Underwater broadband sound source orientation accurate estimation method and system based on stripe slope blind search

Abstract

The invention belongs to the technical field of underwater acoustic signal direction of arrival estimation, underwater acoustic positioning and sonar, and relates to an underwater broadband sound source direction accurate estimation method and system based on stripe slope blind search, and the method comprises the steps: carrying out the time domain sampling of a signal through employing a horizontal array, recording a waveform, and detecting the arrival time of a reference array element signal; fourier transform is carried out on the received signal of each array element to obtain frequency domain data, and phase compensation is carried out on the frequency domain data by using the arrival time; a cost function is constructed by considering beam forming of the fringe slope changing along with frequency, and blind search is conducted on the fringe slope; and extracting the fringe slope and the sound source azimuth value when the cost function is the maximum value, and realizing the accurate estimation of the underwater broadband sound source azimuth. According to the method, only the average sound velocity in water is needed to serve as environmental information input, and the requirements for priori information such as seabed sediments and sound velocity profiles are avoided; an external image processing algorithm is not needed in the stripe slope extraction process, and the influence of the image processing algorithm on orientation estimation performance is avoided.

Description

Technical Field

[0001] This invention belongs to the fields of underwater acoustic signal arrival estimation, underwater acoustic localization and sonar technology. Specifically, it relates to an accurate estimation method and system for the location of underwater broadband sound sources based on fringe slope blind search. Background Technology

[0002] For underwater acoustic signal azimuth estimation, the most commonly used method is conventional beamforming based on a plane wave model. This method utilizes the phase difference generated by the different positions of array elements of the same frequency signal to match the source azimuth information. However, due to the grazing angle differences of the multipath and multimode components of underwater acoustic waveguides, conventional beamforming of horizontal arrays is prone to problems such as beam shift, broadening, splitting, and gain reduction. The main lobe output position will deviate from the source azimuth, leading to azimuth estimation errors.

[0003] To address the above issues, one solution is to utilize comprehensive environmental information to perform matched-field processing on the received signal, thereby eliminating the influence of waveguide effects and restoring the theoretical beam output performance, thus ensuring the accuracy of azimuth estimation. However, this method relies on accurate prior information about the acoustic field environment, which is often difficult to obtain in practice. Another approach is to extract the slope of a single interference fringe from the position-frequency diagram of the narrowband signal array elements using image processing algorithms. This yields frequency-domain amplitude-consistent frequency-shifted sound pressure data, improving signal correlation and thus reducing beamforming output degradation and increasing the accuracy of sound source azimuth estimation. However, this type of method relies on image processing algorithms, and the accuracy of obtaining the fringe slope is affected by the performance of the image processing algorithm. Furthermore, image processing can only obtain a single fringe slope, applicable to narrowband signals, and does not consider the reality that the fringe slope of broadband signals varies with frequency. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and to propose an accurate estimation method and system for the location of underwater broadband sound sources based on stripe slope blind search.

[0005] In view of this, the present invention proposes an accurate underwater broadband sound source location estimation method based on fringe slope blind search, comprising: Step 1: Use a horizontal array to sample the received signal in the time domain and record the time-domain waveform of the reference array element to detect the arrival time of the signal; Step 2: Obtain frequency domain data by performing Fourier transform on the received signals of each element of the horizontal array, and perform phase compensation on the frequency domain data by using the arrival time of the detected reference array element time domain waveform signal. Step 3: Construct a cost function by considering beamforming where the fringe slope varies with frequency; Step 4: Perform a blind search on the fringe slope, extract the fringe slope and sound source orientation when the cost function reaches its maximum value, and achieve accurate estimation of the underwater broadband sound source orientation.

[0006] As an improvement to the above method, the total number of array elements of the horizontal array in step 1 is I The array length is L , No. The position of each element is ,satisfy The unit is meters, and the sampling time for time-domain sampling is... The total duration of the signal is ,satisfy The unit is seconds, the first The time-domain waveform obtained by each array element is .

