CN110221249A - Compressed sensing based broadband sound source localization method - Google Patents
Compressed sensing based broadband sound source localization method Download PDFInfo
- Publication number
- CN110221249A CN110221249A CN201910408263.9A CN201910408263A CN110221249A CN 110221249 A CN110221249 A CN 110221249A CN 201910408263 A CN201910408263 A CN 201910408263A CN 110221249 A CN110221249 A CN 110221249A
- Authority
- CN
- China
- Prior art keywords
- sound source
- sound
- expression model
- frequency point
- frequency
- 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.)
- Pending
Links
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The present invention proposes a kind of compressed sensing based broadband sound source localization method, and the time domain measurement value of sound-source signal is obtained by microphone array, measured value is transformed to frequency domain by Fast Fourier Transform (FFT), the measurement data under frequency band required for selecting.It determines sound source region, grid dividing is carried out to sound source plane according to sound source number.According to the calculation matrix of the Helmoltz establishing equation grid node of free field Green's function to microphone array, compressed sensing based broadband sound source positioning joint sparse expression model is established.Singular value decomposition is carried out to the measured value of the frequency domain of acquisition, obtains deformed broadband sparse expression model.Expression model is solved using DCS-SOMP algorithm, obtains the strength of sound source under each narrowband, the overall sound pressure level of sound source is finally solved, the position of sound source is determined according to the strength of sound source at each grid node.This method increases SVD on the basis of DCS-SOMP algorithm, has not only inherited efficient, the easy convergent characteristic of DCS-SOMP algorithm, but also improves noise immunity.
Description
Technical field
The present invention relates to a kind of broadband sound source localization method, in particular to a kind of compressed sensing based broadband sound source positioning
Method.
Background technique
According to the classification to sound-source signal, sound localization method can be divided into narrowband sound localization method and broadband sound source positioning
Method.Narrowband sound localization method refers to that using some Frequency point, come localization of sound source, broadband sound source localization method refers to utilization
A certain band frequency carrys out localization of sound source.In practical projects, the frequency spectrum of signal is often more complicated, belongs to broadband signal, comprising not
An only frequency.In this case, auditory localization algorithm in narrowband can not determine specifically with which Frequency point, to can not position
Sound source.Broadband sound source localization method is then not limited to some frequency, can choose very wide frequency band, to solve narrowband sound source
The problem of location algorithm is in positioning.Studying broadband sound source location algorithm has higher engineering practical value.
It is directed to the auditory localization problem of narrowband, currently used auditory localization algorithm has beamforming algorithm, Power estimation
Algorithm, compressed sensing class algorithm.This team is in research before, in invention " compressed sensing based three-dimensional auditory localization side
A kind of OMP-SVD method is disclosed in method ", to solve the problems, such as the narrowband auditory localization under low signal-to-noise ratio.But we grind simultaneously
Study carefully discovery, OMP-SVD method can not solve broadband sound source orientation problem, so needing to study the auditory localization side for being directed to broadband
Method.
Existing common broadband sound source localization method mainly includes time domain TIDY algorithm, delay summation beamforming algorithm
And compressed sensing based broadband location algorithm.Wherein, time domain TIDY algorithm needs high sampling rate and processing mass data, because
This needs higher data space, this will cause the waste of resource in engineering.Delay summation beamforming algorithm is being asked
When solving broadband sound source orientation problem, can there are wider main lobe and more secondary lobes in auditory localization result figure, thus can not
Distinguish real sources and falsetto source.Compressed sensing based broadband location algorithm can reduce main lobe width, the shadow of suppressed sidelobes
It rings.But existing compressed sensing algorithm can be with localization of sound source, when signal-to-noise ratio reduces, it may appear that positioning knot when noise is relatively high
The situation of fruit inaccuracy.
Document " Distributed Compressed Sensing of Jointly Sparse Signals [C]
.asilomar conference on signals, systems and computers, 2005:1537-1541 " disclose one
Kind is based on distributed compression perception-synchronous orthogonal matching pursuit (Distributed Compressed Sensing-
Simultaneous Orthogonal Matching Pursuit, DCS-SOMP) broadband sound source localization method.This method is logical
Cross microphone array obtain sound-source signal time domain measurement value, establish microphone array to sound source plane grid node measurement square
Battle array, then establishes compressed sensing based broadband joint sparse expression model, solves finally by DCS-SOMP algorithm each narrow
Strength of sound source under frequency band, so that the sound source to broadband positions.Document the method is in signal-to-noise ratio (signal to
Noise ratio, SNR) it is higher when can obtain it is accurate as a result, but when signal-to-noise ratio is down to -20dB, this method can not be to sound
Source is positioned, i.e. the noise immunity of this method is poor.
