CN104360316A - Array antenna self-adaptive beam forming method based on covariance matrix tapering - Google Patents

Abstract

The invention belongs to the technical field of self-adaptive beam forming and particularly relates to an array antenna self-adaptive beam forming method based on covariance matrix tapering. The array antenna self-adaptive beam forming method specifically includes steps of firstly acquiring sample data by array antennas, calculating a covariance matrix of the sample data, and utilizing the covariance matrix as the estimated value of the covariance matrix of interference and noise; modifying the covariance matrix by the covariance matrix tapering method; solving guiding vector of optimization estimation desired signals with restrictions. The final simulation result proves the effectiveness of the provided two modification methods.

Description

A kind of array antenna Adaptive beamformer method be tapered based on covariance matrix

Technical field

The invention belongs to Adaptive beamformer technical field, particularly a kind of array antenna Adaptive beamformer method be tapered based on covariance matrix.

Background technology

Adaptive beamformer, as one of the mark of Array Signal Processing, has obtained deep research in the past few decades, has been widely used in multiple fields such as radar, wireless telecommunications, sonar, seismic prospecting and medical imaging.Along with the application in engineering practice, because array antenna inevitably exists various error, as contaminated in element position agitation error, mutual coupling, array element response error, training sample and number of training is very few etc., estimated value and the actual value mismatch of the steering vector of wanted signal are caused in capital, impact is in various degree caused on Adaptive beamformer performance, makes the requirement of the robustness to Adaptive beamformer also more and more higher.Therefore, the research of sane Adaptive beamformer (robust adaptivebeamforming, RAB) method is seemed become more and more important.

Array antenna adopts the even linear array along course made good usually, along time dimension, sampling is carried out to reception data and obtains sample data, the sample data obtained is utilized to estimate the covariance matrix of interference plus noise, then utilize the steering vector of the estimated value calculation expectation signal of covariance matrix and the Signal to Interference plus Noise Ratio (SINR) of output, wherein estimated accuracy and performance loss are mainly by the impact of sample number.At present for when comprising wanted signal in small sample number and training sample data, effectively can estimate that the robust adaptive beamforming method of the steering vector of wanted signal is less, and the output performance of these robust ada-ptive beamformer methods also has a certain distance with output performance ideally.

Summary of the invention

The object of the invention is to propose a kind of array antenna Adaptive beamformer method be tapered based on covariance matrix, the present invention utilizes covariance matrix to be tapered technology and first revises sample data covariance, then estimates the steering vector of wanted signal.

For achieving the above object, the sample data covariance matrix obtained under the present invention is directed to small sample number and the contaminated situation of training sample, first utilize covariance matrix to be tapered the covariance matrix of technology to sample data to revise, recycle the steering vector of covariance matrix calculation expectation signal and the Signal to Interference plus Noise Ratio of output of revised sample data, the present invention adopts following technical scheme to be achieved.

A kind of array antenna Adaptive beamformer method be tapered based on covariance matrix comprises the following steps:

Step 1, utilizes array antenna received from the signal of outside, and array antenna is the even linear array be made up of M array element, and the array element distance of array antenna is expressed as d, d=λ/2, and λ represents the wavelength of array antenna received signals; Array antenna is expressed as x (k) at the signal of k reception, and the signal of array antenna received comprises wanted signal, undesired signal and noise;

Step 2, draws the estimated value of the covariance matrix of interference plus noise

Step 3, utilizes covariance matrix to be tapered the estimated value of method to the covariance matrix of interference plus noise improve, draw estimated value after the improvement of the covariance matrix of interference plus noise wherein, the Hadamard of representing matrix amasss, T be M × M dimension be tapered matrix, the element T of capable n-th row of m of matrix T _{mZ (m, n)}for:

$T_{\mathrm{MZ}(m,n)}=\frac{\sin \left((m-n)\mathrm{\Δ}\right)}{(m-n)\mathrm{\Δ}}=\frac{\sin (\mathrm{\πW}(m-n)/2)}{\mathrm{\πW}(m-n)/2}$