[0007] As an improvement to the above method, the reference element in step 1 is the first... Each array element records the time-domain signal for that element. Select the time corresponding to the earliest arriving sample , which is the arrival time of the time domain waveform detection signal.

[0008] As an improvement to the above method, step 2 includes: Perform Fourier transform on the time-domain signals obtained from each element of the horizontal array to obtain the frequency-domain data of the array. ,satisfy For the signal frequency, satisfying Hz, j Represents the imaginary part; detects signal arrival time using time-domain waveforms. For frequency domain data Perform phase compensation to satisfy .

[0009] As an improvement to the above method, the cost function constructed in step 3 for:

[0010] in, From the perspective of searching, it satisfies , It is the slope of the stripes being searched. For the signal frequency, satisfying Hz; [ ] is the frequency band range under study, which meets the requirements. Hz, It is the number of discrete frequency points, satisfying , This is the reference speed of sound, measured in m / s. To process the center frequency of the frequency band, For broadband signals at frequency The slope value of the stripes at that point, For broadband signals at frequency and the position of the array element Frequency offset at that location i Indicates the first Each array element, The data represents the frequency domain of the array, where j represents the imaginary part. I The total number of elements in the horizontal array. The arrival time of the time-domain waveform detection signal. As an improvement to the above method, step 4 includes: Searching for the cost function The slope of the stripes when it reaches its maximum value Sound source location The following equation is satisfied:

[0011] Received This is an accurate estimation of the location of an underwater broadband sound source.

[0012] On the other hand, the present invention provides an accurate underwater broadband sound source location estimation system based on stripe slope blind search, comprising: The time-domain sampling module is used to sample the received signal in the time domain using a horizontal array and record the arrival time of the time-domain waveform of the reference array element. The phase compensation module is used to obtain frequency domain data from the received signals of each element of the horizontal array using Fourier transform, and to perform phase compensation on the frequency domain data using the arrival time of the detected reference array element time domain waveform signal. The cost function construction module is used to construct a cost function by considering beamforming where the fringe slope varies with frequency; The blind search module is used to perform a blind search on the fringe slope, extract the fringe slope and sound source orientation value when the cost function reaches its maximum value, and realize accurate estimation of the underwater broadband sound source orientation.

[0013] Compared with the prior art, the advantages of the present invention are: 1. This invention takes into account the variation of fringe slope with frequency, which is beneficial for processing received signals from underwater broadband sound sources; 2. A blind search process for beamforming is introduced, which uses the maximum beam output to determine the fringe slope of the reference frequency and calculates the frequency shift offset accordingly. This obtains frequency shift sound pressure data with optimal amplitude and phase consistency, eliminating the dependence on external image processing algorithms such as Hough transform in traditional methods. This allows the fringe slope information to be obtained directly through beamforming, avoiding the impact of external image processing algorithms on the orientation estimation performance. 3. By effectively compensating for the phase difference between array signals, the phase consistency of the array signals is restored, thereby achieving accurate estimation of the sound source location. Compared with conventional beamforming methods, it can obtain higher accuracy in sound source location estimation results. 4. Apart from the average sound velocity in water, the present invention does not require knowledge of other sound field environment information, which significantly reduces the dependence on prior environmental information such as seabed sediment and sound velocity profile. Attached Figure Description

