CN116203505A - Orthogonal matching pursuit sound source identification method and device based on block sparse Bayes - Google Patents
Orthogonal matching pursuit sound source identification method and device based on block sparse Bayes Download PDFInfo
- Publication number
- CN116203505A CN116203505A CN202310187866.7A CN202310187866A CN116203505A CN 116203505 A CN116203505 A CN 116203505A CN 202310187866 A CN202310187866 A CN 202310187866A CN 116203505 A CN116203505 A CN 116203505A
- Authority
- CN
- China
- Prior art keywords
- sound source
- sound
- source
- equivalent
- algorithm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 85
- 238000005259 measurement Methods 0.000 claims abstract description 81
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 78
- 238000012546 transfer Methods 0.000 claims abstract description 73
- 230000005405 multipole Effects 0.000 claims abstract description 51
- 230000005855 radiation Effects 0.000 claims abstract description 18
- 230000005404 monopole Effects 0.000 claims description 28
- 230000006870 function Effects 0.000 claims description 18
- 239000013598 vector Substances 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 15
- 238000009826 distribution Methods 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 7
- 230000005540 biological transmission Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 abstract description 7
- 238000012800 visualization Methods 0.000 abstract description 2
- 238000013398 bayesian method Methods 0.000 description 8
- 230000015654 memory Effects 0.000 description 6
- 238000004088 simulation Methods 0.000 description 5
- 238000003860 storage Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012216 screening Methods 0.000 description 4
- 238000001228 spectrum Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000004807 localization Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 230000001427 coherent effect Effects 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000000354 decomposition reaction Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000005236 sound signal Effects 0.000 description 2
- UJCHIZDEQZMODR-BYPYZUCNSA-N (2r)-2-acetamido-3-sulfanylpropanamide Chemical compound CC(=O)N[C@@H](CS)C(N)=O UJCHIZDEQZMODR-BYPYZUCNSA-N 0.000 description 1
- 241001669680 Dormitator maculatus Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/18—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using ultrasonic, sonic, or infrasonic waves
- G01S5/24—Position of single direction-finder fixed by determining direction of a plurality of spaced sources of known location
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses an orthogonal matching pursuit sound source identification method and device based on block sparse Bayes, and relates to the technical field of pneumatic noise field identification and sound field visualization. Comprising the following steps: setting a measuring surface parallel to a sound source surface in a measured sound field of a sound source to be identified; arranging a plurality of equivalent sound sources at the position of a sound source surface, predefining multipole sound source radiation models corresponding to different axial directions for each equivalent source, and establishing a transfer relationship between the equivalent sound source surface and a measurement surface; and obtaining a sound source identification result of the measured sound field of the sound source to be identified according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching pursuit algorithm. The invention has the advantages of convenient realization, suitability for measuring surfaces with arbitrary shapes, good calculation stability and high recognition precision.
Description
Technical Field
The invention relates to the technical field of pneumatic noise field recognition and sound field visualization, in particular to a block sparse Bayesian-based orthogonal matching tracking sound source recognition method and device.
Background
With the intensive development of pneumatic noise related research work, more and more scholars find that although the beam forming technology is mature in the measurement aspect of pneumatic acoustics, due to the fact that pneumatic sound sources generally have the radiation characteristic of strong directivity, the traditional beam forming technology based on monopole source assumption is difficult to obtain the correct distribution of the sound sources of the complex pneumatic sound field. Therefore, aiming at the problem of multi-pole sound source identification such as dipoles, quadrupoles and the like, a plurality of students have developed corresponding research works at home and abroad.
The learner can realize the identification of the sound sources with different directivities by decomposing the multipole sound source into the superposition of the sound source modes with different directions. In 2005, takao Suzuki proposed an adaptive beamforming algorithm for identifying low mach number jet noise sources. Based on the sound pressure data measured by the conical microphone array, carrying out Fourier decomposition on the sound pressure signal of the microphone according to azimuth angles, extracting main feature vectors of each azimuth angle radiation mode, then decomposing the coherent sound signal according to multipole modes under spherical coordinates, and identifying the shape and axial position of the first two azimuth angle modes. The beam forming algorithm is used for measuring jet noise of a low Mach number circular nozzle, and two axisymmetric modes are identified, wherein quadrupole sources account for the main part. However, this method requires that all multipoles have coaxial symmetry, is only applicable to low frequency sound sources with strong far-field directivity, and requires that the sound source be stationary with a high signal-to-noise ratio. In 2008, bouchard proposed a beamforming method for identifying a directed sound source, which defines a set of "sub-beamformers", each for a different spatial mode of the response source. The results show that combining the outputs of the individual sub-beamformers into a weighted sum yields an optimal array gain that results in better recognition than conventional monopole sub-beam forming methods. However, this approach requires an artificial determination of the estimated spherical harmonic mode order based on the expected complexity of the source, and requires more parameters to be estimated for harmonic decomposition of the same order than for the traditional point source model. In 2018, matvey et al generalized a conventional beamforming algorithm to a dipole sound source, and the core idea was to decompose each discrete point into two incoherent dipole sources in mutually perpendicular directions, assuming that the sound source is located in a plane parallel to the microphone array, seeking a source solution corresponding to minimizing the cross-spectral matrix. The algorithm is experimentally verified for buzzer, cylindrical bypass noise and jet plate interaction noise. However, this method is not suitable for recognizing a complex sound source structure such as noise generated by a real airplane, etc., and has a certain requirement for the number of microphones.
In the process of identifying the multipole sound source, if the multipole axis direction can be estimated in advance according to an actual pneumatic sound source model, the multipole characteristic item can be applied to carry out correction, namely a multipole directivity function is added into the beam forming guiding quantity, and a multipole sub-beam forming algorithm corrected based on the transfer relation is established. In 2008, liu et al proposed a beam forming correction method for identifying a dipole source, which produced a method suitable for identifying a dipole source by modifying a signal propagation model of beam forming. Numerical simulation and actual measurement experiments are performed, and the ability of the method to identify dipole sources is verified. In 2015, ric ports proposed and compared four beamforming algorithms for precisely locating dipole sound sources in the spatial domain, including conventional cross-spectrum beamforming, CLEAN-SC (Clean based on Source Coherence, space-coherence-cancellation sound source) deconvolution beamforming, multiplicative cross-spectrum beamforming, and CLEAN-SC-based multiplicative beamforming. Where multiplicative beamforming relies on mutual cancellation between orthogonally arranged microphone arrays to improve the quality of the source map, i.e. to improve the depth resolution of one array based on the lateral resolution of the other. In 2020, zhou et al performed three-dimensional aerodynamic noise source identification based on MEMS microphone tunnel arrays, the sound source was generated by wall-mounted NACA 0012 airfoils in the wind tunnel, and data processing was performed based on the three-dimensional deconvolution method (CLEAN-SC) and beamforming techniques associated with dipole radiation models perpendicular to the airfoils. In the third octave band studied, the sound source recognition results are very consistent with experimental results published in the literature. In 2022, chen proposes a moving dipole source beam forming method based on a framework of a wavelet beam forming method aiming at the problem of three-dimensional noise source identification of a high-speed rotating propeller. The flow field and the noise field are obtained through numerical simulation, and then the physical solution of the source position and the source direction is obtained through acoustic imaging analysis, so that the capability of the method for solving the actual problem is verified. However, the above methods are only suitable for recognizing dipole sound sources, and it is necessary to know the axis direction of the dipole in advance as an input parameter. Although the method can be widely applied to quadrupole sources, the directivity information of all multipole sound sources still needs to be known in advance, and the reconstruction capability of the multipole sound sources with different directivities is greatly different, if the multipole direction is judged to be incorrect, a great phase error is caused, and therefore a completely incorrect sound source identification result is caused.
