Circular acoustic vector-sensor array row near field sources Multiple Parameter Estimation Methods
Technical field
The invention belongs to array signal process technique field, more particularly to the near field sources that a kind of acoustic vector-sensor array is arranged are frequently
The method of estimation of rate, two dimensional arrival angles and distance.
Background technology
Acoustic vector sensors are pointed to by a non-direction sound pressure sensor and three are mutually perpendicular with dipole
The particle vibration velocity sensor of property is composited, can synchro measure acoustic signals sound pressure and particle vibration velocity information, thus
The fields such as radar, sonar, communication, space flight and aviation have obtained increasingly extensive application, when information source falls into the Fresnel of array aperture
It is referred to as near field sources during region, the positioning near field sources needs estimated distance and angle of arrival multiple parameters, based on acoustic vector sensing
The near-field sound source position positioning of device array has turned into the study hotspot of domestic and foreign scholars.
Array high-resolution method for parameter estimation based on subspace theory has been applied to Near-field sources target positioning, Liu Nan
Nanmu is in Jilin University's learned person position scholar's paper (exercise question in 2014:Near field sources multi-parameter inversion based on acoustic vector sensors) middle utilization
Two-dimentional Multiple Signal Classification method carries out the estimation of parameters of near field sources under white Gaussian noise background, and the method can suppress Gaussian stationary and make an uproar
Sound, but need to do the MUSIC search based on biquaternion, it is necessary to above do three-dimensional search, angle of arrival in two dimensional arrival angles and distance
Estimate that the raising of resolution ratio depends on the fine array region of search, thus have the shortcomings that computationally intensive.The present invention is using equal
Nicely rounded ideophone spectra of acoustic vector sensor array, it is proposed that the ESPRIT of near field sources angle of arrival, frequency and distance estimates signal ginseng
Number (ESPRIT) and Multiple Signal Classification method (MUSIC) method for parameter estimation, the present invention makes full use of acoustic vector sensors first
Intrinsic invariable rotary structure itself, parameter Estimation is carried out using ESPRIT algorithms, give near field sources direction of arrival and
The rough estimate value of sound source distance, then searched near coarse value using MUSIC algorithms obtain direction of arrival and sound source away from
From fine estimation, the method need not in population parameter space three-dimensional search, and Parameter automatic pair;Therefore, amount of calculation
Greatly reduce, and in the case where element number of array is limited, improve acoustic vector-sensor array row near field sources under near field sources scene
Parameter Estimation Precision.For near-field sound source signal, the direction of phase difference between array element not only with array element spacing and incoming signal has
Close and with the distance dependent of sound source to array element, so far field condition has the even linear array of translation invariant structure, uniform L gusts
Deng near field without translation invariant structure, so existing document is seldom related to the ESPRIT parameter Estimations of near field sources to calculate
Method, the invariable rotary relation that the present invention makes full use of the array of acoustic vector sensors intrinsic carries out parameter Estimation.
The content of the invention
It is many it is an object of the invention to provide a kind of near field sources circle acoustic vector-sensor array row based on ESPRIT and MUSIC
Parametric joint method of estimation.