In conclusion the voice signal in Practical Project is often broadband signal, therefore study broadband sound source location algorithm
It is worth with more Practical Project.Existing broadband sound source localization method have the defects that it is different, above-mentioned document the method it is anti-
Making an uproar property is poor, will receive larger limitation in engineering application.
Summary of the invention
In order to overcome the shortcomings of that noise immunity is poor in the prior art, the present invention provides one kind to be based on DCS-SOMP algorithm and surprise
Different value decomposes the broadband sound source localization method that (Singular Value Decomposition, SVD) is combined.Pass through microphone
Array obtain sound-source signal time domain measurement value, by Fast Fourier Transform (FFT) (Fast Fourier Transformation,
FFT measured value) is transformed into frequency domain, the measurement data under frequency band required for selecting.Sound source region is determined, according to sound source number
Mesh carries out grid dividing to sound source plane.According to the Helmoltz establishing equation grid node of free field Green's function to microphone
The calculation matrix of array establishes compressed sensing based broadband sound source positioning joint sparse expression model.To the frequency domain of acquisition
Measured value carries out singular value decomposition, obtains deformed broadband sparse expression model.Using DCS-SOMP algorithm to expression model
It is solved, obtains the strength of sound source under each narrowband, the overall sound pressure level of sound source is finally solved, at each grid node
Strength of sound source determines the position of sound source.This method increases SVD on the basis of DCS-SOMP algorithm, both inherits DCS-
Efficient, the easy convergent characteristic of SOMP algorithm, and improve noise immunity.
Based on the above principles, the technical solution of the present invention is as follows:
A kind of compressed sensing based broadband sound source localization method, it is characterised in that: the following steps are included:
Step 1: by the array acquisition sound-source signal containing M microphone, obtaining the measurement data of time domain；
Step 2: piecemeal, adding window and Fast Fourier Transform (FFT) being carried out to the time-domain measurement data that step 1 acquires and handled, is obtained
Obtain frequency-domain measurement data；
Step 3: choosing required frequency segment data, the corresponding data of each Frequency point in selected frequency band are y (fj)=
(y1j,...,yMj)T, wherein the dimension of y is M × B, and B is by the data block total number that divides in step 2, fjFor in selected frequency band
J-th of Frequency point；
Step 4: establishing compressed sensing based broadband sound source positioning joint sparse expression model, specifically include following step
It is rapid:
Step 4.1: determining sound source plane, and grid dividing is carried out to sound source plane, be possible with each grid node
Sound source constructs unknown sound-source signal x, and wherein sound-source signal x is made of the strength of sound source at grid node；
Step 4.2: according to the Helmoltz establishing equation Frequency point f of free field Green's functionjUnder sound source plane and wheat
Calculation matrix A (the f of gram wind arrayj):
Wherein i is imaginary unit, and c is the velocity of sound, dmnFor the distance between m-th of microphone and n-th of grid node；
Step 4.3: establish the auditory localization sparse expression model under each Frequency point:
y(fj)=A (fj)x(fj)+e(fj), j=1,2 ..., J
Wherein x (fj) it is Frequency point fjUnder sound-source signal, e (fj) it is Frequency point fjUnder measured value in include noise
, J is frequency points；
Step 4.4: sparse expression model under different frequency point being combined, compressed sensing based broadband sound source is established
Position joint sparse expression model:
Step 5: to the y (f in joint sparse expression modelj) SVD decomposition is carried out, obtain deformed sparse expression mould
Type:
To the microphone array frequency domain measurement value y (f in joint sparse expression modelj) singular value decomposition is carried out, it is surveyed
Magnitude y (fj), sound-source signal x (fj), noise item e (fj) signal subspace:
y(fj)SV=y (fj)VDK, x (fj)SV=x (fj)VDK, e (fj)SV=e (fj)VDK
Wherein DK=[IK0]T, V is to measured value y (fj) unitary matrice after singular value decomposition, IKFor K × K rank unit matrix,
0 is K × (B-K) rank null matrix, and K is the degree of rarefication of sound-source signal；
Step 6: deformed sparse expression model obtained in step 5 being solved by DCS-SOMP algorithm, is obtained
Obtain the strength of sound source under different frequency point at each grid node；
Step 7: the strength of sound source solved on each Frequency point at each grid node being overlapped, and then is obtained each
Overall sound pressure level OASPL at grid node:
WhereinFor Frequency point fjStrength of sound source at lower n-th of grid node；
Step 8: sound source position is determined according to the overall sound pressure level OASPL at each grid node.