[0014] Figure 1 This is the sound velocity profile of the acoustic field environment in the application of the method of the present invention at a sea depth of 3000 m; Figure 2 This is the time-domain waveform of the signal received by the reference array element in the application of the method of the present invention at a sea depth of 3000 m; Figure 3 The present invention describes the element position-frequency domain acoustic field interference fringe structure obtained by the receiving array in a 3000 m deep sea application using the method of the present invention. Figure 4 This is the beam output result obtained by the method of the present invention when searching for different stripe slopes in an application at a sea depth of 3000 m. Figure 5 The method of the present invention is compared with the azimuth estimation obtained by the receiving array using conventional beamforming, single-slope stripe base beamforming, and the method of the present invention in a 3000 m ocean depth application. Figure 6 This is the sound velocity profile of the acoustic field environment in the application of the method of the present invention at a depth of 5000 m. Figure 7 This is the time-domain waveform of the signal received by the reference array element in the application of the method of the present invention at a sea depth of 5000 m; Figure 8 The present invention describes the element position-frequency domain acoustic field interference fringe structure obtained by the receiving array in a 5000 m deep sea application using the method of the present invention. Figure 9 This is the beam output result obtained by the method of the present invention when searching for different stripe slopes in an application at a sea depth of 5000 m. Figure 10 The method of the present invention is compared with the azimuth estimation obtained by the receiving array using conventional beamforming, single-slope stripe base beamforming, and the method of the present invention in a 5000 m ocean depth application. Figure 11 This is a flowchart of the method of the present invention. Detailed Implementation

[0015] This invention proposes an accurate sound source azimuth estimation method and system based on blind search of fringe slope in array element position-frequency map. This method and system consider the variation of fringe slope with frequency in broadband conditions and obtain the fringe slope based on blind search using beamforming, thus eliminating reliance on external image processing algorithms and extensive prior environmental knowledge. First, the received signal is sampled in the time domain using a horizontal array, and the signal arrival time is detected using the time-domain waveform of a reference array element. Second, Fourier transform is used to obtain frequency domain data for the received signals from different array elements, and phase compensation is performed on the frequency domain data using the signal arrival time. Then, a cost function is constructed by beamforming considering the variation of fringe slope with frequency, and a blind search is performed on the fringe slope. The fringe slope at the maximum value of the cost function and the sound source azimuth value are extracted to achieve accurate estimation of the underwater broadband sound source azimuth. This method considers the real-world situation of the fringe slope of underwater broadband sound source signals varying with frequency and is applicable to underwater broadband sound sources. Compared with conventional beamforming, this method can obtain more accurate sound source azimuth estimation results. Moreover, this invention only requires the average sound speed in water as environmental information input, avoiding the need for prior information such as seabed sediment and sound speed profile; the process of extracting the fringe slope does not require external image processing algorithms, thus avoiding the impact of image processing algorithms on the orientation estimation performance.

[0016] This invention provides an accurate method for estimating the azimuth of underwater broadband sound sources based on blind search of element position-frequency map fringe slope, such as... Figure 11 As shown: Step 1: A horizontal array records the received signal, the first... The time-domain waveform obtained by each array element is ,in The total number of array elements; the array length is L , No. i The position of each element is ,satisfy The unit is meters. The signal sampling time is [value], and the total signal duration is [value]. ,satisfy The unit is seconds.

[0017] Step 2: Select the first Each array element serves as a reference element, recording the time-domain signal for that element. Select the time corresponding to the earliest arriving sample The frequency domain data of the array is obtained by performing Fourier transform on the time-domain waveforms obtained from each array element. ,satisfy For the signal frequency, satisfying Hz. For frequency domain data Perform phase compensation to satisfy .

[0018] Step 3: Perform a blind search for the maximum beam output value under different fringe slope values, and construct the cost function through beamforming as follows. in [ ] is the frequency band range under study, which meets the requirements. Hz, It is the number of discrete frequency points, satisfying . This is the reference speed of sound, measured in m / s. The average value of the waveguide sound velocity profile is usually chosen. When this value is unknown, 1500 m / s can be selected. From the perspective of searching, it satisfies . This refers to the slope of the stripes being searched; the search range is typically set to satisfy... Hz / m, the specific value can be modified according to the preliminary judgment of the target distance and orientation in the actual application scenario. To process the center frequency of the frequency band, For broadband signals at frequency The slope value of the stripes at that point, For broadband signals at frequency and the position of the array element Frequency offset at that location.