Aiming at the problem, a learner proposes to build an acoustic transfer relation model between each channel data of the microphone array and the multi-type sound source by predefining possible types of the sound source, and further solves an overdetermined equation set containing the sound source intensity and directivity information by using an inverse method, so that the intensities respectively corresponding to the multi-pole sound source can be obtained. In 2011, takao was based on L 1 The generalized inverse technology develops a multipole pneumatic sound source recognition algorithm. Firstly, predefining a sound source model comprising monopoles and multipoles, decomposing a cross-spectrum matrix into eigenmodes, extracting each coherent signal, and then using an iterative re-weighted least square method to take source distribution as L 1 And solving a norm problem to obtain amplitude distribution corresponding to each eigenmode. The proposed algorithm is checked by using the airfoil model test data in the wind tunnel, and the relevant research conclusion of the jet-flap interaction is obtained. However, the sound source model which can be distinguished by the method only comprises monopole and dipole sources, the quadrupole sound source is not considered, and only the sound source with the sound source surface parallel to the microphone array can be identified. For example, a dipole whose axis direction is perpendicular to the microphone array direction cannot be distinguished, which may be misidentified as a monopole. In 2016, madoliat also predefines the placement of different types of sound sources, and solves for their corresponding intensities by the inverse method. If it is known in advance Some type of sound source does not participate in sound production, it can be removed from the arrangement. By optimizing the arrangement of different types of sound sources, the mismatch problem between the measured value and the reconstructed value can also be reduced. However, the above method also requires that the measurement plane is parallel to the sound source plane, and that the axis of the dipole sound source is only on the sound source plane, otherwise a reconstruction result of misleading sound source characteristics may be obtained. In addition, because of the high probability of pre-assumption of source type, the inversion process is a serious underdetermined problem, and the small measurement error easily leads to failure of sound source identification. In 2020, gao et al propose a beamforming method based on dipole assumptions, where the dipole source can be located without prior knowledge of the source direction, the position of the dipole source is determined by calculating the beamforming results at predefined directions and positions using a propagation function based on the dipole, and the final beamforming result at each scan point is determined by the maximum at the predefined direction. Numerical simulations and experiments were performed on the rotating dipole source to obtain satisfactory dipole source positioning results in different directions. It can be seen that the dipole-based beamforming method can accurately locate the dipole sound source, whereas the conventional beamforming method splits the dipole source into two sound sources, resulting in erroneous decisions. However, the method can only realize sound source localization, does not further judge directivity and other information of the sound source, is limited by the nature of a beam forming algorithm, and is only suitable for identifying high-frequency noise.
In consideration of radiation directivity of complex pneumatic sound sources, students also directly utilize the radiation property of multipole sub-sources to identify and judge multipole sub-sound sources. In 2016, mimani et al used the aeroacoustic time reversal method in wind tunnels to improve the identification resolution of dipole sources. A two-dimensional cylinder flow induced noise identification experiment in the uniform cross flow wind tunnel is performed, and the effectiveness of the method in the aspect of improving the pneumatic noise source identification resolution is proved. In the sound source identification result of the method, two instantaneous maximum value areas representing the sound source appear near the cylinder axis, the intensities are almost the same, but the phases are opposite, so as to represent the dipole source. However, this method is equally applicable only to identifying dipole sources with axes parallel to the array, and is not applicable to identifying more complex aerodynamic noise structures. And the scholars can eliminate the influence of the directivity of the sound source on the sound source positioning and identification by carrying out transformation processing on the array sound signal cross spectrum matrix. In 2019, pan et al have proposed a multipole sub-beam forming algorithm that incorporates the inverse method for detecting multipole sub-sources with different radiation directions. Source directivity is removed by normalizing and squaring elements of the feature vector to achieve sound source localization. Each source is represented by spherical harmonics after positioning, from which the type and direction of the source can be deduced. The algorithm was validated using numerical simulations and experiments of speaker and cylinder deswirle. Because the method only reserves the phase information in the measured data, the robustness under the test conditions of different working conditions needs to be further improved.
In summary, the present multipole sound source identification method still has limitations. The existing high-resolution wave beam forming method can well realize real-time accurate imaging of monopole sound sources. However, for dipole or quadrupole sound sources with strong directivity characteristics, part of methods need to acquire priori knowledge of sound fields in advance, so that sound field transfer functions are adjusted according to directivity information of the multipole sound sources, and stability and robustness of the method in inversion are poor, so that accurate sound source position information is difficult to acquire. In addition, most of the existing multipole sound source identification methods are proposed for dipole sound sources, and for quadrupole sound sources, only a single low-frequency sound source with obvious far-field directivity can be analyzed. Therefore, the multipole sound source identification method has the advantages of good robustness, wide application range and high calculation precision, and has higher theoretical and engineering significance.
Disclosure of Invention
The invention provides the multi-pole sound source identification method aiming at the problems of how to provide the multi-pole sound source identification method with good robustness, wide application range and high calculation precision.
In order to solve the technical problems, the invention provides the following technical scheme:
in one aspect, the invention provides a block sparse Bayesian-based orthogonal matching pursuit sound source identification method, which is implemented by electronic equipment and comprises the following steps:
S1, setting a measuring surface parallel to a sound source surface in a measured sound field of a sound source to be identified.
S2, arranging a plurality of equivalent sound sources at the position of the sound source surface, and establishing a transfer relation between the equivalent sound source surface and the measuring surface.
And S3, obtaining a sound source identification result of the measured sound field of the sound source to be identified according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching pursuit algorithm.
Optionally, in the measured sound field of the sound source to be identified in S1, a measurement plane parallel to the sound source plane is set, including:
in a radiation sound field of a multipole sound source of a sound source to be identified, a measurement surface parallel to the sound source surface is provided, the measurement surface being an archimedes spiral array comprising a plurality of sound pressure sensors.
Optionally, disposing a plurality of equivalent sound sources at positions of the sound source surface in S2 and establishing a transfer relationship between the equivalent sound source surface and the measurement surface includes:
s21, collecting sound pressure signals on the measuring surface.
S22, arranging a plurality of equivalent sound sources at the positions of the sound source surface according to the sound pressure signals.
S23, predefining multipole sub-sound source radiation models corresponding to different axial directions for each equivalent sound source in the plurality of equivalent sound sources, and establishing a transfer relation between an equivalent sound source surface and a measurement surface.