Beneficial effect
The beneficial effects of the present invention are: method provided by the invention combines SVD and DCS-SOMP algorithm, DCS- is inherited
Efficient, the easy convergent characteristic of SOMP algorithm.Simultaneously by the signal subspace of measurement data being extracted, to improve after SVD
Noise immunity.Compared with the method described in the document, method provided by the invention, still can be right when signal-to-noise ratio is down to -20dB
Sound source is positioned.
Additional aspect and advantage of the invention will be set forth in part in the description, and will partially become from the following description
Obviously, or practice through the invention is recognized.
Detailed description of the invention
Above-mentioned and/or additional aspect of the invention and advantage will become from the description of the embodiment in conjunction with the following figures
Obviously and it is readily appreciated that, in which:
Attached drawing 1 is the flow chart of compressed sensing based broadband sound source positioning.
Attached drawing 2 is the flow chart of DCS-SOMP algorithm.
Attached drawing 3 is the propagation model figure of auditory localization.
Attached drawing 4 is the spectrogram of emulation broadband sound source signal used.
Attached drawing 5 is positioning result of the method provided by the present invention in the case where signal-to-noise ratio is -20dB.
Attached drawing 6 is positioning result of the DCS-SOMP algorithm in the case where signal-to-noise ratio is -20dB.
Specific embodiment
The embodiment of the present invention is described below in detail, the embodiment is exemplary, it is intended to it is used to explain the present invention, and
It is not considered as limiting the invention.
Referring to Fig.1-6, performance evaluation is carried out to the method proposed by emulating.Using be located at S1 (- 0.2,0,0.8),
Four monopole sound sources of S2 (0.2,0,0.8), S3 (- 0.2,0,0.8) and S4 (- 0.2,0,0.8) are as the sound source in emulation
Signal, attached drawing 4 are the spectrogram of broadband sound source signal.Auditory localization process is specific as follows:
Step 1: sound-source signal being acquired by the array that the bore being made of 60 microphones is 1m, sample rate is
44.1kHz samples duration 10s, obtains the measurement data of time domain.
Step 2: piecemeal, adding window and Fast Fourier Transform (FFT) being carried out to the time-domain measurement data that step 1 acquires and handled, is obtained
Obtain frequency-domain measurement data.
When piecemeal, each data block includes 1024 data points, carries out to each data block plus Hanning window is handled, data block
Duplication be 50%, the measurement data of frequency domain is obtained by Fast Fourier Transform (FFT).In emulation, white Gaussian noise is added,
Setting signal-to-noise ratio is -20dB.
Step 3: choosing required frequency segment data, the corresponding data of each Frequency point in selected frequency band are y (fj)=
(y1j,...,yMj)T, wherein the dimension of y is M × B, and B is by the data block total number that divides in step 2, fjFor in selected frequency band
J-th of Frequency point.
It is 2000-2500Hz that frequency band is arranged in the present embodiment, chooses the frequency-domain measurement data in this section of frequency range, obtains frequency
Measurement data in section at each Frequency point.
Step 4: establishing compressed sensing based broadband sound source positioning joint sparse expression model, the net including sound source plane
Lattice divide and the foundation of calculation matrix.Specifically includes the following steps:
Step 4.1: determining sound source plane, and grid dividing is carried out to sound source plane, be possible with each grid node
Sound source constructs unknown sound-source signal x, and wherein sound-source signal x is made of the strength of sound source at grid node.
Fig. 3 gives the propagation model of auditory localization.Selection sound source plane is 0.8 × 0.8m2Rectangular area, sound source is flat
Face is 0.8m at a distance from microphone array.Sound source number is 4 in emulation, and sound source plane is divided into the net of 60 × 60=3600
Lattice.
Step 4.2: according to the Helmoltz establishing equation Frequency point f of free field Green's functionjUnder sound source plane and wheat
Calculation matrix A (the f of gram wind arrayj):
Wherein i is imaginary unit, and c is the velocity of sound, dmnFor the distance between m-th of microphone and n-th of grid node.
Step 4.3: establish the auditory localization sparse expression model under each Frequency point:
y(fj)=A (fj)x(fj)+e(fj), j=1,2 ..., J
Wherein x (fj) it is Frequency point fjUnder sound-source signal, e (fj) it is Frequency point fjUnder measured value in Gauss white noise
Sound, J are frequency points.