Wherein, m=1,2 ..., M, n=1,2 ..., M; Δ=W pi/2 W represents zero width being trapped in sin θ ', and the span of θ ' is 0 ° ~ 180 °;

Step 4, represents the positive semidefinite matrix of M × M dimension by matrix A, set up the following Optimized model about matrix A:

subject to Tr(A)＝M

$\mathrm{Tr}\left(\tilde{C}A\right)\≤{\mathrm{\Δ}}_0$

A≥0

Wherein, the mark of Tr () representing matrix, subscript-1 representing matrix inverse, Θ represents the angular regions only comprising the incident angle of wanted signal of setting, b represents the angular regions of the incident angle of arbitrary signal; D (θ)=[1, e ^{j2 π d/ λ sin θ}..., e ^{j2 π d/ λ (M-1) sin θ}] ^h, θ represents the incident angle of wanted signal, the conjugate transpose of subscript H representing matrix; the implication of A>=0 is: A is positive semidefinite matrix;

Step 5, the optimization problem described in solution procedure 4, draws the optimum solution A of matrix A ^*, by matrix A ^*decompose by following formula, A ^*=Y*Y ^h, wherein, Y represents the sequency spectrum matrix that M × r ties up;

Draw the estimated value of the steering vector of the wanted signal of array antenna received if r=1, if r>1, then wherein, v is the vector that r × 1 is tieed up, and $v^H{Y}^H\tilde{C}\mathrm{Yv}=\mathrm{Tr}\left(Y^H\tilde{C}Y\right),$ Wherein, || || implication be modulus value;

Step 6, the estimated value of the steering vector of the wanted signal of the array antenna received drawn according to step 5 carry out array antenna Adaptive beamformer.

Beneficial effect of the present invention is: when comprising desired signal components in training sample, and the present invention effectively can estimate the steering vector of wanted signal.Compare existing robust adaptive beamforming method, the prior imformation wanted required for the present invention is less, and less demanding to the levels of precision of prior imformation.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of a kind of array antenna Adaptive beamformer method be tapered based on covariance matrix of the present invention;

Fig. 2 is the array structure schematic diagram observing reception in emulation experiment 1 and emulation experiment 2;

Fig. 3 is the change curve schematic diagram of the emulation experiment 1 output Signal to Interference plus Noise Ratio that adopts the present invention and existing several method to draw respectively with input signal-to-noise ratio;

Fig. 4 is the change curve schematic diagram of the emulation experiment 1 output Signal to Interference plus Noise Ratio that adopts the present invention and existing several method to draw under fixing input signal-to-noise ratio is the condition of 10dB respectively with fast umber of beats;

Fig. 5 is the change curve schematic diagram of the emulation experiment 2 output Signal to Interference plus Noise Ratio that adopts the present invention and existing several method to draw respectively with input signal-to-noise ratio;

Fig. 6 is the change curve schematic diagram of the emulation experiment 2 output Signal to Interference plus Noise Ratio that adopts the present invention and existing several method to draw under fixing input signal-to-noise ratio is the condition of 10dB respectively with fast umber of beats.

Embodiment

Below in conjunction with accompanying drawing, the invention will be further described:

With reference to Fig. 1, it is the process flow diagram of a kind of array antenna Adaptive beamformer method be tapered based on covariance matrix of the present invention.This Adaptive beamformer method being tapered technology based on covariance matrix comprises the following steps:

Step 1, utilizes array antenna received from the signal of outside, and array antenna is the even linear array (being designated as ULA) be made up of M array element, and the array element distance of array antenna is expressed as d, d=λ/2, and λ represents the wavelength of array antenna received signals; Array antenna is expressed as x (k) at the signal of k reception, and the signal of array antenna received comprises wanted signal, undesired signal and noise.