Optionally, establishing a transfer relationship between the equivalent sound source surface and the measurement surface in S2 includes:
predefining a sound source type and a target domain, calculating a transmission matrix, and establishing a transfer relation between an equivalent sound source surface and a measurement surface, wherein the transfer relation is shown in the following formula (1):
wherein G is mon ，G dip ，
Representing the free field green's function, q, between the equivalent sound source and the measurement surface of monopole, dipole, longitudinal quadrupole, transverse quadrupole, respectively mon ，q dip ，/>
Respectively representing the source strengths corresponding to the monopole sound source, the dipole sound source, the longitudinal quadrupole sound source and the transverse quadrupole sound source, wherein n is the number corresponding to the equivalent sound sources with different sound source types, and n is mon ，n dip ，/>
Is the total number of equivalent sound sources corresponding to monopole sound sources, dipole sound sources, longitudinal quadrupole sound sources and transverse quadrupole sound sources, P m To measure the surface acoustic pressure.
Wherein G is mon ，G dip ，
The expression of (2) to (5) below:
wherein r is the spatial distance between the measuring point position and the equivalent sound source position, and k is the wave of the analysis frequencyThe number i is an imaginary number unit, r is the spatial distance between the equivalent sound source and the measuring point position, and theta d Is the included angle between the normal vector of the dipole axis and the vector of the sound source pointing to the measuring point position, theta qlo Is the included angle between the transverse axis of the longitudinal quadrupoles and the vector of the equivalent sound source pointing to the measuring point position, 
Is the included angle between the longitudinal axis of the longitudinal quadrupole and the measuring point, theta qla Is the included angle between the axis of the transverse quadrupoles and the measuring point.
Optionally, obtaining a sound source identification result of the measured sound field of the sound source to be identified according to the transfer relationship between the equivalent sound source surface and the measurement surface, the block sparse bayesian algorithm and the orthogonal matching pursuit algorithm in S3 includes:
s31, obtaining the source strengths of different sound source types corresponding to each axis in different axes according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching tracking algorithm.
S32, superposing the source intensities to obtain a sound source identification result of the detected sound field of the sound source to be identified.
Optionally, obtaining the source strengths of different sound source types corresponding to each axis in the different axes according to the transfer relationship between the equivalent sound source surface and the measurement surface, the block sparse bayesian algorithm and the orthogonal matching pursuit algorithm in S31 includes:
s311, setting initial iteration times l=0, and taking an initial residual signal as a measurement plane P m The support set index is empty set
S312, solving a transfer relation between an equivalent sound source surface and a measurement surface based on a block sparse Bayesian algorithm to obtain source strength, preliminarily estimating the number and position information of sound sources, and endowing the number of the sound sources within a dynamic range of 6dB to S.
S313, solving to obtain source intensity q by utilizing block sparse Bayesian algorithm l Find q l Adding the atom corresponding to the position into the support set index at the position of the largest element in the support set index, and updating the residual signal.
S314, judging whether the iteration number l is smaller than S, if l is smaller than S, making l=l+1, and turning to execute S313; and if l is more than or equal to s, stopping iteration to obtain the source strengths of different sound source types corresponding to each axis in different axes.
Optionally, solving the transfer relationship between the equivalent sound source surface and the measurement surface based on the block sparse bayesian algorithm in S312 to obtain a source strength includes:
s3121, establishing a compressed sensing model according to the transfer relation between the equivalent sound source surface and the measurement surface.
S3122, establishing a layered Bayesian model according to the compressed sensing model and the sparse Bayesian learning algorithm.
And S3123, optimizing the posterior probability density and likelihood function in the hierarchical Bayesian model to obtain the source intensity.
Optionally, the step S32 of superposing the source intensities to obtain a sound source identification result of the measured sound field of the sound source to be identified includes:
s321, obtaining sound source types according to the source strengths of different sound source types corresponding to each axis, and positioning the sound source distribution position of each sound source type.
S322, superposing the source intensities corresponding to each sound source type to obtain a sound source identification result of the detected sound field of the sound source to be identified.
On the other hand, the invention provides an orthogonal matching pursuit sound source recognition device based on block sparse Bayes, which is applied to realizing an orthogonal matching pursuit sound source recognition method based on block sparse Bayes, and comprises the following steps:
and the measuring surface setting module is used for setting a measuring surface parallel to the sound source surface in the measured sound field of the sound source to be identified.
And the transfer relation establishing module is used for arranging a plurality of equivalent sound sources at the position of the sound source surface and establishing a transfer relation between the equivalent sound source surface and the measuring surface.
And the output module is used for obtaining a sound source identification result of the detected sound field of the sound source to be identified according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching pursuit algorithm.
Optionally, the measurement face setting module is further configured to:
in a radiation sound field of a multipole sound source of a sound source to be identified, a measurement surface parallel to the sound source surface is provided, the measurement surface being an archimedes spiral array comprising a plurality of sound pressure sensors.
Optionally, the transfer relation establishment module is further configured to:
S21, collecting sound pressure signals on the measuring surface.
S22, arranging a plurality of equivalent sound sources at the positions of the sound source surface according to the sound pressure signals.
S23, predefining multipole sub-sound source radiation models corresponding to different axial directions for each equivalent sound source in the plurality of equivalent sound sources, and establishing a transfer relation between an equivalent sound source surface and a measurement surface.
Optionally, the transfer relation establishment module is further configured to:
predefining a sound source type and a target domain, calculating a transmission matrix, and establishing a transfer relation between an equivalent sound source surface and a measurement surface, wherein the transfer relation is shown in the following formula (1):
wherein G is mon ，G dip ，
Representing the free field green's function, q, between the equivalent sound source and the measurement surface of monopole, dipole, longitudinal quadrupole, transverse quadrupole, respectively mon ，q dip ，/>
Respectively representing the source intensity corresponding to the monopole sound source, the dipole sound source, the longitudinal quadrupole sound source and the transverse quadrupole sound source, wherein n is notNumber n of equivalent sound sources of the same sound source type respectively corresponding to mon ，n dip ，/>
Is the total number of equivalent sound sources corresponding to monopole sound sources, dipole sound sources, longitudinal quadrupole sound sources and transverse quadrupole sound sources, P m To measure the surface acoustic pressure.
Wherein G is mon ，G dip ，
The expression of (2) to (5) below:
Wherein r is the spatial distance between the position of the measuring point and the position of the equivalent sound source, k is the wave number of the analysis frequency, i is the imaginary unit, r is the spatial distance between the equivalent sound source and the position of the measuring point, θ d Is the included angle between the normal vector of the dipole axis and the vector of the sound source pointing to the measuring point position, theta qlo Is the included angle between the transverse axis of the longitudinal quadrupoles and the vector of the equivalent sound source pointing to the measuring point position,
is the included angle between the longitudinal axis of the longitudinal quadrupole and the measuring point, theta qla Between the axis of the transverse quadrupole and the measuring pointIs included in the bearing.
Optionally, the output module is further configured to:
s31, obtaining the source strengths of different sound source types corresponding to each axis in different axes according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching tracking algorithm.
S32, superposing the source intensities to obtain a sound source identification result of the detected sound field of the sound source to be identified.