Step 4.4: sparse expression model under different frequency point being combined, compressed sensing based broadband sound source is established
Position joint sparse expression model:
Step 5: to the y (f in joint sparse expression modelj) SVD decomposition is carried out, obtain deformed sparse expression mould
Type:
To the microphone array frequency domain measurement value y (f in joint sparse expression modelj) singular value decomposition is carried out, it is surveyed
Magnitude y (fj), sound-source signal x (fj), noise item e (fj) signal subspace:
y(fj)SV=y (fj)VDK, x (fj)SV=x (fj)VDK, e (fj)SV=e (fj)VDK
Wherein DK=[IK 0]T, V is to measured value y (fj) unitary matrice after singular value decomposition, IKFor K × K rank unit square
Battle array, 0 is K × (B-K) rank null matrix, and K is the degree of rarefication of sound-source signal, in the present embodiment, K=4.
Step 6: deformed sparse expression model obtained in step 5 being solved by DCS-SOMP algorithm, is obtained
Obtain the strength of sound source under different frequency point at each grid node, detailed process are as follows:
(a) it initializes: enabling iteration count value t=1.For a certain Frequency point j ∈ (1,2 ..., J), regularization is initialized
Because of subvectorInitialize the column index collection selectedEnable rj,tIndicate measured value y (fj) passing through t iteration
Remaining residual error afterwards, and initialize rj,1=y (fj)。
(b) it selects: the residual error under different frequency is projected to respectively in each column vector of corresponding calculation matrix, and will be different
The corresponding projection value of same column index is summed under frequency.Selection, and should so that the maximum matrix column index of the sum of projection
Index is added in the indexed set of previous step selection.That is:
In formula, A (fj)nFor frequency fjLower calculation matrix A (fj) the n-th column vector.
(c) regularization: carry out Schmidt's regularization, relative in previous ones selectively by the measurement of regularization
The calculation matrix column vector selected in current iteration is carried out regularization by matrix column vector.Have:
In formula, γj,tThe calculation matrix column vector selectedRegularization as a result, its be equal to willIt carries out
After QR is decomposed, Q value and vectorThe product of mould.
(d) iteration: the corresponding factor estimated value of selected vector is updatedWith residual error rj,tValue.Factor estimated valueSolve system of equation can be passed throughLeast square solutionIt obtains.Residual error rj,tIt is solved equal to previous step
Residual error afterwards subtracts least square solution in current iterationCorresponding measured value.That is:
In formula:For selected row vector in calculation matrixAfter carrying out QR decomposition, gained R value
With vectorMould is divided by resulting quotient.
(e) convergence judgement: if t≤K, return step (b).Otherwise, step (f) is executed；
(f) regularization is solved: to matrixQR decomposition is carried out, is had:
Wherein, Γ (fj)=[γj,1,γj,2,...,γj,T],
The estimated value of signalIt can be solved by following formula:
After obtaining the sound source approximate solution under each frequency point, the strength of sound source under different frequent points is calculated:
Step 7: the strength of sound source solved on each Frequency point at each grid node being overlapped, and then is obtained each
Overall sound pressure level OASPL at grid node:
WhereinFor Frequency point fjStrength of sound source at lower n-th of grid node；
Step 8: sound source position is determined according to the overall sound pressure level OASPL at each grid node.
Attached drawing 5 and attached drawing 6 are set forth provided in broadband sound source localization method and background technology document of the invention
Positioning result of the method in the case where signal-to-noise ratio is -20dB.Square indicates that real sources employed in emulation, dot indicate logical in figure
Cross sound localization method reconstruct sound source obtained.It can be found that being mentioned in background technology document when signal-to-noise ratio is equal to -20dB
The method of confession is excessive to auditory localization error, and method provided by the present invention can position each sound source.With background
Sound localization method in technical literature is compared, and method provided by the present invention has stronger noise immunity.
Although the embodiments of the present invention has been shown and described above, it is to be understood that above-described embodiment is example
Property, it is not considered as limiting the invention, those skilled in the art are not departing from the principle of the present invention and objective
In the case where can make changes, modifications, alterations, and variations to the above described embodiments within the scope of the invention.