Specifically, in step 1, utilize array antenna received from the signal of outside, array antenna is the even linear array (being designated as ULA) be made up of M array element, the array element distance of array antenna is expressed as d, d=λ/2, and λ represents the wavelength of array antenna received signals.In the embodiment of the present invention, the signal of array antenna received is arrowband plane wave.Then array antenna is expressed as at signal x (k) of k reception:

x(k)＝s(k)+i(k)+n(k)

Wherein, k=1,2 ..., K, K represent the frame number of the signal of array antenna received, and s (k) represents the wanted signal (be such as echo signal) of array antenna at k reception, and it is the vector that M × 1 is tieed up; I (k) represents the undesired signal of array antenna at k reception, and it is the vector that M × 1 is tieed up; N (k) represents the noise of array antenna at k reception, and it is the vector that M × 1 is tieed up.In the embodiment of the present invention, wanted signal, undesired signal and noise are uncorrelated mutually.

Array antenna is expressed as at wanted signal s (k) of k reception:

s(k)＝s(k)a，

Wherein, s (k) represents the amplitude data of array antenna at the wanted signal of k reception, a represents the steering vector of the wanted signal of array antenna received, and a is the function of incident angle (wanted signal is incident to the angle of the array antenna) θ of wanted signal, can be expressed as:

a＝[1,e ^{j2πd/λsinθ},…,e ^{j2πd/λ(M-1)sinθ}] ^H

Wherein, d is the array element distance of array antenna, the conjugate transpose of subscript H representing matrix, and M is the array number of array antenna, and obviously, a is the vector that M × 1 is tieed up.

Step 2, draws the estimated value of the covariance matrix of interference plus noise

Wherein, K represents the frame number of the signal of array antenna received, and x (i) represents the signal of array antenna at i reception, i=1,2 ..., the conjugate transpose of K, subscript H representing matrix.The signal in each moment of array antenna received represents a training sample, then K represents number of training, and x (i) represents i-th training sample.

Step 3, utilizes covariance matrix to be tapered (CMT, covariance matrix taper) method to the estimated value of the covariance matrix of interference plus noise improve, draw estimated value after the improvement of the covariance matrix of interference plus noise wherein, the Hadamard of representing matrix amasss, and T is and matrix what have identical dimensional is tapered matrix, and the structure of matrix T adopts MZ method, i.e. the element T of capable n-th row of the m of matrix T _{mZ (m, n)}for:

$T_{\mathrm{MZ}(m,n)}=\frac{\sin \left((m-n)\mathrm{\Δ}\right)}{(m-n)\mathrm{\Δ}}=\frac{\sin (\mathrm{\πW}(m-n)/2)}{\mathrm{\πW}(m-n)/2}$

Wherein, m=1,2 ..., M, n=1,2 ..., M; Δ=W pi/2, W represents zero width being trapped in sin θ ', and the span of θ ' is 0 ° ~ 180 °.

Step 4, represents the positive semidefinite matrix of M × M dimension by matrix A, set up the following Optimized model about matrix A:

subject to Tr(A)＝M

$\mathrm{Tr}\left(\tilde{C}A\right)\≤{\mathrm{\Δ}}_0$

A≥0

Wherein, the mark of Tr () representing matrix, M is the array number of array antenna, represent estimated value after the improvement of the covariance matrix of interference plus noise, subscript-1 representing matrix inverse, Θ represents the angular regions only comprising the incident angle of wanted signal of setting, the supplementary angle of Θ, b represents the angular regions of the incident angle of arbitrary signal; D (θ) is the relevant steering vector with angle θ, and its structure is determined by the geometry of array antenna; D (θ)=[1, e ^{j2 π d/ λ sin θ}..., e ^{j2 π d/ λ (M-1) sin θ}] ^h, θ represents the incident angle of wanted signal, the conjugate transpose of subscript H representing matrix. that is, for arbitrarily maximal value be Δ ₀, the conjugate transpose of subscript H representing matrix; , the implication of A>=0 is: A is positive semidefinite matrix; for define symbol; it is the estimated value of the steering vector of the wanted signal of array antenna received.

Step 5, the optimization problem described in solution procedure 4, draws the optimum solution A of matrix A ^*.Due to by matrix A ^*order be designated as r, by matrix A ^*decompose by following formula, A ^*=Y*Y ^h, wherein, Y represents that (namely the order of Y is r) to the sequency spectrum matrix that M × r ties up, the conjugate transpose of subscript H representing matrix.