Optionally, the output module is further configured to:
s311, setting initial iteration times l=0, and taking an initial residual signal as a measurement plane P m The support set index is empty set
S312, solving a transfer relation between an equivalent sound source surface and a measurement surface based on a block sparse Bayesian algorithm to obtain source strength, preliminarily estimating the number and position information of sound sources, and endowing the number of the sound sources within a dynamic range of 6dB to S.
S313, solving to obtain source intensity q by utilizing block sparse Bayesian algorithm l Find q l Adding the atom corresponding to the position into the support set index at the position of the largest element in the support set index, and updating the residual signal.
S314, judging whether the iteration number l is smaller than S, if l is smaller than S, making l=l+1, and turning to execute S313; and if l is more than or equal to s, stopping iteration to obtain the source strengths of different sound source types corresponding to each axis in different axes.
Optionally, the output module is further configured to:
s3121, establishing a compressed sensing model according to the transfer relation between the equivalent sound source surface and the measurement surface.
S3122, establishing a layered Bayesian model according to the compressed sensing model and the sparse Bayesian learning algorithm.
And S3123, optimizing the posterior probability density and likelihood function in the hierarchical Bayesian model to obtain the source intensity.
Optionally, the output module is further configured to:
s321, obtaining sound source types according to the source strengths of different sound source types corresponding to each axis, and positioning the sound source distribution position of each sound source type.
S322, superposing the source intensities corresponding to each sound source type to obtain a sound source identification result of the detected sound field of the sound source to be identified.
In one aspect, an electronic device is provided, which includes a processor and a memory, where at least one instruction is stored in the memory, where the at least one instruction is loaded and executed by the processor to implement the above-mentioned block sparse bayesian-based orthogonal matching pursuit sound source identification method.
In one aspect, a computer readable storage medium is provided, in which at least one instruction is stored, and the at least one instruction is loaded and executed by a processor to implement the above-mentioned block sparse bayesian-based quadrature matching pursuit sound source recognition method.
Compared with the prior art, the technical scheme has at least the following beneficial effects:
according to the scheme, the problem of complex sound source identification under the coexistence of multipoles of different types such as monopole, dipole and quadrupole can be simultaneously realized, and the limitation that the traditional method can only identify the dipole sound source or a single low-frequency quadrupole sound source with obvious far-field directivity is solved.
According to the method, a block sparse Bayesian method with higher resolution is introduced, and the conventional atomic screening criterion is replaced by the block sparse Bayesian method with stronger sparsity, so that the sparsity is ensured, and meanwhile, the calculation accuracy is greatly improved. The method solves the problems that in the traditional method, the block sparse Bayesian algorithm is not suitable for the environment with low signal to noise ratio, when the noise influence is obvious, serious ghost phenomenon easily occurs in the identification result, and under the condition of strong correlation of the sensing matrix atoms, the orthogonal matching pursuit method cannot accurately screen out atoms, so that algorithm failure is caused, and the like.
The invention solves the problem that the prior knowledge of the sound field needs to be acquired in advance in the traditional method. On the one hand, the multi-pole sound source identification is realized by constructing the inverse problem reflecting the sound field transfer relation, and the equivalent sources are predefined to be a plurality of sound source types in different axial directions, so that the sound field transfer function is not required to be adjusted again according to the directivity information of the multi-pole sound source. On the other hand, the position and the size of the non-zero element in the signal to be reconstructed can be automatically determined through iteration by the block sparse Bayesian algorithm, so that the source strong sparsity value does not need to be specified in advance.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of an orthogonal matching pursuit sound source identification method based on block sparse Bayes, which is provided by the embodiment of the invention;
FIG. 2 is a flow chart of an orthogonal matching pursuit multipole sound source identification method based on a block sparse Bayesian prior provided by an embodiment of the invention;
Fig. 3 is a schematic diagram of a block sparse bayesian prior-based orthogonal matching pursuit multipole sound source identification method according to an embodiment of the present invention;
FIG. 4 is a graph of recognition results for monopole and dipole sound sources provided by an embodiment of the present invention;
FIG. 5 is a graph of the recognition results for transverse quadrupole and longitudinal quadrupole sound sources provided by an embodiment of the present invention;
FIG. 6 is a block diagram of an orthogonal matching pursuit sound source recognition device based on block sparse Bayes provided by an embodiment of the invention;
fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. It will be apparent that the described embodiments are some, but not all, embodiments of the invention. All other embodiments, which can be made by a person skilled in the art without creative efforts, based on the described embodiments of the present invention fall within the protection scope of the present invention.
As shown in fig. 1, the embodiment of the invention provides a block sparse bayesian-based orthogonal matching pursuit sound source identification method, which can be implemented by electronic equipment. The flow chart of the block sparse Bayes-based orthogonal matching pursuit sound source identification method shown in fig. 1 can comprise the following steps:
S1, setting a measuring surface parallel to a sound source surface in a measured sound field of a sound source to be identified.
Optionally, in the measured sound field of the sound source to be identified in S1, a measurement plane parallel to the sound source plane is set, including:
in a radiation sound field of a multipole sound source of a sound source to be identified, a measurement surface parallel to the sound source surface is provided, the measurement surface being an archimedes spiral array comprising a plurality of sound pressure sensors.
In a possible embodiment, as shown in fig. 2 and 3, in the radiated sound field of the multipole sound source, a measuring plane P parallel to the sound emission plane is arranged m The measuring surface can be an Archimedes spiral array formed by 63 sound pressure sensors; the aperture of the measuring surface can be 1.5, the unit m, and the distance between adjacent measuring points is smaller than half wavelength of the analysis frequency; the data acquisition system based on the NI-PXIe bus is used for completing multichannel synchronous acquisition of the microphone array, and the sampling frequency is 44100Hz; the sound field under test may be a sound field of a steady state.
S2, arranging a plurality of equivalent sound sources at the position of the sound source surface, and establishing a transfer relation between the equivalent sound source surface and the measuring surface.
Optionally, the step S2 may include the following steps S21 to S23:
s21, collecting sound pressure signals on the measuring surface.
S22, arranging a plurality of equivalent sound sources at the positions of the sound source surface according to the sound pressure signals.
S23, predefining multipole sub-sound source radiation models corresponding to different axial directions for each equivalent sound source in the plurality of equivalent sound sources, and establishing a transfer relation between an equivalent sound source surface and a measurement surface.
Optionally, establishing a transfer relationship between the equivalent sound source surface and the measurement surface includes:
predefining a sound source type and a target domain, calculating a transmission matrix, and establishing a transfer relation between an equivalent sound source surface and a measurement surface, wherein the transfer relation is shown in the following formula (1):
wherein G is mon ，G dip ，
Representing the free field green's function, q, between the equivalent sound source and the measurement surface of monopole, dipole, longitudinal quadrupole, transverse quadrupole, respectively mon ，q dip ，/>
Respectively representing the source strengths corresponding to the monopole sound source, the dipole sound source, the longitudinal quadrupole sound source and the transverse quadrupole sound source, wherein n is the number corresponding to the equivalent sound sources with different sound source types, and n is mon ，n dip ，/>
Is the total number of equivalent sound sources corresponding to monopole sound sources, dipole sound sources, longitudinal quadrupole sound sources and transverse quadrupole sound sources, P m To measure the surface acoustic pressure. />
Wherein G is mon ，G dip ，
The expression of (2) to (5) below:
wherein r is the spatial distance between the position of the measuring point and the position of the equivalent sound source, k is the wave number of the analysis frequency, i is the imaginary unit, r is the spatial distance between the equivalent sound source and the position of the measuring point, θ d Is the included angle between the normal vector of the dipole axis and the vector of the sound source pointing to the measuring point position, theta qlo Is the included angle between the transverse axis of the longitudinal quadrupoles and the vector of the equivalent sound source pointing to the measuring point position,
is the included angle between the longitudinal axis of the longitudinal quadrupole and the measuring point, theta qla Is the included angle between the axis of the transverse quadrupoles and the measuring point.
