Through this method, the optimal waveguide invariants and reference sound velocities can be searched under conditions lacking environmental information, reducing the error in source distance estimation.

Description

Technical Field

[0001] This application belongs to the field of marine acoustics, specifically relating to a method and system for locating low-frequency broadband pulse signals under unknown environmental information. Background Technology

[0002] Acoustic passive localization has always been an important problem in shallow water acoustics. Over the past few decades, researchers have proposed many passive localization methods based on the sound field propagation mechanism. These methods mainly include localization methods based on sound field models, normal mode models, and waveguide properties.

[0003] The core idea of ​​matched-field localization is to construct an acoustic field model using ocean environmental parameters, calculate a copy field vector, and match it with the received field to achieve passive target localization. Matched-field localization incorporates an ocean waveguide model, representing a significant advancement over traditional localization methods that do not consider waveguide models. However, in practical applications, due to environmental uncertainties and mismatches between target motion and array manifold vectors, the matched-field output often fails to determine the target's location.

[0004] Matched-field localization methods utilize full-wavelength information, thus requiring ample spatial sampling. Matched-mode methods, based on matched-field methods, separate different modes from the received sound field. Modal information contains the location information of the sound source; selecting stable modes for matching localization can improve the signal-to-noise ratio, reduce the impact of environmental mismatch, and suppress interfering sound sources. Localization methods based on normal mode models all assume modal separability, generally requiring a relatively small number of modes. Mode separation typically requires the array aperture to meet certain distance sampling requirements. To address the issue of insufficient aperture, existing methods can be broadly categorized into two approaches: one is to improve beamforming resolution, and the other is to reduce accuracy requirements, simplifying the estimation problem into a classification problem.

[0005] Furthermore, in shallow sea environments, the interference between different modes causes broadband moving sound sources to exhibit alternating bright and dark fringes in the range-frequency plane. The slopes of these fringes remain essentially constant within a specific waveguide and can be described using waveguide invariants. Based on the range and depth information contained in the slope of the interference structure, it can be used for passive localization of moving sound sources. Localization methods based on waveguide invariants are all built upon a certain signal-to-noise ratio and are more robust to environmental mismatches compared to matched-field methods. However, waveguide invariant-based methods also require prior information about the waveguide invariants, resulting in larger localization errors for broadband pulse signals when the shallow sea environment is unknown. Summary of the Invention

[0006] The purpose of this application is to overcome the shortcomings of existing technologies in terms of large positioning errors of broadband pulse signals when there is unknown shallow sea environment information.

[0007] To achieve the above objectives, this application proposes a method for locating low-frequency broadband pulse signals under unknown environmental information, the method comprising: Step 1: Perform a Fourier transform on the time-domain waveform received by the array sensor to obtain the frequency-domain snapshot signal; Step 2: Using the assumed de-dispersion formula, obtain the broadband discrete wavenumber spectrum pointing to the pulse signal; Step 3: Use the variational Bayes method to obtain preliminary estimation results for each mode; Step 4: Based on the preliminary estimation results of each mode, search for the optimal waveguide invariants and reference sound velocity; Step 5: Substitute the optimal waveguide invariants and reference sound velocity after the search into the de-dispersion formula, and repeat steps 2 to 4 to re-estimate each mode, waveguide invariants and reference sound velocity until the set number of iterations is met. Step 6: Input the wavenumber and phase information of each mode into the cost function of distance estimation to estimate the distance of the pulse sound source.

[0008] As an improvement to the above method, step 1 includes: determining the time length. T Fourier transform of data containing pulse signals: ; in, For the depth of the sound source is Spatial location frequency The received sound pressure field; The radial distance of the receiving point centered on the sound source; The depth of the receiving point; For array element sequence number, J The number of array elements; Indicates the first j The time-domain waveform received by each array element; i It is the imaginary unit.