[0019] Step 4: Search The beam output estimation result when the maximum value is taken satisfies

[0020] in For selection As the center frequency The sound source orientation obtained by beamforming at the fringe slope. That is, the final estimate of the sound source location. That is, the center frequency is directly obtained from the blind search process of beamforming. The final estimated value of the fringe slope. The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0021] Embodiment 1 of this invention proposes an accurate estimation method for the location of underwater broadband sound sources based on fringe slope blind search. This embodiment is an application scenario of a 3000 m deep-sea waveguide with a munk depth. The sound velocity profile is as follows: Figure 1 Seawater density 1.0 g / Seawater absorbs sound silently. The speed of sound on the seabed is 1555 m / s, and the density is 1.8 g / L. Absorption of 0.3 dB / m² in the seabed medium , Represents the wavelength of sound waves; the spectrum of a sound source. =1, sound source depth is 300 m; the aperture of the uniform horizontal receiving array is 1000 m, and the element spacing is... The array is deployed at a depth of 3000 m; the horizontal distance between the sound source and the center point of the receiving array is 7 km. The azimuth of the sound source is 90°. The study focuses on broadband signals in the 30-70 Hz range, and background noise is not considered in the simulation.

[0022] Step 1: Receive the signal from the horizontal array Record it, among which 201; The total signal length is ,satisfy s, the unit is seconds (symbol) s The array element length is... L =500 m, the first The position of each element is = m.

[0023] Step 2: Select the first Each array element serves as a reference element, recording the time-domain signal for that element. Waveform analysis is performed; the time-domain waveform of the reference array element in this embodiment is as follows: Figure 2 As shown. Estimated signal arrival time. =4.70 s. Fourier transforms were performed on the time-domain waveforms obtained from each array element to obtain the array's frequency-domain data. ,satisfy . For the signal frequency, satisfying 70 Hz. Constructing a phase compensation factor. For frequency domain data Perform phase compensation to satisfy The sound pressure amplitude spectrum distributions obtained under different array element positions and signal frequencies are as follows: Figure 3 As shown.

[0024] Step 3: Estimate the fringe slope by constructing a cost function through beamforming and considering the variation of fringe slope with frequency, and select an average sound velocity of 1521 in water. The frequency band under study is used as a reference speed of sound. The discrete frequency is 401, and the reference frequency is 50. Construct a cost function that satisfies From the perspective of searching, it satisfies . Indicates the first The signal of each element is at frequency Frequency offset at that location. The search term is for the slope of the interference fringes at the reference frequency of 50Hz in the element position-frequency plot. The range of slopes being searched satisfies... .

[0025] Step 4: Search for a cost function that makes the cost function... The slope of the stripes when it reaches its maximum value Sound source location These are used as estimates of the fringe slope and the sound source location, respectively, satisfying the following conditions: like Figure 4 As shown, the cost function is The maximum value is taken at 0.0069 Hz / m, which is the fringe slope at the reference frequency of 50 Hz. The fringe slopes at other frequencies satisfy the following condition. Hz / m. Figure 3 The middle dashed line satisfies This gives the frequency shift of the fringes at different element positions. For example... Figure 5 As shown, the maximum beam output value searched by the method of this invention corresponds to a sound source azimuth of 85.0°, which is very close to the true azimuth of 90°; while the sound source azimuth estimate obtained by the conventional beamforming method is 70.0°, which is significantly different from the true sound source azimuth; and the sound source azimuth estimate obtained by the method that does not consider the change of fringe slope with frequency is 74.5°, which is still different from the true azimuth. At the same time, the above processing does not use image processing algorithms to extract fringe slope, nor does it require knowledge of other accurate environmental information (such as seabed sediment) other than the average sound speed of seawater.

[0026] Example 2 Embodiment 2 of the present invention presents an accurate estimation method for the location of an underwater broadband sound source based on fringe slope blind search. This embodiment is an application scenario of a 5000m deep-sea Munk waveguide. The Munk sound velocity profile is as follows: Figure 6 As shown. The horizontal receiving array is deployed at a depth of 5000 m, the horizontal distance between the sound source and the center point of the receiving array is 8 km, and the azimuth of the sound source is 60°. Other simulation parameters are the same as in Example 1.