In a possible embodiment, if the sound source of a known partial type does not participate in the construction of the radiated sound field during the actual measurement, the equivalent source type corresponding thereto can be removed from the above equation. In the method, for the dipole, the transverse quadrupole sound source and the longitudinal quadrupole sound source, 4 equivalent sources in the axial direction are selected, namely, the axial direction comprises 0 degree, 45 degrees, 90 degrees and 135 degrees, and through linear combination of the two equivalent sources, any source on a plane can be identified, so that redundancy is avoided, and meanwhile, the calculation accuracy is high.
And S3, obtaining a sound source identification result of the measured sound field of the sound source to be identified according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching pursuit algorithm.
Optionally, the step S3 may include the following steps S31 to S32:
s31, obtaining the source strengths of different sound source types corresponding to each axis in different axes according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching tracking algorithm.
Optionally, the step S31 may include the following steps S311 to S314:
s311, setting initial iteration times l=0, and taking an initial residual signal as a measurement plane P m The support set index is empty set
S312, solving a transfer relation between an equivalent sound source surface and a measurement surface based on a block sparse Bayesian algorithm to obtain source strength, preliminarily estimating the number and position information of sound sources, and endowing the number of the sound sources within a dynamic range of 6dB to S.
S313, solving to obtain source intensity q by utilizing block sparse Bayesian algorithm l Find q l Adding the atom corresponding to the position into the support set index at the position of the largest element in the support set index, and updating the residual signal.
S314, judging whether the iteration number l is smaller than S, if l is smaller than S, making l=l+1, and turning to execute S313; and if l is more than or equal to s, stopping iteration to obtain the source strengths of different sound source types corresponding to each axis in different axes.
In a feasible implementation manner, a sound source identification result obtained by a block sparse Bayesian algorithm is used as an atomic screening principle of an orthogonal matching pursuit algorithm, the number of sound sources preliminarily determined by the block sparse Bayesian algorithm is used as iteration times, and the source intensity corresponding to each predefined axis direction is finally obtained through repeated iteration.
Further, in the process of obtaining source intensity through iteration, setting initial iteration frequency l=0, performing sound source localization through a block sparse Bayesian method, endowing s with the number of sound sources within a dynamic range of 6dB, and increasing the iteration frequencyl=l+1, and the initial residual signal is set as the measurement sound pressure P m The support set index is empty set
Then, comparing the iteration times l with the iteration times s, and if l is smaller than s, solving by using a block sparse Bayesian method to obtain a source strength q l Find q l Adding the atom corresponding to the position of the largest element in the position into the support set index, updating the residual error, and recalculating the source intensity. Stopping iteration until the iteration stopping condition is met, namely, when l is more than or equal to s, outputting an accurate solution of the source intensity of the sound source, and finally obtaining the source intensity corresponding to each predefined axis direction.
Optionally, solving the transfer relationship between the equivalent sound source surface and the measurement surface based on a block sparse bayesian algorithm to obtain a source strength, including:
s3121, establishing a compressed sensing model according to the transfer relation between the equivalent sound source surface and the measurement surface.
In a possible implementation manner, a compressed sensing model is built towards a source intensity solving process, and the compressed sensing model is shown in the following formula (6):
P m ＝Gq+v (6)
wherein q.epsilon.R N For the signal to be reconstructed, i.e. the source strength to be solved, i.e.
G∈R M×N For the sense matrix, i.e. green's function, i.e +.>
P m ∈R M For observing vectors, i.e. measuring sound pressure, v e R M To observe noise. Where φ is typically an underdetermined matrix, i.e., M < N.
S3122, establishing a layered Bayesian model according to the compressed sensing model and the sparse Bayesian learning algorithm.
In a possible implementation, the sparse Bayesian learning algorithm is implemented by observing P m Establishing a layered Bayesian model with the signal q to be solved, wherein the layered Bayesian model is represented by the following formula (7)The illustration is:
wherein the signal q is divided into several blocks, q i Represents the ith signal block, N is the number of signal blocks, beta -1 Is the noise variance, I is the identity matrix, gamma i Representing the i-th block signal q i And observation matrix P m Is a correlation of (3).
And S3123, optimizing the posterior probability density and likelihood function in the hierarchical Bayesian model to obtain the source intensity.
In one possible embodiment, the probability density is determined by applying a probability density P (q|P m ；{|γ i -beta) and likelihood function P (P m ；{|γ i And } beta) optimizing to realize sparse reconstruction of signals.
S32, superposing the source intensities to obtain a sound source identification result of the detected sound field of the sound source to be identified.
Optionally, the step S32 may include the following steps S321 to S322:
S321, obtaining sound source types according to the source strengths of different sound source types corresponding to each axis, and positioning the sound source distribution position of each sound source type.
S322, superposing the source intensities corresponding to each sound source type to obtain a sound source identification result of the detected sound field of the sound source to be identified.
In a feasible implementation manner, after the source strengths of different types of sound sources corresponding to each axis direction are obtained respectively, the type of sound source can be judged respectively, the distribution position of each type of sound source is positioned, and after the source strengths corresponding to each predefined sound source type are further overlapped, the accurate positioning of any type of multipole sound source can be realized according to the source strength distribution.
For example, as shown in fig. 3, the measuring surface is an archimedes spiral array with an aperture of 1.5m, which is composed of 63 array elements, a geometric center O point of the measuring surface is taken as a spatial coordinate origin, a spatial coordinate system is established, a distance L between the sound source surface and the measuring surface is 0.75m, a sound source frequency is set to 3000Hz, a signal to noise ratio is set to 15dB, and simulation cases of four sound source sound fields are given in the present invention. The invention gives out the identification results of monopole and dipole sound source, and transverse quadrupole sound source and longitudinal quadrupole sound source, and marks the real position of sound source with "+" sign in the figure.
Further, referring to fig. 4, it can be seen that when the radiation sound field includes two monopole and two dipole sound sources, the invention can accurately identify the sound source position, and accurately determine the directivity information of the sound source. Similarly, referring to fig. 5, it can be seen that when the radiation sound field includes two propagation signals of the transverse quadrupoles and two longitudinal quadrupoles, the identification of the positions of the four quadrupoles sound sources can be accurately realized based on the invention, and meanwhile, the directivity information of the sound source can be accurately judged, which proves that whether the sound source is a monopole or a multipole sound source with obvious directivity information, the method provided by the invention can effectively realize that a plurality of target multipole sound sources are accurately identified in the free sound field without prior knowledge of the sound field, and the identification result is stable and accurate.