[0009] As an improvement to the above method, step 2 includes: Based on the estimated reference speed of sound Discrete wavenumber spectrum at the reference frequency for: ; in, ; For the selected reference frequency; For frequency The horizontal wavenumber at that location; Frequency for array reception Narrowband signal vector at the location; Weighted vector In frequency The conjugate transpose at that point. k Indicates the horizontal wavenumber to which the beam output points. For array element j Radial distance difference from the reference position; Vector transpose; For the conjugate transpose; based on the estimated reference speed of sound and waveguide invariants ,frequency f Discrete wavenumber spectrum at ... for: ; in, For frequency The horizontal wavenumber at that location.

[0010] As an improvement to the above method The value is taken as the empirical minimum value of the seawater sound speed in the current sea area; The value is 1; frequency f Less than the reference frequency.

[0011] As an improvement to the above method, step 3 includes: Step 3-1: Initialize the parameters of the gamma distribution and ; Step 3-2: Update the selected frequency band range Approximate probability distribution : ; in, Indicates taking the expected value. Represents the reciprocal of the noise variance; This represents the amplitude corresponding to each horizontal wavenumber. For frequency f First n The amplitude corresponding to each horizontal wavenumber. N The number of equally spaced divisions for the horizontal wavenumber range of the mode to be searched; , indicating that A matrix of elements , For the first m The horizontal wavenumber of the first mode; K The number of values ​​for the horizontal wavenumber discrete grid; Represents the measured wavenumber spectrum; Weighted vector for all wavenumbers The matrix formed; It represents the reciprocal of the prior variance of the amplitude corresponding to each level wavenumber; Step 3-3: Update the selected frequency band range Approximate probability distribution : ; in, Indicates the number of frequency points within the frequency band; Gamma() represents the gamma distribution; Indicates the first n The reciprocal of the prior variance of the amplitude corresponding to each horizontal wavenumber; Steps 3-4: Calculate the noise power using the estimated prior variance of the signal. : ; in, Indicates taking the trace. Indicates the estimated number of modes; This represents the matrix constructed using the estimated modal level wavenumbers; Represents the identity matrix; Step 3-5: Repeat steps 3-2 to 3-4 until... The approximate probability distribution converges; Steps 3-6: According to The peak energy and position are used to obtain preliminary estimates of the modes of each order.

[0012] As an improvement to the above method, step 4 includes: ; in, , For the estimated modes; For the actual reference speed of sound, For real waveguide invariants; This represents the magnitude vector corresponding to each estimated mode.

[0013] As an improvement to the above method, the dispersion removal formula in step 5 is as follows:

[0014] in, Based on the estimated reference speed of sound Calculated wavenumber; .

[0015] As an improvement to the above method, step 6 includes: Cost function for distance estimation for: ; in, Indicates the first m The amplitude of the first mode estimation; , For the first m The horizontal wavenumber of the first mode at the reference frequency; Make Distance to obtain the maximum value r This is the final estimated target distance.

[0016] This application also provides a low-frequency broadband pulse signal positioning system for unknown environmental information, implemented based on the above method, the system comprising: The frequency domain snapshot signal acquisition module is used to perform Fourier transform on the time domain waveform received by the array sensor to obtain the frequency domain snapshot signal; The broadband discrete wavenumber spectrum acquisition module is used to obtain the broadband discrete wavenumber spectrum pointing to the pulse signal using the assumed de-dispersion formula. The module for obtaining preliminary estimation results of each mode order is used to obtain preliminary estimation results of each mode order using the variational Bayes method. The module for obtaining the optimal waveguide invariants and reference sound velocity is used to search for the optimal waveguide invariants and reference sound velocity based on the preliminary estimation results of each mode. The iterative calculation module is used to substitute the optimal waveguide invariants and reference sound velocity after the search into the de-dispersion formula, and repeatedly call the modules for obtaining broadband discrete wavenumber spectrum, obtaining preliminary estimation results of each mode, and obtaining the optimal waveguide invariants and reference sound velocity to re-estimate each mode, waveguide invariants, and reference sound velocity until the set number of iterations is met. The pulse source distance estimation module is used to input the wavenumber and phase information of each mode into the distance estimation cost function to estimate the distance of the pulse source.