[0027] Step 1: Receive the signal from the horizontal array Record it, among which 201; The total signal length is ,satisfy s, the unit is seconds (symbol s). The array element length is L =1000 m, the first The position of each element is = .

[0028] Step 2: Select the first Each array element serves as a reference element, recording the time-domain signal for that element. Waveform analysis is performed; the time-domain waveform of the reference array element in this embodiment is as follows: Figure 2 As shown, the estimated signal arrival time is... =5.26 s. Fourier transforms were performed on the time-domain waveforms obtained from each array element to obtain the array's frequency-domain data. ,satisfy . For the signal frequency, satisfying 70 Hz. Arrival time obtained from step 2. Constructing phase compensation factor For frequency domain data Perform phase compensation to satisfy The sound pressure amplitude spectrum distributions obtained under different array element positions and signal frequencies are as follows: Figure 3 As shown.

[0029] Step 3: Estimate the fringe slope by constructing a cost function through beamforming and considering the variation of fringe slope with frequency, and select an average sound velocity of 1521 in water. The frequency band under study is used as a reference speed of sound. The discrete frequency point is 401, and the center frequency is 50. Then the cost function satisfies From the perspective of searching, it satisfies . Indicates the relationship with the first Frequency offset related to each array element. The search is for the slope of the interference fringes in the element position-frequency plot, and the range of slopes being searched satisfies... .

[0030] Step 4: Search for a cost function that makes the cost function... The slope of the stripes when it reaches its maximum value Sound source location These are used as estimates of the fringe slope and the sound source location, respectively, satisfying the following conditions: like Figure 4 As shown, the cost function Searching for stripe slope The maximum value is taken at 0.0044 Hz / m, which is the fringe slope at the reference frequency of 50 Hz. The fringe slopes at other frequencies satisfy the following condition. Hz / m. Figure 3 The middle dashed line satisfies This gives the frequency shift of the fringes at different element positions. For example... Figure 5As shown, the maximum beam output value searched by the method of this invention corresponds to a sound source azimuth of 60.5°, which is very close to the true azimuth of 60°; while the sound source azimuth estimate obtained by the conventional beamforming method is 48.0°, which differs significantly from the true sound source azimuth; and the sound source azimuth estimate obtained by the method that does not consider the change of fringe slope with frequency is 50.5°, which still differs significantly from the true azimuth; at the same time, the above processing does not use image processing algorithms to extract fringe slope, nor does it require knowledge of other accurate environmental information (such as seabed sediment) besides the average sound speed of seawater. Figure 7 The figure shows the time-domain waveform of the signal received by the reference array element in the application of the method of the present invention at a sea depth of 5000 m; Figure 8 The present invention describes the element position-frequency domain acoustic field interference fringe structure obtained by the receiving array in a 5000 m deep sea application using the method of the present invention. Figure 9 This is the beam output result obtained by the method of the present invention when searching for different stripe slopes in an application at a sea depth of 5000 m. Figure 10 The method of the present invention is compared with the azimuth estimation obtained by the receiving array using conventional beamforming, single-slope stripe-based beamforming, and the method of the present invention in a 5000 m ocean depth application.

[0031] Example 3 Embodiment 3 of the present invention provides an accurate underwater broadband sound source location estimation system based on stripe slope blind search, used to implement the above-described method steps, including: The time-domain sampling module is used to sample the received signal in the time domain using a horizontal array and record the arrival time of the time-domain waveform of the reference array element. The phase compensation module is used to obtain frequency domain data from the received signals of each element of the horizontal array using Fourier transform, and to perform phase compensation on the frequency domain data using the arrival time of the detected reference array element time domain waveform signal. Cost function building module for constructing cost functions by considering beamforming where the fringe slope varies with frequency; and The blind search module is used to perform a blind search on the fringe slope, extract the fringe slope and sound source orientation value when the cost function reaches its maximum value, and realize accurate estimation of the underwater broadband sound source orientation.