Then the estimated value of the steering vector of the wanted signal of array antenna received is drawn if r=1, if r>1, then wherein, v is the vector that r × 1 is tieed up, and v meets following two relational expressions:

$\mathrm{Yv}=\sqrt{M}$

$v^H{Y}^H\tilde{C}\mathrm{Yv}=\mathrm{Tr}\left(Y^H\tilde{C}Y\right)$

Wherein, || || implication be modulus value.In the embodiment of the present invention, v a kind of may solution be proportional to all proper vectors of matrix D and, matrix D is:

$D=\frac{1}{M}{Y}^{H}Y-\frac{Y^H\tilde{C}Y}{\mathrm{Tr}\left(Y^H\tilde{C}Y\right)}.$

Effect of the present invention can be further illustrated by emulation experiment 1 and emulation experiment 2.

Step 6, the estimated value of the steering vector of the wanted signal of the array antenna received drawn according to step 5 carry out array antenna Adaptive beamformer.

Specifically, Wave beam forming can be carried out by MUSIC (multiple signal classification) method, draw beam pattern p (θ),

$p\left(\mathrm{\θ}\right)=\frac{1}{{\hat{a}}^H{U}_n{U}_n^H\hat{a}}$

Wherein, U _nrepresent known noise subspace, the peak value of output represents the fine estimation of arrival bearing.

Emulation experiment 1: the steering vector mismatch considering the wanted signal caused by wavefront distortion.In this emulation experiment, we consider that the steering vector distortion of signal is caused by the propagation of ripple in nonhomogeneous media.

1) experiment condition, is provided with two experiment conditions altogether: experiment condition 1 and experiment condition 2.

Experiment condition 1: with reference to Fig. 2, for observing the array structure schematic diagram of reception in emulation experiment 1.In emulation experiment 1, the array antenna of Received signal strength is even linear array (many vertical lines drawn under horizontal linear in corresponding diagram 2), and the array number of array antenna is 10, and these 10 array elements are evenly placed in the horizontal direction, array element distance d=λ/2.Be provided with two interference sources, their incident direction is respectively 30 ° and 50 °, dryly makes an uproar than being 30dB, and the actual incident angle of wanted signal is θ _p=3 °, the spatial noise of interpolation is average to be zero variance be 1 white Gaussian noise, comprise desired signal components all the time in training sample.Simulation result obtains by 200 Monte-Carlo experiment statisticses, and in emulation experiment 1, wanted signal, interference and noise are mutual incoherent random Gaussian narrow band processs.Following multiple method is carried out com-parison and analysis by emulation experiment 1: the present invention, SDPR (Semi-definite Programming Relaxation) robust adaptive beamforming method, SMI (invert by Direct Sampling covariance matrix, Sample Matrix Inversion) method and SQP (seqential quadratic programming, Sequential Quadratic Programming) robust adaptive beamforming method, export Signal to Interference plus Noise Ratio (SINR) as a reference with optimum.For the present invention, SDPR robust adaptive beamforming method and SQP robust adaptive beamforming method, only comprise the angular regions of the incident angle of wanted signal for [θ _p-5 °, θ _p+ 5 °], for SQP robust adaptive beamforming method, Matrix C (list of references: A.Hassanien, S.A.Vorobyov, and K.M.Wong, " Robustadaptive beamforming using sequential programming:An iterative solution to themismatch problem; " IEEE Signal Processing Letters, vol.15, pp.733 – 736,2008.) dominant eigenvalue number is taken as 8, slack variable δ=0.1.For above often kind of method, number of training is 6.

Experiment condition 2: establish phase increment to be all constant in each Monte-Carlo experiment, separate and obey the Gaussian distribution that zero-mean variance is 0.04 between phase place.When adopting of the present invention, the value of W/2 is 0.02 (angle that is tapered being tapered matrix is 0.02).

2) experimental result.

Adopt the present invention and existing several method to obtain the estimated value of the steering vector of the wanted signal of array antenna received respectively, and then the performance of the output Signal to Interference plus Noise Ratio of the adaptive beam formed is compared.With reference to Fig. 3, the output Signal to Interference plus Noise Ratio adopting the present invention and existing several method to draw respectively for emulation experiment 1 is with the change curve schematic diagram of input signal-to-noise ratio.In Fig. 3, transverse axis represents input signal-to-noise ratio, and unit is dB, and the longitudinal axis represents output Signal to Interference plus Noise Ratio, and unit is dB.Optimum expression optimum exports Signal to Interference plus Noise Ratio curve (desirable output Signal to Interference plus Noise Ratio curve), improve one's methods and two represent the present invention, SDPR method represents SDPR robust adaptive beamforming method, and SQP method represents SQP robust adaptive beamforming method.As seen from Figure 3, along with the increase gradually of input signal-to-noise ratio, compared with SMI method, SDPR robust adaptive beamforming method and SQP robust adaptive beamforming method, Wave beam forming of the present invention can obtain higher output Signal to Interference plus Noise Ratio, wherein, SDPR robust adaptive beamforming method, compared with the present invention, exports the loss of Signal to Interference plus Noise Ratio or obvious.

Be under the condition of 10dB at fixing input signal-to-noise ratio, by the estimated value adopting the present invention and existing several method to obtain the steering vector of the wanted signal of array antenna received respectively, and then the performance of the output Signal to Interference plus Noise Ratio of the adaptive beam formed is compared.With reference to Fig. 4, for emulation experiment 1 is the output Signal to Interference plus Noise Ratio that adopts the present invention and existing several method to draw respectively under the condition of the 10dB change curve schematic diagram with fast umber of beats (number of training) at fixing input signal-to-noise ratio.In Fig. 4, transverse axis represents fast umber of beats, and the longitudinal axis represents output Signal to Interference plus Noise Ratio, and unit is dB.Optimum expression optimum exports Signal to Interference plus Noise Ratio curve (desirable output Signal to Interference plus Noise Ratio curve, be a horizontal linear), improve one's methods and two represent the present invention's (for curve), SDPR method represents SDPR robust adaptive beamforming method, and SQP method represents SQP robust adaptive beamforming method.As seen from Figure 4, the output Signal to Interference plus Noise Ratio of Wave beam forming of the present invention when less fast umber of beats (number of training) will significantly better than SMI method and SDPR robust adaptive beamforming method, also SQP robust adaptive beamforming method is better than, when fast umber of beats (number of training) is about 7, the loss of the output Signal to Interference plus Noise Ratio of SDPR robust adaptive beamforming method is very high, and the performance of the output Signal to Interference plus Noise Ratio of Wave beam forming of the present invention will significantly better than SDPR robust adaptive beamforming method.When fast umber of beats (number of training) is more than or equal to 10, performance and the SDPR robust adaptive beamforming method of the output Signal to Interference plus Noise Ratio of Wave beam forming of the present invention are consistent, and illustrate that the present invention has good stability under different fast umber of beats.

Emulation experiment 2: the steering vector mismatch considering the wanted signal caused due to angle mismatching.In emulation experiment 2, consider the incidence angle θ due to arrival bearing _pthe steering vector evaluated error of the wanted signal that mismatch produces.

1) experiment condition, is provided with two experiment conditions altogether: experiment condition 1 and experiment condition 2.

Experiment condition 1: identical with the experiment condition 1 of emulation experiment 1.

Experiment condition 2: mismatched angles size is set to 2 °.When adopting of the present invention, when adopting of the present invention, the value of W/2 is 0.02 (angle that is tapered being tapered matrix is 0.02)