To achieve these goals, the present invention takes following technical solution:
Circular acoustic vector-sensor array row near field sources Multiple Parameter Estimation Methods, K different frequency, orthogonal arrowband, with
The steady near-field sound source signal of machine respectively from different directions with different distance (θk, φk, rk) incide circular acoustic vector sensing
On device array, θkIt is the angle of pitch of incoming signal, φkIt is the azimuth of incoming signal, rkIt is k-th signal and origin of coordinates battle array
The distance between unit, the circular array is equidistantly spaced from radius for the array element on the circumference of R is constituted, with circumference by M
The center of circle is the origin of coordinates and places identical acoustic vector sensors as reference array element in the origin of coordinates, and the array element is by acoustic pressure
Sensor and x-axis, y-axis and z-axis direction vibration velocity sensor group into acoustic vector sensors, the respective channel of all the sensors is mutual
It is parallel:All of sound pressure sensor is parallel to each other, and all of x-axis direction vibration velocity sensor is parallel to each other, and all of y-axis direction shakes
Fast sensor is parallel to each other, and all of z-axis direction vibration velocity sensor is parallel to each other, and x-axis, y-axis and z-axis vibration velocity sensor
It is mutually perpendicular to two-by-two;On circular array between adjacent array element at intervals of λmin/ (8sin (π/M)), λminIt is incident acoustic wave signal
Meet Near Field between minimum wavelength, and radius of circle R and array element interval and the wavelength of incident acoustic wave signal and the distance of sound source;
The step of near field sources Multiple Parameter Estimation Methods, is as follows:
Step one, the reception data that near field acoustic signals are obtained using circular array;
The n times snapshot data of the circular reception array received signal constitutes direct sampled data Z1, receiving signal delayed Δ T
N times synchronization snapshot data afterwards constitutes time delay sampled data Z2, by Z1And Z2This two groups of data are constituted and receive the total evidence of signalWhereinfsIt is Nyquist sampling frequency;
Step 2, using receiving the full data acquisition signal subspace of signal and noise subspace;
Estimate full data correlation matrixWherein, A is full data-signal steering vector
Matrix, Rs=SSH/ N is incoming signal correlation matrix, σ2It is the power of white Gaussian noise, I is the unit matrix of 8M × 8M, according to
Subspace theory, to data correlation matrix RZCarry out feature decomposition and obtain signal subspace UsWith noise subspace UN,Wherein, EVD represents feature decomposition, λiIt is ith feature value that feature decomposition is obtained,
viIt is the corresponding ith feature vector of characteristic value, Us=[v1..., vK] it is the corresponding characteristic vector composition of the individual characteristic values greatly of K
Signal subspace, UN=[vK+1..., v8M] noise subspace that constitutes for the corresponding characteristic vector of the small characteristic values of 8M-K;
Step 3, estimation signal guide vector matrixAnd signal frequency
By the signal subspace U of 8M × KsIt is divided into the matrix U of upper and lower two pieces of 4M × K1And U2, closed using time invariable rotary
Architecture, by U1And U2Ψ T=Ω T are obtained by matrix operation, whereinFeature decomposition is carried out to matrix Ψ, it is special
Value indicative constitutes matrixCharacteristic vector constitutes matrixWhereinIt is the estimate of Ω,It is the estimate of T, so as to obtain letter
The estimation of number steering vector matrix and signal frequency:
Wherein, arg () expressions take argument,Representing matrixThe row k kth column element for taking,Us
=AT, U1=A1T, U2=A2T, A are exactly the full data array steering vector matrix in step 2, A1It is direct sampled data array
Steering vector matrix, A2It is time delay sampled data array steering vector matrix, T is the nonsingular matrix of K × K,It is matrix U1Pseudo inverse matrix,
Step 4, composition and array arrangement according to signal guide vector, obtain k-th x-axis of signal, y-axis vibration velocity and
Acoustic pressure obtains the rough estimate value of angle of arrival and sound source distance using normalized vector to the normalized vector of z-axis vibration velocity;
The arrangement form of structure and array according to acoustic vector sensors, by signal guide vector estimateKth (1
≤ k≤K) row be divided into the corresponding matrix-block of each array element,RepresentKth row,
M-th kth row of array element is represented,X-axis, three components of y-axis direction vibration velocity and acoustic pressure all
Normalized vector is obtained compared with the vibration velocity component of z-axis directionK-th signal is averagely obtained by the M normalized vector of array element
Normalized vectorDirection of arrival rough estimate value is obtained according to these ratio relationsHarmony
Source is apart from rough estimate value
The normalized vector of k-th signal is:
So as to obtain angle of arrival and
The estimate of distance:
Wherein,It is the distance between k-th signal and origin of coordinates array element, λk
It is k-th wavelength of signal, ρ0It is environment liquid density, c is acoustic wave propagation velocity, Vectorial Γk
The 1st, 2,4 elements, exp () is to ask exponent arithmetic, tan () and arctan () to represent ask tangent and arc tangent respectively
Computing;
Step 5, the fine estimation for searching for direction of arrival and distance near coarse value using MUSIC algorithms;
Using the structure type of circular array, the full data array be given in the zonule near rough estimate value is oriented to arrow
AmountUsing MUSIC spectrum peak search methods
Search obtains the fine estimation of signal near coarse value;
Wherein,It is the phase difference structure of signal arrival array element m and reference array element
Into spatial domain steering vector,For incident sound source is believed
Number reach the phase difference of array element m and reference array element, UnIt is noise subspace that step 2 is obtained, the x-axis of unit energy signal, y
The vibration velocity component and acoustic pressure scalar in axle and z-axis direction be:
θ, φ, r are search variables,
WithPoint
It is not the rough estimate value at the azimuth, the angle of pitch and distance in step 4, εθ、εφAnd εrIt is respectively intended to set the angle of pitch, orientation
Angle and the region of search length of distance;
K=1 ..., K, m=1 ..., M in abovementioned steps, n=1 ..., M, j represent imaginary unit.