[0032] It is worth noting that in the embodiments of the above system, the modules included are divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional module are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0033] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention 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 the present invention do not depart from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for accurate estimation of the location of an underwater broadband sound source based on fringe slope blind search, comprising: Step 1: Use a horizontal array to sample the received signal in the time domain and record the time-domain waveform of the reference array element to detect the arrival time of the signal; Step 2: Obtain frequency domain data by performing Fourier transform on the received signals of each element of the horizontal array, and perform phase compensation on the frequency domain data by using the arrival time of the detected reference array element time domain waveform signal. Step 3: Construct a cost function by considering beamforming where the fringe slope varies with frequency; Step 4: Perform a blind search on the fringe slope, extract the fringe slope and sound source orientation values ​​when the cost function reaches its maximum value, and achieve accurate estimation of the underwater broadband sound source orientation.

2. The method for accurate estimation of underwater broadband sound source location based on fringe slope blind search according to claim 1, characterized in that, The total number of array elements in the horizontal array of step 1 is I The array length is L , No. The position of each element is ,satisfy The unit is meters, and the sampling time for time-domain sampling is... The total duration of the signal is ,satisfy The unit is seconds, the first The time-domain waveform obtained by each array element is .

3. The method for accurate estimation of underwater broadband sound source location based on fringe slope blind search according to claim 2, characterized in that, The reference array element in step 1 is the first... Each array element records the time-domain signal for that element. Select the time corresponding to the earliest arriving sample , which is the arrival time of the time domain waveform detection signal.

4. The method for accurate estimation of underwater broadband sound source location based on fringe slope blind search according to claim 3, characterized in that, Step 2 includes: Perform Fourier transform on the time-domain signals obtained from each element of the horizontal array to obtain the frequency-domain data of the array. ,satisfy For the signal frequency, satisfying Hz, j Represents the imaginary part; detects signal arrival time using time-domain waveforms. For frequency domain data Perform phase compensation to satisfy .

5. The method for accurate estimation of underwater broadband sound source location based on fringe slope blind search according to claim 1, characterized in that, The cost function constructed in step 3 for: in, From the perspective of searching, it satisfies , It is the slope of the stripes being searched. For the signal frequency, satisfying Hz; [ ] is the frequency band range under study, which meets the requirements. Hz, It is the number of discrete frequency points, satisfying , This is the reference speed of sound, measured in m / s. To process the center frequency of the frequency band, For broadband signals at frequency The slope value of the stripes at that point, For broadband signals at frequency and the position of the array element Frequency offset at that location i Indicates the first Each array element, The data represents the frequency domain of the array, where j represents the imaginary part. I The total number of elements in the horizontal array. The arrival time of the time-domain waveform detection signal.

6. The method for accurate estimation of underwater broadband sound source location based on fringe slope blind search according to claim 5, characterized in that, Step 4 includes: Searching for the cost function The slope of the stripes when it reaches its maximum value Sound source location The following equation is satisfied: Received This is an accurate estimation of the location of an underwater broadband sound source.

7. A system for accurately estimating the location of an underwater broadband sound source based on fringe slope blind search, characterized in that, include: The time-domain sampling module is used to sample the received signal in the time domain using a horizontal array and record the arrival time of the time-domain waveform of the reference array element. The phase compensation module is used to obtain frequency domain data from the received signals of each element of the horizontal array using Fourier transform, and to perform phase compensation on the frequency domain data using the arrival time of the detected reference array element time domain waveform signal. The cost function construction module is used to construct a cost function by considering beamforming where the fringe slope varies with frequency; and The blind search module is used to perform a blind search on the fringe slope, extract the fringe slope and sound source orientation value when the cost function reaches its maximum value, and realize accurate estimation of the underwater broadband sound source orientation.

Images