Claims (1)
1. a kind of compressed sensing based broadband sound source localization method, it is characterised in that: the following steps are included:
Step 1: by the array acquisition sound-source signal containing M microphone, obtaining the measurement data of time domain；
Step 2: piecemeal, adding window and Fast Fourier Transform (FFT) being carried out to the time-domain measurement data that step 1 acquires and handled, frequency is obtained
Domain measurement data；
Step 3: choosing required frequency segment data, the corresponding data of each Frequency point in selected frequency band are y (fj)=
(y1j,...,yMj)T, wherein the dimension of y is M × B, and B is by the data block total number that divides in step 2, fjFor in selected frequency band
J-th of Frequency point；
Step 4: compressed sensing based broadband sound source positioning joint sparse expression model is established, specifically includes the following steps:
Step 4.1: it determines sound source plane, and grid dividing is carried out to sound source plane, with each grid node for possible sound source,
Unknown sound-source signal x is constructed, wherein sound-source signal x is made of the strength of sound source at grid node；
Step 4.2: according to the Helmoltz establishing equation Frequency point f of free field Green's functionjUnder sound source plane and microphone array
Calculation matrix A (the f of columnj):
Wherein i is imaginary unit, and c is the velocity of sound, dmnFor the distance between m-th of microphone and n-th of grid node；
Step 4.3: establish the auditory localization sparse expression model under each Frequency point:
y(fj)=A (fj)x(fj)+e(fj), j=1,2 ..., J
Wherein x (fj) it is Frequency point fjUnder sound-source signal, e (fj) it is Frequency point fjUnder measured value in include noise item, J is
Frequency points；
Step 4.4: sparse expression model under different frequency point being combined, compressed sensing based broadband sound source positioning is established
Joint sparse expression model:
Step 5: to the y (f in joint sparse expression modelj) SVD decomposition is carried out, obtain deformed sparse expression model:
To the microphone array frequency domain measurement value y (f in joint sparse expression modelj) singular value decomposition is carried out, obtain measured value y
(fj), sound-source signal x (fj), noise item e (fj) signal subspace:
y(fj)SV=y (fj)VDK, x (fj)SV=x (fj)VDK, e (fj)SV=e (fj)VDK
Wherein DK=[IK0]T, V is to measured value y (fj) unitary matrice after singular value decomposition, IKIt is K for K × K rank unit matrix, 0
× (B-K) rank null matrix, K are the degree of rarefication of sound-source signal；
Step 6: deformed sparse expression model obtained in step 5 being solved by DCS-SOMP algorithm, is obtained not
Strength of sound source under same frequency point at each grid node；
Step 7: the strength of sound source solved on each Frequency point at each grid node being overlapped, and then obtains each grid
Overall sound pressure level OASPL at node:
WhereinFor Frequency point fjStrength of sound source at lower n-th of grid node；
Step 8: sound source position is determined according to the overall sound pressure level OASPL at each grid node.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910408263.9A CN110221249A (en) | 2019-05-16 | 2019-05-16 | Compressed sensing based broadband sound source localization method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910408263.9A CN110221249A (en) | 2019-05-16 | 2019-05-16 | Compressed sensing based broadband sound source localization method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110221249A true CN110221249A (en) | 2019-09-10 |
Family
ID=67821250
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910408263.9A Pending CN110221249A (en) | 2019-05-16 | 2019-05-16 | Compressed sensing based broadband sound source localization method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110221249A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112946577A (en) * | 2021-02-01 | 2021-06-11 | 东南大学 | Ultra-short baseline underwater sound source positioning method based on broadband compressed sensing |
| CN113376576A (en) * | 2020-07-23 | 2021-09-10 | 郑州大学 | Positioning method of sound source positioning sensor based on small-aperture microphone array |
| CN113721194A (en) * | 2021-07-30 | 2021-11-30 | 南京师范大学 | MWCS-based near-field speech signal three-dimensional positioning algorithm |
| CN115201753A (en) * | 2022-09-19 | 2022-10-18 | 泉州市音符算子科技有限公司 | Low-power-consumption multi-spectral-resolution voice positioning method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100054085A1 (en) * | 2008-08-26 | 2010-03-04 | Nuance Communications, Inc. | Method and Device for Locating a Sound Source |
| CN102565759A (en) * | 2011-12-29 | 2012-07-11 | 东南大学 | Binaural sound source localization method based on sub-band signal to noise ratio estimation |
| CN103592628A (en) * | 2013-11-12 | 2014-02-19 | 上海大学 | Multi-sound-source positioning method based on formation of real value weight beam in spherical harmonic domain |
| CN107247251A (en) * | 2017-06-20 | 2017-10-13 | 西北工业大学 | Three-dimensional sound localization method based on compressed sensing |
-
2019
- 2019-05-16 CN CN201910408263.9A patent/CN110221249A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100054085A1 (en) * | 2008-08-26 | 2010-03-04 | Nuance Communications, Inc. | Method and Device for Locating a Sound Source |
| CN102565759A (en) * | 2011-12-29 | 2012-07-11 | 东南大学 | Binaural sound source localization method based on sub-band signal to noise ratio estimation |
| CN103592628A (en) * | 2013-11-12 | 2014-02-19 | 上海大学 | Multi-sound-source positioning method based on formation of real value weight beam in spherical harmonic domain |
| CN107247251A (en) * | 2017-06-20 | 2017-10-13 | 西北工业大学 | Three-dimensional sound localization method based on compressed sensing |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113376576A (en) * | 2020-07-23 | 2021-09-10 | 郑州大学 | Positioning method of sound source positioning sensor based on small-aperture microphone array |
| CN112946577A (en) * | 2021-02-01 | 2021-06-11 | 东南大学 | Ultra-short baseline underwater sound source positioning method based on broadband compressed sensing |
| CN112946577B (en) * | 2021-02-01 | 2023-12-22 | 东南大学 | Ultra-short baseline underwater sound source positioning method based on broadband compressed sensing |
| CN113721194A (en) * | 2021-07-30 | 2021-11-30 | 南京师范大学 | MWCS-based near-field speech signal three-dimensional positioning algorithm |
| CN113721194B (en) * | 2021-07-30 | 2024-04-09 | 南京师范大学 | Near-field voice signal three-dimensional positioning method based on MWCS |
| CN115201753A (en) * | 2022-09-19 | 2022-10-18 | 泉州市音符算子科技有限公司 | Low-power-consumption multi-spectral-resolution voice positioning method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110221249A (en) | Compressed sensing based broadband sound source localization method | |
| CN108375763B (en) | Frequency division positioning method applied to multi-sound-source environment | |
| CN109407045B (en) | Non-uniform sensor array broadband signal direction-of-arrival estimation method | |
| EP2628316B1 (en) | Apparatus and method for deriving a directional information and computer program product | |
| CN103886207B (en) | Nested MIMO radar DOA estimation method based on compressed sensing | |
| CN110045323B (en) | Matrix filling-based co-prime matrix robust adaptive beamforming algorithm | |
| CN109407046A (en) | A kind of nested array direction of arrival angle estimation method based on variational Bayesian | |
| Moore et al. | Direction of arrival estimation using pseudo-intensity vectors with direct-path dominance test | |
| CN110208735A (en) | A kind of DOA Estimation in Coherent Signal method based on management loading | |
| CN102147458B (en) | Method and device for estimating direction of arrival (DOA) of broadband sound source | |
| Ning et al. | Sound source localization of non-synchronous measurements beamforming with block Hermitian matrix completion | |
| CN109298383A (en) | A kind of relatively prime battle array direction of arrival angle estimation method based on variational Bayesian | |
| CN106021637A (en) | DOA estimation method in co-prime array based on iteration sparse reconstruction | |
| Loesch et al. | Blind source separation based on time-frequency sparseness in the presence of spatial aliasing | |
| Tourbabin et al. | Theoretical framework for the optimization of microphone array configuration for humanoid robot audition | |
| CN107576931A (en) | A kind of correlation based on the sparse reconstruct of covariance low dimensional iteration/coherent signal Wave arrival direction estimating method | |
| Leong et al. | Multiple target localization using wideband echo chirp signals | |
| CN105929386A (en) | Wave arrival estimation method based on high-order accumulated amount | |
| CN109343003B (en) | Method for identifying sound source formed by fast iterative shrinking wave beams | |
| CN103837858B (en) | A kind of far field direction of arrival estimation method for planar array and system | |
| CN109884592A (en) | A kind of auditory localization emulation mode towards low frequency Gaussian noise source | |
| CN110907892B (en) | Method for estimating arrival angle of voice signal of ball microphone array | |
| Nesta et al. | Enhanced multidimensional spatial functions for unambiguous localization of multiple sparse acoustic sources | |
| Shi et al. | Spatial time-frequency DOA estimation based on joint diagonalization using Jacobi rotation | |
| CN108564962A (en) | Unmanned plane voice signal Enhancement Method based on tetrahedron microphone array |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190910 |
|
| WD01 | Invention patent application deemed withdrawn after publication |