The array that the present invention is used is Homogeneous Circular array, and the array element of array is by sound pressure sensor and x-axis, y-axis and z-axis
The acoustic vector sensors that the vibration velocity sensor in direction is constituted, and all of sound pressure sensor is parallel to each other, all of x-axis direction
Vibration velocity sensor is parallel to each other, and all of y-axis direction vibration velocity sensor is parallel to each other, all of z-axis direction vibration velocity sensor phase
It is mutually parallel.
The present invention gives a kind of circular acoustic vector-sensor array row near field sources multi-parameter ESPRIT algorithm for estimating, for one
As scalar sensors array such as microphone array cannot utilize ESPRIT algorithm estimation of near field sound-source signals parameter because closely
The corrugated of field is spherical wave;Phase between array element it is not only relevant with the direction of array element spacing and incoming signal and also with sound source to battle array
The distance dependent of unit, so far field condition has the even linear array of translation invariant structure, uniform L gusts etc., does not have near field
Translation invariant structure, it is impossible to utilize ESPRIT algorithms, the present invention makes full use of acoustic vector sensors rotation intrinsic in itself not
Structure changes, using ESPRIT algorithms enter line parameter rough estimate and with MUSIC zonules essence search method be combined realize angle of arrival and away from
From accurate estimation, compared with existing fourth order statistic and second-order statistic method, the inventive method amount of calculation is small, and has
Simple, the easy to use advantage of algorithm.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
The accompanying drawing for having technology to be needed to use in describing does simple introduction, it should be apparent that, drawings in the following description are only the present invention
Some embodiments, for those of ordinary skill in the art, on the premise of not paying creative work, can also basis
These accompanying drawings obtain other accompanying drawings.
Fig. 1 is the schematic diagram of embodiment of the present invention acoustic vector-sensor array row;
Fig. 2 is the flow chart of the inventive method;
Fig. 3 is the angle-of- arrival estimation scatter diagram of the inventive method of emulation experiment;
Fig. 4 is the orientation angular estimation root-mean-square error of the inventive method with the change curve of signal to noise ratio;
Fig. 5 is the pitching angular estimation root-mean-square error of the inventive method with the change curve of signal to noise ratio;
Fig. 6 is the angle-of- arrival estimation root-mean-square error of the inventive method with the change curve of signal to noise ratio;
Fig. 7 is the distance estimations root-mean-square error of the inventive method with the change curve of signal to noise ratio;
Fig. 8 is the angle-of- arrival estimation probability of success of the inventive method with the change curve of signal to noise ratio.
Specific embodiment
In order to above and other objects of the present invention, feature and advantage can be become apparent from, the embodiment of the present invention cited below particularly,
And coordinate appended diagram, it is described below in detail.
Fig. 1 show the schematic diagram of the acoustic vector-sensor array row of the embodiment of the present invention.Acoustic vector sensors of the invention
Array is equidistantly spaced from radius for the array element on the circumference of R is constituted by M, and the center of circle with circumference is as the origin of coordinates and in coordinate
Origin places identical acoustic vector sensors as reference array element, between the adjacent array element of the circular array at intervals of λmin/
(8sin (π/M)), the array element of array be by sound pressure sensor and x-axis, y-axis and z-axis direction vibration velocity sensor group into acoustic vector
Sensor, wherein, λminIt is the minimum wavelength of incoming signal.