In the embodiment of the invention, the complex sound source identification problem under the coexistence condition of multipoles of different types such as monopole, dipole, quadrupole and the like can be simultaneously realized, and the limitation that the traditional method can only identify the dipole sound source or a single low-frequency quadrupole sound source with obvious far-field directivity is solved.
According to the method, a block sparse Bayesian method with higher resolution is introduced, and the conventional atomic screening criterion is replaced by the block sparse Bayesian method with stronger sparsity, so that the sparsity is ensured, and meanwhile, the calculation accuracy is greatly improved. The method solves the problems that in the traditional method, the block sparse Bayesian algorithm is not suitable for the environment with low signal to noise ratio, when the noise influence is obvious, serious ghost phenomenon easily occurs in the identification result, and under the condition of strong correlation of the sensing matrix atoms, the orthogonal matching pursuit method cannot accurately screen out atoms, so that algorithm failure is caused, and the like.
The invention solves the problem that the prior knowledge of the sound field needs to be acquired in advance in the traditional method. On the one hand, the multi-pole sound source identification is realized by constructing the inverse problem reflecting the sound field transfer relation, and the equivalent sources are predefined to be a plurality of sound source types in different axial directions, so that the sound field transfer function is not required to be adjusted again according to the directivity information of the multi-pole sound source. On the other hand, the position and the size of the non-zero element in the signal to be reconstructed can be automatically determined through iteration by the block sparse Bayesian algorithm, so that the source strong sparsity value does not need to be specified in advance.
As shown in fig. 6, an embodiment of the present invention provides an apparatus 600 for identifying an orthogonal matching pursuit sound source based on block sparse bayesian, where the apparatus 600 is applied to implement an orthogonal matching pursuit sound source identification method based on block sparse bayesian, and the apparatus 600 includes:
The measurement surface setting module 610 is configured to set a measurement surface parallel to a sound source surface in a measured sound field of a sound source to be identified.
The transfer relation establishment module 620 is configured to arrange a plurality of equivalent sound sources at positions of the sound source surface, and establish a transfer relation between the equivalent sound source surface and the measurement surface.
And the output module 630 is configured to obtain a sound source identification result of the measured sound field of the sound source to be identified according to the transfer relationship between the equivalent sound source surface and the measurement surface, the block sparse bayesian algorithm, and the orthogonal matching pursuit algorithm.
Optionally, the measurement surface setting module 610 is further configured to:
in a radiation sound field of a multipole sound source of a sound source to be identified, a measurement surface parallel to the sound source surface is provided, the measurement surface being an archimedes spiral array comprising a plurality of sound pressure sensors.
Optionally, the transfer relationship establishment module 620 is further configured to:
s21, collecting sound pressure signals on the measuring surface.
S22, arranging a plurality of equivalent sound sources at the positions of the sound source surface according to the sound pressure signals.
S23, predefining multipole sub-sound source radiation models corresponding to different axial directions for each equivalent sound source in the plurality of equivalent sound sources, and establishing a transfer relation between an equivalent sound source surface and a measurement surface.
Optionally, the transfer relationship establishment module 620 is further configured to:
predefining a sound source type and a target domain, calculating a transmission matrix, and establishing a transfer relation between an equivalent sound source surface and a measurement surface, wherein the transfer relation is shown in the following formula (1):
wherein G is mon ，G dip ，
Representing the free field green's function, q, between the equivalent sound source and the measurement surface of monopole, dipole, longitudinal quadrupole, transverse quadrupole, respectively mon ，q dip ，/>
Respectively representing the source strengths corresponding to the monopole sound source, the dipole sound source, the longitudinal quadrupole sound source and the transverse quadrupole sound source, wherein n is the number corresponding to the equivalent sound sources with different sound source types, and n is mon ，n dip ，/>
Is the total number of equivalent sound sources corresponding to monopole sound sources, dipole sound sources, longitudinal quadrupole sound sources and transverse quadrupole sound sources, P m To measure the surface acoustic pressure.
Wherein G is mon ，G dip ，
The expression of (2) to (5) below:
wherein r is the spatial distance between the position of the measuring point and the position of the equivalent sound source, k is the wave number of the analysis frequency, i is the imaginary unit, r is the spatial distance between the equivalent sound source and the position of the measuring point, θ d Is the included angle between the normal vector of the dipole axis and the vector of the sound source pointing to the measuring point position, theta qlo Is the included angle between the transverse axis of the longitudinal quadrupoles and the vector of the equivalent sound source pointing to the measuring point position, 
Is the included angle between the longitudinal axis of the longitudinal quadrupole and the measuring point, theta qla Is the included angle between the axis of the transverse quadrupoles and the measuring point.
Optionally, the output module 630 is further configured to:
s31, obtaining the source strengths of different sound source types corresponding to each axis in different axes according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching tracking algorithm.
S32, superposing the source intensities to obtain a sound source identification result of the detected sound field of the sound source to be identified.
Optionally, the output module 630 is further configured to:
s311, setting initial iteration times l=0, and taking an initial residual signal as a measurement plane P m The support set index is empty set
S312, solving a transfer relation between an equivalent sound source surface and a measurement surface based on a block sparse Bayesian algorithm to obtain source strength, preliminarily estimating the number and position information of sound sources, and endowing the number of the sound sources within a dynamic range of 6dB to S.
S313, solving to obtain source intensity q by utilizing block sparse Bayesian algorithm l Find q l Adding the atom corresponding to the position into the support set index at the position of the largest element in the support set index, and updating the residual signal.
S314, judging whether the iteration number l is smaller than S, if l is smaller than S, making l=l+1, and turning to execute S313; and if l is more than or equal to s, stopping iteration to obtain the source strengths of different sound source types corresponding to each axis in different axes.
Optionally, the output module 630 is further configured to:
s3121, establishing a compressed sensing model according to the transfer relation between the equivalent sound source surface and the measurement surface.
S3122, establishing a layered Bayesian model according to the compressed sensing model and the sparse Bayesian learning algorithm.
And S3123, optimizing the posterior probability density and likelihood function in the hierarchical Bayesian model to obtain the source intensity.
Optionally, the output module 630 is further configured to:
s321, obtaining sound source types according to the source strengths of different sound source types corresponding to each axis, and positioning the sound source distribution position of each sound source type.
S322, superposing the source intensities corresponding to each sound source type to obtain a sound source identification result of the detected sound field of the sound source to be identified.
In the embodiment of the invention, the complex sound source identification problem under the coexistence condition of multipoles of different types such as monopole, dipole, quadrupole and the like can be simultaneously realized, and the limitation that the traditional method can only identify the dipole sound source or a single low-frequency quadrupole sound source with obvious far-field directivity is solved.
According to the method, a block sparse Bayesian method with higher resolution is introduced, and the conventional atomic screening criterion is replaced by the block sparse Bayesian method with stronger sparsity, so that the sparsity is ensured, and meanwhile, the calculation accuracy is greatly improved. The method solves the problems that in the traditional method, the block sparse Bayesian algorithm is not suitable for the environment with low signal to noise ratio, when the noise influence is obvious, serious ghost phenomenon easily occurs in the identification result, and under the condition of strong correlation of the sensing matrix atoms, the orthogonal matching pursuit method cannot accurately screen out atoms, so that algorithm failure is caused, and the like.