[0017] Compared with existing technologies, the advantages of this application are: 1. This invention utilizes the de-dispersion formula and variational Bayesian method to achieve mode separation of broadband pulse signals when the horizontal array aperture is short, and uses the phase difference between different modes to estimate the sound source distance; 2. The method of this invention can search for the optimal waveguide invariants and reference sound velocity under conditions where environmental information is lacking, thereby reducing the error in sound source distance estimation. Attached Figure Description

[0018] Figure 1 The diagram shows a flowchart of a low-frequency broadband pulse signal localization method under unknown environmental information. Figure 2 The figures show the simulation parameters for the marine environment. Figure 3 The diagram shows the array configuration of the underwater acoustic array, which has 128 elements. Figure 4 The results shown are the modal estimation results obtained when the environmental information is known. Figure 5 The figure shows the preliminary modal estimation results obtained when environmental information is unknown; Figure 6 The image shows the final modal estimation results obtained when environmental information is unknown; Figure 7 The figure shows the cost function of the reference sound velocity and waveguide invariants when the environmental information is unknown; Figure 8 The image shows the sound source distance estimation results obtained when environmental information is unknown. Detailed Implementation

[0019] The technical solution of this application will be described in detail below with reference to the accompanying drawings.

[0020] Example 1 This invention proposes a method for locating low-frequency broadband pulse signals under unknown environmental information, based on the shallow sea waveguide dispersion assumption and combined with variational Bayesian theory. The method includes: Step 1: Perform Fourier transform on the time-domain waveform received by the array sensor to obtain the frequency-domain snapshot signal; Step 2: Using the assumed de-dispersion formula, obtain the broadband discrete wavenumber spectrum pointing to the pulse signal; Step 3: Use the variational Bayes method to obtain preliminary estimation results for each mode; Step 4: Based on the preliminary estimation results of each mode, search for the optimal waveguide invariants and reference sound velocity; Step 5: Substitute the optimal waveguide invariants and reference sound velocity after the search into the de-dispersion formula, and repeat steps 2 to 4 to re-estimate each mode, waveguide invariants and reference sound velocity until the number of iterations is satisfied. Step 6: Input the wavenumber and phase information of each mode into the cost function of distance estimation to estimate the distance of the pulse sound source.

[0021] According to the normal mode model, when the sound source depth is... Spatial location frequency The received sound pressure field is (1) in, The radial distance of the receiving point centered on the sound source. For the depth of the receiving point, For array element sequence number, J The number of array elements , The first m The eigenfunctions and level wavenumbers of the first mode. iThe unit is the imaginary number. For a horizontal array receiving signal, after conventional beamforming, its beam output is: (2) in, For the first m First mode amplitude, For weighted vectors, k Indicates the horizontal wavenumber to which the beam output points. For vector transpose, It is the conjugate transpose. The distance from the reference position to the sound source. For array element j The radial distance difference from the reference position. Let: (3) The wavenumber spectrum can be expressed as: (4) The biggest challenge in separating modes in the wavenumber domain using horizontal array beamforming results is that insufficient array aperture makes it difficult to meet the resolution requirements for mode separation. Variational Bayesian estimation, a technique in compressed sensing, has the advantages of high resolution and the ability to separate coherent signals. To extend the original model using compressed sensing theory, the horizontal wavenumber range of the mode to be searched is first divided into equal intervals. N Parts, forming , indicating that A matrix of elements K The number of values ​​for the discrete grid of the horizontal wavenumbers is given, and the amplitude corresponding to each horizontal wavenumber becomes... The new wavenumber spectrum model is: (5) At this point, the magnitude vector is sparse. Assume the noise mean is zero and the variance is... For wavenumber spectra, there is a likelihood function (6) In the formula, CN represents a complex Gaussian distribution. It is a weighted vector of all wavenumbers. The matrix formed. Prior obedience to Gamma distribution with parameters: (7) signal vector Follow the Student-t stratified prior: (8) for The prior variance, and follows the principle of... The gamma distribution with parameter is: (9) Using the variational Bayesian estimation method, the log-likelihood function It can be divided into two parts (10) in, (11) (12) Let be any probability density function. Let KL represent the KL divergence between two probability density functions. When the KL divergence is 0, ... It equals the posterior probability density. Therefore This is the lower bound of the log-likelihood function. Solving for the posterior probability becomes making The biggest optimization problem, therefore, needs to be found Make Maximum. According to mean-field theory, assuming... The following decomposition can be performed: (13) Substituting into equation (11), we can obtain the maximum value. of , (14) Therefore, the variational Bayesian method involves iteratively performing the above equation on the corresponding N+1 steps to obtain the maximization. of .

[0022] Considering broadband signals, due to dispersion, the same mode will undergo nonlinear changes with frequency in the wavenumber domain. If modes at different frequencies are at the same wavenumber position, then the broadband signal... This can be viewed as multiple samplings conforming to the same probability distribution. Therefore, aligning wavenumbers of different frequencies during beamforming transformation to the wavenumber domain can also be called de-dispersion. The transformation formula is: (15) in, Based on the estimated reference speed of sound Calculated wavenumber, For the selected reference frequency, For the predicted waveguide invariants, The new wavenumber spectrum output is (16) in, For the actual reference speed of sound, For true waveguide invariants, , For the first m The horizontal wavenumber of the first mode at the reference frequency, Equation (16) can be approximated as: (17) The constant coefficient, For the array aperture, different frequencies are... The extreme value will be obtained, and the interval between the extreme values ​​corresponding to different modes is: The error in the wavenumber difference between different modes is only related to the error in the waveguide invariants. Therefore, after obtaining the estimated beam positions for each mode, the dispersion curve is shifted and a search is performed. The reference sound velocity and waveguide invariant with the minimum reconstruction error are obtained, and then a new round of variational Bayes estimation is performed using the new reference sound velocity and waveguide invariant to obtain the final estimation result.

[0023] After obtaining the dispersion structure, the corresponding modal amplitudes can also be obtained, and the cost function for constructing the target distance is: (18) Based on the above theoretical analysis, this invention proposes a method for locating low-frequency broadband pulse signals under unknown environmental information, comprising the following steps: Step 1: Perform a Fourier transform on the time-domain waveform received by the array sensor to obtain the frequency-domain snapshot signal.

[0024] The specific process is as follows: Regarding the time length T Fourier transform of data containing pulse signals:

[0025] In the formula, Indicates the first j The time-domain waveform received by each array element.

[0026] Step 2: Using the assumed de-dispersion formula (Equation (15)), obtain the broadband discrete wavenumber spectrum pointing to the pulse signal.

[0027] The specific process is as follows: based on the estimated reference speed of sound The discrete wavenumber spectrum at the reference frequency is:

[0028] in, Based on the estimated reference speed of sound and waveguide invariants ,frequency f The discrete wavenumber spectrum at point is:

[0029] Among them, frequency f Horizontal wavenumber at , It is the narrowband signal vector received by the array. yes In frequency f The conjugate transpose at that point.

[0030] Preferred, The typical value is the empirical minimum value of the sound speed of seawater in the current sea area.

[0031] Preferred, The typical value is 1.

[0032] Preferred frequency f It should be less than the reference frequency.

[0033] Step 3: Use the variational Bayes method to obtain preliminary estimation results for each mode.

[0034] The specific process is as follows: (1) Initialize the parameters of the gamma distribution and This parameter is usually chosen based on experience; in this embodiment, it can be taken as (1). -3 1 -3 ) and (1 -6 1 -3 ); (2) Update the selected frequency band range Approximate probability distribution;

[0035] in, Indicates taking the expected value. This represents the reciprocal of the noise variance. , , For frequency f First n The amplitude corresponding to each horizontal wavenumber. Represents the measured wavenumber spectrum; (3) Update the selected frequency band range Approximate probability distribution;

[0036] in Indicates the number of frequency points within the frequency band. for The prior variance; Gamma() represents the gamma distribution; It represents the reciprocal of the prior variance of the amplitude corresponding to each level wavenumber; Indicates the first n The reciprocal of the prior variance of the amplitude corresponding to each horizontal wavenumber.

[0037] (4) Calculate the noise power using the estimated prior variance of the signal. ;

[0038] in Indicates taking the trace. This represents the estimated number of modes. The matrix is ​​constructed using the estimated modal level wavenumbers. ; (5) Repeat (2) to (4) until The approximate probability distribution converges; (6) According to The peak energy and position are used to obtain preliminary estimates of the modes of each order.

[0039] Step 4: Based on the preliminary estimation results of each mode, search for the optimal waveguide invariants and reference sound velocity.

[0040] The specific process is as follows:

[0041] in, , For the estimated modes, This represents the magnitude vector corresponding to each estimated mode.

[0042] Step 5: Substitute the optimal waveguide invariants and reference sound velocity obtained from the search into the de-dispersion formula, and repeat steps 2 to 4 to re-estimate each mode, waveguide invariants, and reference sound velocity until the required number of iterations is met.

[0043] Step 6: Input the wavenumber and phase information of each mode into the cost function of distance estimation to estimate the distance of the pulse sound source.

[0044] The specific process is as follows: the cost function for distance estimation for:

[0045] Make Distance to obtain the maximum value r This is the final estimated target distance. Indicates the first m The amplitude of the first mode estimation.

[0046] The technical solution of the present invention will now be illustrated with reference to the accompanying drawings and specific examples.

[0047] Suppose a target detection operation is being conducted in a certain sea area, and the marine environment is as follows: Figure 2 As shown, an acoustic array is deployed on the seabed, with the array pattern as follows. Figure 3 As shown, the array has 128 elements. The received signal bandwidth is 30-100Hz, with a total of 351 frequency points, assuming random phase at each frequency point. The estimated reference sound velocity is 1525m / s, and the waveguide invariant is 1. The signal processor will process the signal according to the specific steps described above. Figure 4 The modal estimation results are obtained when the environmental information is known. The black dashed line is the true modal dispersion curve. It can be seen that the estimated dispersion curve is basically consistent with the true value. Figure 5 The results are preliminary modal estimations obtained when environmental information is unknown, and the dispersion curves differ significantly from the true values. Figure 6 The final modal estimation result is obtained when the environmental information is unknown. The reference sound velocity and waveguide invariant dispersion curve after the search are close to the true value. Figure 7 This is the cost function distribution obtained after searching for two parameters; Figure 8 The sound source distance estimation result obtained under unknown environmental information is consistent with the actual distance of 20 kilometers.

[0048] Example 2 This application also provides a low-frequency broadband pulse signal positioning system for unknown environmental information, implemented based on the above method, the system comprising: The frequency domain snapshot signal acquisition module is used to perform Fourier transform on the time domain waveform received by the array sensor to obtain the frequency domain snapshot signal; The broadband discrete wavenumber spectrum acquisition module is used to obtain the broadband discrete wavenumber spectrum pointing to the pulse signal using the assumed de-dispersion formula. The module for obtaining preliminary estimation results of each mode order is used to obtain preliminary estimation results of each mode order using the variational Bayes method. The module for obtaining the optimal waveguide invariants and reference sound velocity is used to search for the optimal waveguide invariants and reference sound velocity based on the preliminary estimation results of each mode. The iterative calculation module is used to substitute the optimal waveguide invariants and reference sound velocity after the search into the de-dispersion formula, and repeatedly call the modules for obtaining broadband discrete wavenumber spectrum, obtaining preliminary estimation results of each mode, and obtaining the optimal waveguide invariants and reference sound velocity to re-estimate each mode, waveguide invariants, and reference sound velocity until the set number of iterations is met. The pulse source distance estimation module is used to input the wavenumber and phase information of each mode into the distance estimation cost function to estimate the distance of the pulse source.

[0049] This application may also provide a computer device, including: at least one processor, memory, at least one network interface, and a user interface. The various components in this device are coupled together via a bus system. It is understood that the bus system is used to implement communication between these components. In addition to a data bus, the bus system also includes a power bus, a control bus, and a status signal bus.

[0050] The user interface can include a display, keyboard, or clicking device. Examples include a mouse, trackball, touchpad, or touchscreen.

[0051] It is understood that the memory in the embodiments disclosed in this application may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory may be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM). The memories described herein are intended to include, but are not limited to, these and any other suitable types of memory.

[0052] In some implementations, the memory stores elements such as executable modules or data structures, or subsets thereof, or extended sets thereof: operating systems and applications.

[0053] The operating system includes various system programs, such as the framework layer, core library layer, and driver layer, used to implement various basic business functions and handle hardware-based tasks. The application programs include various applications, such as media players and browsers, used to implement various application functions. Programs implementing the methods of the embodiments of this disclosure can be included in the application programs.

[0054] In the above embodiments, the processor can also invoke programs or instructions stored in memory, specifically programs or instructions stored in an application program, for the following purposes: Follow the steps described above.

[0055] The above methods can be applied to or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above methods can be completed by integrated logic circuits in the processor's hardware or by software instructions. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic diagrams disclosed above. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the disclosed methods can be directly implemented by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above methods.

[0056] It is understood that the embodiments described in this application can be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in this application, or combinations thereof.

[0057] For software implementation, the technology of this application can be implemented by executing the functional modules (e.g., procedures, functions, etc.) of this application. The software code can be stored in memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0058] This application may also provide a non-volatile storage medium for storing a computer program. When the computer program is executed by a processor, it can implement the steps in the above method embodiments.

[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to the embodiments, those skilled in the art should understand that modifications or equivalent substitutions to the technical solutions of this application do not depart from the spirit and scope of the technical solutions of this application, and should all be covered within the scope of the claims of this application.

Claims

1. A low-frequency broadband pulse signal positioning method in an unknown environment, the method comprising: Step 1: Fourier transforming a time-domain waveform received by an array sensor to obtain a frequency-domain snapshot signal; Step 2: using a hypothetical de-dispersion formula to obtain a broadband discrete wave number spectrum of a directional pulse signal; Step 3: using a variational Bayesian method to obtain a preliminary estimation result of each order mode; Step 4: searching for optimal waveguide invariants and a reference sound speed according to the preliminary estimation result of each order mode; Step 5: inputting the searched optimal waveguide invariants and the reference sound speed into the de-dispersion formula, and repeating Steps 2 to 4 to re-estimate each order mode, the waveguide invariants and the reference sound speed until a set number of iterations is met; Step 6: inputting wave number and phase information of each order mode into a distance estimation cost function to estimate the distance of a pulse sound source.

2. The method of claim 1, wherein, The step 1 includes: performing Fourier transform on the data of the time length T performing Fourier transform on the data containing the pulse signal ； wherein, is the depth of the sound source, is the spatial position of the sound source, is the frequency of the received sound pressure field at the spatial position ; is the radial distance of the receiving point from the sound source; is the depth of the receiving point; is the array element number, J is the number of array elements; represents the time-domain waveform received by the j th array element; i is the imaginary unit.

3. The method of claim 2, wherein, The Step 2 comprises: According to the estimated reference sound speed , the discrete wave number spectrum at the reference frequency is: ； in, ; For the selected reference frequency; For frequency The horizontal wavenumber at that location; Frequency for array reception Narrowband signal vector at the location; Weighted vector In frequency The conjugate transpose at that point. k Indicates the horizontal wavenumber to which the beam output points. For array element j Radial distance difference from the reference position; Vector transpose; For the conjugate transpose; based on the estimated reference speed of sound and waveguide invariants ,frequency f Discrete wavenumber spectrum at ... for: ； wherein is the frequency is the horizontal wave number at the location 4. The low-frequency wideband pulse signal positioning method in an unknown environment according to claim 3, characterized in that, is equal to the empirical minimum value of the sea water sound speed in the current sea area; is equal to 1; the frequency f is less than the reference frequency.

5. The method of claim 3, wherein, The Step 3 comprises: Step 3-1: Initialization of parameters of the gamma distribution and ; Step 3-2: Update the approximate probability distribution over the selected band range :​ ； wherein, represents the inverse of the expectation, represents the inverse of the noise variance; represents the amplitude corresponding to the respective horizontal wavenumber, is the frequency f is the amplitude corresponding to the respective horizontal wavenumber, n is the amplitude corresponding to the respective horizontal wavenumber, N is the number of partitions of the range of modal horizontal wavenumbers to be searched; represents a matrix with as elements, , is the horizontal wavenumber of the modal of order m ; K is the number of values of the discrete grid of horizontal wavenumbers; represents the measured wavenumber spectrum; is the matrix composed of the weighting vectors for all wavenumbers; represents the inverse of the prior variance of the amplitude corresponding to the respective horizontal wavenumber; Step 3-3: Update the approximate probability distribution over the selected band range :​ ； wherein, represents the number of frequency points within the frequency band; Gamma() represents a gamma distribution; represents the inverse of the prior variance of the amplitude corresponding to the n th horizontal wave number. Step 3-4: Compute noise power using the estimated signal prior variance : ； wherein, denotes the trace, denotes the estimated number of modes; denotes the matrix constructed with the estimated modal wavenumbers; denotes the identity matrix; Step 3-5: Repeat Step 3-2 to Step 3-4 until the approximate probability distribution of converges; Step 3-6: Obtain the preliminary estimation of each order mode according to the peak energy and location of the peak energy and location of 6. The method of claim 5, wherein, The Step 4 comprises: ； wherein, , are the estimated modal orders; is the true reference sound speed, is the true waveguide invariants; denotes the estimated amplitude vector corresponding to the estimated modal orders.

7. The method of claim 6, wherein, The de-dispersion formula in the Step 5 is: wherein is the estimated reference sound speed calculated wave number; .

8. The method of claim 7, wherein, The Step 6 comprises: Cost function for distance estimation is: ； wherein, represents the amplitude of the estimated mode of order m , represents the horizontal wave number of the mode of order m at the reference frequency.​ such that the distance that attains a maximum r is the final estimated target distance.

9. A low-frequency wideband pulse signal positioning system in unknown environment information, based on the method of any one of claims 1-8, characterized in that, The system comprises: a frequency-domain snapshot signal acquisition module configured to Fourier transform a time-domain waveform received by an array sensor to obtain a frequency-domain snapshot signal; a broadband discrete wave number spectrum acquisition module configured to use a hypothetical de-dispersion formula to obtain a broadband discrete wave number spectrum of a directional pulse signal; a preliminary estimation result of each order mode acquisition module configured to use a variational Bayesian method to obtain a preliminary estimation result of each order mode; an optimal waveguide invariant and reference sound speed acquisition module configured to search for optimal waveguide invariants and a reference sound speed according to the preliminary estimation result of each order mode; an iterative calculation module configured to input the searched optimal waveguide invariants and the reference sound speed into the de-dispersion formula, repeatedly call the broadband discrete wave number spectrum acquisition module, the preliminary estimation result of each order mode acquisition module and the optimal waveguide invariant and reference sound speed acquisition module, and re-estimate each order mode, the waveguide invariants and the reference sound speed until a set number of iterations is met; and a pulse sound source distance estimation module configured to input wave number and phase information of each order mode into a distance estimation cost function to estimate the distance of a pulse sound source.

Images