The step of reference picture 2, Multiple Parameter Estimation Methods of the invention, is as follows:Circular acoustic vector-sensor array row receive K
Different frequency, orthogonal arrowband, random steady near-field sound source signal, K is the quantity of incident sound-source signal, K≤M-1,
Step one, the reception data that near field acoustic signals are obtained using circular array;
The n times snapshot data of the circular reception array received signal constitutes direct sampled data Z1, receiving signal delayed Δ T
N times synchronization snapshot data afterwards constitutes time delay sampled data Z2, by Z1And Z2This two groups of data are constituted and receive the total evidence of signalWhereinfsIt is Nyquist sampling frequency;
Step 2, using receiving the full data acquisition signal subspace of signal and noise subspace;
Estimate full data correlation matrixWherein, A is full data-signal steering vector
Matrix, Rs=SSH/ N is incoming signal correlation matrix, σ2It is the power of white Gaussian noise, I is the unit matrix of 8M × 8M, according to
Subspace theory, to data correlation matrix RZCarry out feature decomposition and obtain signal subspace UsWith noise subspace UN,Wherein, EVD represents feature decomposition, λiIt is ith feature that feature decomposition is obtained
Value, viIt is the corresponding ith feature vector of characteristic value, Us=[v1..., vK] it is the K corresponding characteristic vector structure of big characteristic value
Into signal subspace, UN=[vK+1..., v8M] noise that constitutes for the corresponding characteristic vector of the small characteristic values of 8M-K is empty
Between;
Step 3, estimation signal guide vector matrixAnd signal frequency
By the signal subspace U of 8M × KsIt is divided into the matrix U of upper and lower two pieces of 4M × K1And U2, closed using time invariable rotary
Architecture, by U1And U2Ψ T=Ω T are obtained by matrix operation, whereinFeature decomposition is carried out to matrix Ψ, it is special
Value indicative constitutes matrixCharacteristic vector constitutes matrixWhereinIt is the estimate of Ω,It is the estimate of T, so as to obtain letter
The estimation of number steering vector matrix and signal frequency:
Wherein, arg () expressions take argument,Representing matrixThe row k kth column element for taking,
Us=AT, U1=A1T, U2=A2T, A are exactly the full data array steering vector matrix in step 2, A1It is direct sampled data battle array
Row steering vector matrix, A2It is time delay sampled data array steering vector matrix, T is the nonsingular matrix of K × K,It is matrix U1Pseudo inverse matrix,
Step 4, composition and array arrangement according to signal guide vector, obtain k-th x-axis of signal, y-axis vibration velocity and
Acoustic pressure obtains the rough estimate value of angle of arrival and sound source distance using normalized vector to the normalized vector of z-axis vibration velocity;
The arrangement form of structure and array according to acoustic vector sensors, by signal guide vector estimateKth (1
≤ k≤K) row be divided into the corresponding matrix-block of each array element,RepresentKth row,
M-th kth row of array element is represented,X-axis, three components of y-axis direction vibration velocity and acoustic pressure all
Normalized vector is obtained compared with the vibration velocity component of z-axis directionK-th signal is averagely obtained by the M normalized vector of array element
Normalized vectorDirection of arrival rough estimate value is obtained according to these ratio relationsHarmony
Source is apart from rough estimate value
The normalized vector of k-th signal is:
So as to obtain angle of arrival and
The estimate of distance:
Wherein,It is the distance between k-th signal and origin of coordinates array element, λk
It is k-th wavelength of signal, ρ0It is environment liquid density, c is acoustic wave propagation velocity, Vectorial Γk
The 1st, 2,4 elements, exp () is to ask exponent arithmetic, tan () and arctan () to represent ask tangent and arc tangent respectively
Computing;
Step 5, the fine estimation for searching for direction of arrival and distance near coarse value using MUSIC algorithms;
Using the structure type of circular array, the full data array be given in the zonule near rough estimate value is oriented to arrow
AmountUsing MUSIC spectrum peak search methods
Search obtains the fine estimation of signal near coarse value;
Wherein,It is the phase difference structure of signal arrival array element m and reference array element
Into spatial domain steering vector,It is incident sound source
Signal reaches the phase difference of array element m and reference array element, UnIt is noise subspace that step 2 is obtained, the x-axis of unit energy signal,
The vibration velocity component and acoustic pressure scalar in y-axis and z-axis direction be:
θ, φ, r are search variables,
WithPoint
It is not the rough estimate value at the azimuth, the angle of pitch and distance in step 4, εθ、εφAnd εrIt is respectively intended to set the angle of pitch, orientation
Angle and the region of search length of distance;
K=1 ..., K, m=1 ..., M in abovementioned steps, n=1 ..., M, j represent imaginary unit.