The invention solves the problem that the prior knowledge of the sound field needs to be acquired in advance in the traditional method. On the one hand, the multi-pole sound source identification is realized by constructing the inverse problem reflecting the sound field transfer relation, and the equivalent sources are predefined to be a plurality of sound source types in different axial directions, so that the sound field transfer function is not required to be adjusted again according to the directivity information of the multi-pole sound source. On the other hand, the position and the size of the non-zero element in the signal to be reconstructed can be automatically determined through iteration by the block sparse Bayesian algorithm, so that the source strong sparsity value does not need to be specified in advance.
Fig. 7 is a schematic structural diagram of an electronic device 700 according to an embodiment of the present invention, where the electronic device 700 may have a relatively large difference due to different configurations or performances, and may include one or more processors (central processing units, CPU) 701 and one or more memories 702, where at least one instruction is stored in the memories 702, and the at least one instruction is loaded and executed by the processors 701 to implement the following block sparse bayesian-based quadrature matching pursuit sound source recognition method:
s1, setting a measuring surface parallel to a sound source surface in a measured sound field of a sound source to be identified.
S2, arranging a plurality of equivalent sound sources at the position of the sound source surface, and establishing a transfer relation between the equivalent sound source surface and the measuring surface.
And S3, obtaining a sound source identification result of the measured sound field of the sound source to be identified according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching pursuit algorithm.
In an exemplary embodiment, a computer readable storage medium, such as a memory, comprising instructions executable by a processor in a terminal to perform the above-described block sparse bayesian-based quadrature matching pursuit sound source identification method is also provided. For example, the computer readable storage medium may be ROM, random Access Memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program for instructing relevant hardware, where the program may be stored in a computer readable storage medium, and the storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.
Claims (10)
1. An orthogonal matching pursuit sound source identification method based on block sparse Bayes, which is characterized by comprising the following steps:
s1, setting a measuring surface parallel to a sound source surface in a measured sound field of a sound source to be identified;
s2, arranging a plurality of equivalent sound sources at the position of the sound source surface, and establishing a transfer relationship between the equivalent sound source surface and the measurement surface;
and S3, obtaining a sound source identification result of the detected sound field of the sound source to be identified according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching pursuit algorithm.
2. The method according to claim 1, wherein the step of setting a measurement plane parallel to the sound source plane in the measured sound field of the sound source to be identified in S1 includes:
in a radiated sound field of a multipole sound source of a sound source to be identified, a measurement surface parallel to the sound source surface is provided, the measurement surface being an archimedes spiral array including a plurality of sound pressure sensors.
3. The method according to claim 1, wherein arranging a plurality of equivalent sound sources at the position of the sound source face in the S2, and establishing a transfer relationship between the equivalent sound source face and the measurement face, comprises:
S21, collecting sound pressure signals on the measuring surface;
s22, arranging a plurality of equivalent sound sources at the position of the sound source surface according to the sound pressure signal;
s23, predefining multipole sound source radiation models corresponding to different axial directions for each equivalent sound source in the plurality of equivalent sound sources, and establishing a transfer relation between an equivalent sound source surface and the measuring surface.
4. The method according to claim 1, wherein establishing a transfer relationship between the equivalent sound source surface and the measurement surface in S2 comprises:
predefining a sound source type and a target domain, calculating a transmission matrix, and establishing a transfer relation between an equivalent sound source surface and the measurement surface, wherein the transfer relation is shown in the following formula (1):
wherein G is mon ，G dip ，
Representing the free field green's function, q, between the equivalent sound source and the measurement surface of monopole, dipole, longitudinal quadrupole, transverse quadrupole, respectively mon ，q dip ，/>
Respectively representing the source strengths corresponding to the monopole sound source, the dipole sound source, the longitudinal quadrupole sound source and the transverse quadrupole sound source, wherein n is the number corresponding to the equivalent sound sources with different sound source types, and n is mon ，n dip ，/>
Is the total number of equivalent sound sources corresponding to monopole sound sources, dipole sound sources, longitudinal quadrupole sound sources and transverse quadrupole sound sources, P m To measure the surface sound pressure;
wherein G is mon ，G dip ，
The expression of (2) to (5) below:
wherein r is the spatial distance between the position of the measuring point and the position of the equivalent sound source, k is the wave number of the analysis frequency, i is the imaginary unit, r is the spatial distance between the equivalent sound source and the position of the measuring point, θ d Is the included angle between the normal vector of the dipole axis and the vector of the sound source pointing to the measuring point position, theta qlo Is the included angle between the transverse axis of the longitudinal quadrupoles and the vector of the equivalent sound source pointing to the measuring point position,
is the included angle between the longitudinal axis of the longitudinal quadrupole and the measuring point, theta qla Is the included angle between the axis of the transverse quadrupoles and the measuring point.
5. The method according to claim 1, wherein the obtaining the sound source recognition result of the measured sound field of the sound source to be recognized in S3 according to the transfer relationship between the equivalent sound source surface and the measurement surface, the block sparse bayesian algorithm, and the orthogonal matching pursuit algorithm includes:
s31, obtaining source strengths of different sound source types corresponding to each axis in different axes according to the transfer relation between the equivalent sound source surface and the measurement surface, a block sparse Bayesian algorithm and an orthogonal matching tracking algorithm;
s32, superposing the source intensities to obtain a sound source identification result of the detected sound field of the sound source to be identified.
6. The method according to claim 5, wherein the obtaining the source strengths of different source types corresponding to each axis of the different axes in S31 according to the transfer relationship between the equivalent sound source surface and the measurement surface, the block sparse bayesian algorithm, and the orthogonal matching pursuit algorithm includes:
s311, setting initial iteration times l=0, and taking an initial residual signal as a measurement plane P m The support set index is empty set
S312, solving a transfer relation between the equivalent sound source surface and the measurement surface based on a block sparse Bayesian algorithm to obtain source strength, preliminarily estimating the number and position information of sound sources, and endowing the number of the sound sources within a dynamic range of 6dB to S;
s313, solving to obtain source intensity q by utilizing block sparse Bayesian algorithm l Find the q l Adding atoms corresponding to the positions to the support set index at the positions of the largest elements in the support set index, and updating residual signals;
s314, judging whether the iteration number l is smaller than S, if l is smaller than S, making l=l+1, and turning to execute S313; and if l is more than or equal to s, stopping iteration to obtain the source strengths of different sound source types corresponding to each axis in different axes.
7. The method of claim 6, wherein the solving the transfer relationship between the equivalent sound source surface and the measurement surface based on the block sparse bayesian algorithm in S312 to obtain a source strength includes:
S3121, establishing a compressed sensing model according to the transfer relation between the equivalent sound source surface and the measurement surface;
s3122, establishing a layered Bayesian model according to the compressed sensing model and a sparse Bayesian learning algorithm;
and S3123, optimizing the posterior probability density and likelihood function in the hierarchical Bayesian model to obtain a strong source.