2) experimental result

Under small sample said conditions, adopt the present invention and existing several method to obtain the estimated value of the steering vector of the wanted signal of array antenna received respectively, and then the performance of the output Signal to Interference plus Noise Ratio of the adaptive beam formed is compared.With reference to Fig. 5, the output Signal to Interference plus Noise Ratio adopting the present invention and existing several method to draw respectively for emulation experiment 2 is with the change curve schematic diagram of input signal-to-noise ratio.In Fig. 3, transverse axis represents input signal-to-noise ratio, and unit is dB, and the longitudinal axis represents output Signal to Interference plus Noise Ratio, and unit is dB.Optimum expression optimum exports Signal to Interference plus Noise Ratio curve (desirable output Signal to Interference plus Noise Ratio curve), improve one's methods and two represent the present invention, SDPR method represents SDPR robust adaptive beamforming method, and SQP method represents SQP robust adaptive beamforming method.As seen from Figure 5, along with the increase gradually of input signal-to-noise ratio, compared with SMI method, SDPR robust adaptive beamforming method and SQP robust adaptive beamforming method, Wave beam forming of the present invention can obtain higher output Signal to Interference plus Noise Ratio, wherein, SDPR robust adaptive beamforming method, compared with the present invention, exports the loss of Signal to Interference plus Noise Ratio or obvious.

Be under the condition of 10dB at fixing input signal-to-noise ratio, by the estimated value adopting the present invention and existing several method to obtain the steering vector of the wanted signal of array antenna received respectively, and then the performance of the output Signal to Interference plus Noise Ratio of the adaptive beam formed is compared.With reference to Fig. 6, for emulation experiment 2 is the output Signal to Interference plus Noise Ratio that adopts the present invention and existing several method to draw respectively under the condition of the 10dB change curve schematic diagram with fast umber of beats at fixing input signal-to-noise ratio.In Fig. 6, transverse axis represents fast umber of beats, and the longitudinal axis represents output Signal to Interference plus Noise Ratio, and unit is dB.Optimum expression optimum exports Signal to Interference plus Noise Ratio curve (desirable output Signal to Interference plus Noise Ratio curve, be a horizontal linear), improve one's methods and two represent the present invention's (for curve), SDPR method represents SDPR robust adaptive beamforming method, and SQP method represents SQP robust adaptive beamforming method.As seen from Figure 6, the output Signal to Interference plus Noise Ratio of Wave beam forming of the present invention when less fast umber of beats significantly better than SMI method and SDPR robust adaptive beamforming method, will slightly be better than SQP robust adaptive beamforming method.When fast umber of beats is more than or equal to 10, performance and the SDPR robust adaptive beamforming method of the output Signal to Interference plus Noise Ratio of Wave beam forming of the present invention are consistent, and illustrate that the present invention has good stability under different fast umber of beats.That is, the present invention is not only applicable to the condition of little number of training, and its Wave beam forming also has higher output Signal to Interference plus Noise Ratio under the condition of sufficient number of training, and usable range of the present invention is larger.

In sum, for robust adaptive beamforming, in existing method, although there are some solutions, major part is all process when sample number is more sufficient.In actual environment, in order to raise the efficiency, it is desirable to realize robust adaptive beamforming when only having a small amount of sample, the rate of exchange being made to the steering vector of wanted signal and estimates accurately.But when small sample number, when even comprising desired signal components in sample, the performance loss of the existing robust adaptive beamforming method of major part is very serious, and make estimated value and the actual value mismatch of the steering vector of wanted signal, evaluated error is large, precision is low.The present invention improves the covariance matrix of sample data by using covariance matrix to be tapered, effectively reduce the evaluated error of the covariance matrix of interference plus noise under small sample number, thus improve the estimated accuracy of the steering vector of wanted signal, and there is robustness.

Obviously, those skilled in the art can carry out various change and modification to the present invention and not depart from the spirit and scope of the present invention.Like this, if these amendments of the present invention and modification belong within the scope of the claims in the present invention and equivalent technologies thereof, then the present invention is also intended to comprise these change and modification.

Claims (4)

1., based on the array antenna Adaptive beamformer method that covariance matrix is tapered, it is characterized in that, comprise the following steps:

Step 1, utilizes array antenna received from the signal of outside, and array antenna is the even linear array be made up of M array element, and the array element distance of array antenna is expressed as d, d=λ/2, and λ represents the wavelength of array antenna received signals; Array antenna is expressed as x (k) at the signal of k reception, and the signal of array antenna received comprises wanted signal, undesired signal and noise;

Step 2, draws the estimated value of the covariance matrix of interference plus noise

Step 3, utilizes covariance matrix to be tapered the estimated value of method to the covariance matrix of interference plus noise improve, draw estimated value after the improvement of the covariance matrix of interference plus noise wherein, the Hadamard of representing matrix amasss, T be M × M dimension be tapered matrix, the element T of capable n-th row of m of matrix T _{mZ (m, n)}for:

$T_{\mathrm{MZ}(m,n)}=\frac{\sin \left((m-n)\mathrm{\Δ}\right)}{(m-n)\mathrm{\Δ}}=\frac{\sin (\mathrm{\πW}(m-n)/2)}{\mathrm{\πW}(m-n)/2}$

Wherein, m=1,2 ..., M, n=1,2 ..., M; Δ=W pi/2, W represents zero width being trapped in sin θ ', and the span of θ ' is 0 ° ~ 180 °;

Step 4, represents the positive semidefinite matrix of M × M dimension by matrix A, set up the following Optimized model about matrix A:

subject to Tr(A)＝M

$\mathrm{Tr}\left(\tilde{C}A\right)\≤{\mathrm{\Δ}}_0$

Wherein, the mark of Tr () representing matrix, subscript-1 representing matrix inverse, Θ represents the angular regions only comprising the incident angle of wanted signal of setting, b represents the angular regions of the incident angle of arbitrary signal; D (θ)=[1, e ^{j2 π d/ λ sin θ}..., e ^{j2 π d/ λ (M-1) sin θ}] ^h, θ represents the incident angle of wanted signal, the conjugate transpose of subscript H representing matrix; implication be: A is positive semidefinite matrix;

Step 5, the optimization problem described in solution procedure 4, draws the optimum solution A of matrix A ^*, by matrix A ^*decompose by following formula, A ^*=Y*Y ^h, wherein, Y represents the sequency spectrum matrix that M × r ties up;

Draw the estimated value of the steering vector of the wanted signal of array antenna received if r=1, if r>1, then wherein, v is the vector that r × 1 is tieed up, and $v^H{Y}^H\tilde{C}\mathrm{Yv}=\mathrm{Tr}\left(Y^H\tilde{C}Y\right),$ Wherein, || || implication be modulus value;

Step 6, the estimated value of the steering vector of the wanted signal of the array antenna received drawn according to step 5 carry out array antenna Adaptive beamformer.

2. a kind of array antenna Adaptive beamformer method be tapered based on covariance matrix as claimed in claim 1, it is characterized in that, in step 1, array antenna is expressed as at signal x (k) of k reception:

x(k)＝s(k)+i(k)+n(k)

Wherein, k=1,2, ..., K, K represent the frame number of the signal of array antenna received, and s (k) represents the wanted signal of array antenna at k reception, i (k) represents the undesired signal of array antenna at k reception, and n (k) represents the noise of array antenna at k reception;

Array antenna is expressed as at wanted signal s (k) of k reception:

s(k)＝s(k)a，

Wherein, s (k) represents the amplitude data of array antenna at the wanted signal of k reception, and a represents the steering vector of the wanted signal of array antenna received,

a＝[1,e ^{j2πd/λsinθ},…,e ^{j2πd/λ(M-1)sinθ}] ^H

Wherein, d is the array element distance of array antenna, the conjugate transpose of subscript H representing matrix, and M is the array number of array antenna.

3. a kind of array antenna Adaptive beamformer method be tapered based on covariance matrix as claimed in claim 1, is characterized in that, in step 2, and the estimated value of the covariance matrix of described interference plus noise for:

Wherein, K represents the frame number of the signal of array antenna received, and x (i) represents the signal of array antenna at i reception, i=1,2 ..., the conjugate transpose of K, subscript H representing matrix.

4. a kind of array antenna Adaptive beamformer method be tapered based on covariance matrix as claimed in claim 1, is characterized in that, in steps of 5, if r>1, then v and matrix D all proper vectors and be directly proportional, matrix D is:

$D=\frac{1}{M}{Y}^{H}Y-\frac{Y^H\tilde{C}Y}{\mathrm{Tr}\left(Y^H\tilde{C}Y\right)}.$