The present invention estimates signal array steering vector and letter using the time invariable rotary structure that data before and after time delay have
Number frequency, is divided into x-axis, y-axis, four submatrixs of z-axis and acoustic pressure, using the ratio of corresponding element between submatrix by by array steering vector
Value relation obtains the rough estimate value of direction of arrival and sound source distance, and MUSIC algorithms are utilized in rough estimate value near zone
Precise search is carried out to obtain accurate direction of arrival and range estimation, the algorithm avoids answering for fourth order statistic algorithm
The universe three-dimensional search problem of polygamy and simple MUSIC algorithms, combines the having of ESPRIT and MUSIC algorithms and does not need
Parameter matches computing.
Effect of the invention can be further illustrated by following simulation result:
Emulation experiment condition is as follows:
Two different frequencies, orthogonal arrowband, random steady near-field sound source signal incide by 9 be equidistantly spaced from
Radius for R circumference on array element constitute circular acoustic vector-sensor array row, as shown in figure 1, adjacent array element at intervals of
λmin/ (8sin (π/9)), the parameter of incoming signal is:(θ1, φ120 °, 50 ° of)=(), (θ2, φ230 °, 70 ° of)=(), it is returned
One changes frequency for (f1, f2)=(0.3,0.4), fast umber of beats is 1024 times, 200 independent experiments.
The simulation experiment result as shown in Figures 3 to 8, Fig. 3 for signal to noise ratio be 15dB when, the inventive method angle-of- arrival estimation
Scatter diagram, the as can be seen from Figure 3 smaller range of the azimuth of the inventive method and angle of pitch estimate all near actual value
Interior distribution, the angle-of- arrival estimation of the inventive method has Parameter Estimation Precision higher;Present invention side is can be seen that from Fig. 4 and Fig. 7
The root-mean-square error of the azimuth of method, the angle of pitch, angle of arrival and distance estimations is smaller, that is, estimate near true value compared with
A small range is disturbed, and the angle of arrival and distance estimations of the inventive method have Parameter Estimation Precision higher, because of the invention
Method takes full advantage of acoustic vector sensors invariable rotary structure intrinsic in itself, and it is thick to enter line parameter using ESPRIT algorithms
Estimate, and processed by the precise search of MUSIC small areas and improve Parameter Estimation Precision;The angle-of- arrival estimation probability of success is
Refer to that the angle of pitch and azimuth estimate meet in 200 independent experimentsExperiment number account for always
The percentage of experiment number;Wherein, θ0And φ0It is true value,WithRefer to the estimate of i & lt experiment, from figure 8, it is seen that
The inventive method has the probability of success higher, and when signal to noise ratio is -10dB, the probability of success of signal one and signal two is respectively
20% and 30%, and signal to noise ratio is when being 0dB, two probability of succesies of signal are all higher than 80%.
The above, is only presently preferred embodiments of the present invention, and any formal limitation is not done to the present invention, though
So the present invention is disclosed above with preferred embodiment, but is not limited to the present invention, any to be familiar with this professional technology people
Member, without departing from the scope of the present invention, when making a little change or modification using the technology contents of the disclosure above
It is the Equivalent embodiments of equivalent variations, as long as being the content without departing from technical solution of the present invention, according to technical spirit of the invention
Any simple modification, equivalent variations and the modification made to above example, still fall within the range of technical solution of the present invention.