8. The method according to claim 5, wherein the step of superimposing the source intensities in S32 to obtain a sound source identification result of the detected sound field of the sound source to be identified includes:
s321, obtaining sound source types according to the source strengths of different sound source types corresponding to each axis, and positioning the sound source distribution position of each sound source type;
s322, superposing the source intensities corresponding to each sound source type to obtain a sound source identification result of the detected sound field of the sound source to be identified.
9. An orthogonal matching pursuit sound source recognition device based on block sparse bayesian, the device comprising:
the measuring surface setting module is used for setting a measuring surface parallel to the sound source surface in the measured sound field of the sound source to be identified;
the transmission relation establishing module is used for arranging a plurality of equivalent sound sources at the position of the sound source surface and establishing a transmission relation between the equivalent sound source surface and the measuring surface;
And the output module is used for obtaining a sound source identification result of the detected sound field of the sound source to be identified according to the transfer relation between the equivalent sound source surface and the measurement surface, the block sparse Bayesian algorithm and the orthogonal matching pursuit algorithm.
10. The apparatus of claim 9, wherein the output module is further configured to:
s31, obtaining source strengths of different sound source types corresponding to each axis in different axes according to the transfer relation between the equivalent sound source surface and the measurement surface, a block sparse Bayesian algorithm and an orthogonal matching tracking algorithm;
s32, superposing the source intensities to obtain a sound source identification result of the detected sound field of the sound source to be identified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310187866.7A CN116203505B (en) | 2023-02-22 | 2023-02-22 | Orthogonal matching pursuit sound source identification method and device based on block sparse Bayes |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310187866.7A CN116203505B (en) | 2023-02-22 | 2023-02-22 | Orthogonal matching pursuit sound source identification method and device based on block sparse Bayes |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116203505A true CN116203505A (en) | 2023-06-02 |
| CN116203505B CN116203505B (en) | 2024-02-13 |
Family
ID=86514415
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310187866.7A Active CN116203505B (en) | 2023-02-22 | 2023-02-22 | Orthogonal matching pursuit sound source identification method and device based on block sparse Bayes |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116203505B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20100092593A (en) * | 2009-02-13 | 2010-08-23 | 한국과학기술원 | Active noise control method by using acoustical holography and apparatus using it |
| CN102121847A (en) * | 2010-12-16 | 2011-07-13 | 合肥工业大学 | Method for reestablishing transient sound field |
| CN103728013A (en) * | 2013-12-25 | 2014-04-16 | 广西科技大学 | Noise source recognizing method |
| CN106289505A (en) * | 2016-07-21 | 2017-01-04 | 合肥工业大学 | A kind of method separating static sound source radiation sound field and rotation sound source radiated sound field |
| US20170094421A1 (en) * | 2015-09-25 | 2017-03-30 | Ritwik Giri | Dynamic relative transfer function estimation using structured sparse bayesian learning |
| CN111664932A (en) * | 2020-05-22 | 2020-09-15 | 重庆大学 | Sound source identification method based on Bayesian compressed sensing |
-
2023
- 2023-02-22 CN CN202310187866.7A patent/CN116203505B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20100092593A (en) * | 2009-02-13 | 2010-08-23 | 한국과학기술원 | Active noise control method by using acoustical holography and apparatus using it |
| CN102121847A (en) * | 2010-12-16 | 2011-07-13 | 合肥工业大学 | Method for reestablishing transient sound field |
| CN103728013A (en) * | 2013-12-25 | 2014-04-16 | 广西科技大学 | Noise source recognizing method |
| US20170094421A1 (en) * | 2015-09-25 | 2017-03-30 | Ritwik Giri | Dynamic relative transfer function estimation using structured sparse bayesian learning |
| CN106289505A (en) * | 2016-07-21 | 2017-01-04 | 合肥工业大学 | A kind of method separating static sound source radiation sound field and rotation sound source radiated sound field |
| CN111664932A (en) * | 2020-05-22 | 2020-09-15 | 重庆大学 | Sound source identification method based on Bayesian compressed sensing |
Non-Patent Citations (3)
| Title |
|---|
| 张亚虎 等: "基于等效源法的近场声全息技术研究与应用", 赤峰学院学报（自然科学版）, vol. 37, no. 7, pages 27 - 29 * |
| 张宏宇 等: "基于镜像源方法抑制混响干扰的声全息算法改进", 科学技术与工程, vol. 19, no. 33, pages 113 - 118 * |
| 陈汉涛 等: "低信噪比条件下的平面等效源近场声全息方法", 科学技术与工程, vol. 18, no. 27, pages 116 - 124 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116203505B (en) | 2024-02-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108845325B (en) | Towed line array sonar subarray error mismatch estimation method | |
| CN108181557B (en) | Method for determining ultrahigh frequency partial discharge signal direction | |
| CN109375198B (en) | Low-frequency sonar array impedance characteristic calibration method | |
| CN111812581B (en) | Spherical array sound source direction-of-arrival estimation method based on atomic norms | |
| CN106970348A (en) | Electromagnetic Vector Sensor Array decorrelation LMS two dimension MUSIC method for parameter estimation | |
| CN110095755B (en) | Sound source positioning method | |
| CN110736976A (en) | sonar beam former performance estimation method of arbitrary array | |
| CN103048641B (en) | Vector array steady focusing processing method with combined constraint optimization form | |
| CN116203505B (en) | Orthogonal matching pursuit sound source identification method and device based on block sparse Bayes | |
| CN115201821B (en) | Small target detection method based on strong target imaging cancellation | |
| CN114252148B (en) | Sound field reconstruction method based on prolate ellipsoid wave superposition | |
| Minvielle et al. | Indoor 3-D radar imaging for low-RCS analysis | |
| Zhang et al. | DOA estimation method for underwater acoustic signals based on two-dimensional power distribution (TPD) for few element array | |
| CN110824484B (en) | Array element position estimation method based on constant modulus algorithm | |
| RU2385467C1 (en) | Method for spatial polarisation-sensitive localisation of multibeam radio signals | |
| Jia et al. | Bearing-only multiple sources localization and the spatial spectrum | |
| Zhang et al. | Research on continuous leakage location of stiffened structure based on frequency energy ratio mapping method | |
| Zhang et al. | A vortex electromagnetic imaging method based on two-dimensional matrix compensation for off-grid targets | |
| Wang et al. | A fast irregular microphone array design method based on acoustic beamforming | |
| CN111965602A (en) | Method and system for detecting amplitude-phase consistency of phased array radar | |
| Shi et al. | An improved generalized inverse beamforming-noise source localization method using acoustic vector sensor arrays | |
| CN111735996A (en) | Multipath interference suppression method and device for mathematical wave-absorbing darkroom construction | |
| Nagaraju et al. | Performance Analysis of Sparse Array Using Compressive Sensing in A Closed Multi-Path Environment | |
| CN113702939B (en) | Near-field local irradiation target scattering near-far field conversion method | |
| Minvielle et al. | Fast 3D Synthetic Aperture Radar Imaging from Polarization-Diverse Measurements |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |