CN109639333A - A kind of Beamforming Method based on effective reconstruct covariance matrix - Google Patents
A kind of Beamforming Method based on effective reconstruct covariance matrix Download PDFInfo
- Publication number
- CN109639333A CN109639333A CN201811480843.0A CN201811480843A CN109639333A CN 109639333 A CN109639333 A CN 109639333A CN 201811480843 A CN201811480843 A CN 201811480843A CN 109639333 A CN109639333 A CN 109639333A
- Authority
- CN
- China
- Prior art keywords
- covariance matrix
- signal
- desired signal
- noise
- array
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/08—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station
- H04B7/0837—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station using pre-detection combining
- H04B7/0842—Weighted combining
- H04B7/086—Weighted combining using weights depending on external parameters, e.g. direction of arrival [DOA], predetermined weights or beamforming
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a kind of based on effective robust adaptive beamforming method for rebuilding covariance matrix.This method judges whether the factor reconstructs the standard of covariance matrix as algorithm by introducing, desired signal angle direction is estimated first with Power Spectrum Estimation Method, can desired signal angle be estimated the standard rebuild as covariance matrix, then the covariance matrix under different state of signal-to-noise is constructed respectively, without reconstructing covariance matrix i.e. under low signal-to-noise ratio, covariance matrix reconstructing method is used under high s/n ratio, array weight vector finally is sought, is finally reached and guarantees to reduce its computation complexity while Beam-former output performance.The present invention still has preferable output performance in high s/n ratio.Invention increases a threshold decision processes, can be estimated using desired signal as critical condition, can balance the relationship between complexity and output performance well.
Description
Technical field
The present invention relates to beam-forming technology field and field of signal processing, and in particular to one kind is based on effectively reconstruct association side
The Beamforming Method of poor matrix.
Background technique
Adaptive beamformer technology can adaptively change according to external environment variation characteristic with characteristics of signals is received
Become weighted vector size, and then guarantee that desired signal undistorted can receive and inhibit jamming performance farthest,
Therefore it can be suitable for more complicated engineer application environment, have broad application prospects.
Beam-former reached by array weight vector main lobe wave beam alignment desired signal come to and null be aligned
Interference signal come to purpose, and best initial weights vector is based on interference noise matrix and all accurate nothing of desired signal steering vector
In the case where accidentally.Interference noise matrix hardly results in actual application environment, typically by sample covariance matrix come generation
It replaces, but would generally include desired signal in sampled data, when desired signal power is smaller, the sampling covariance square sought at this time
Battle array is smaller with the gap of interference noise covariance matrix, the difference between the corresponding array weight vector sought and best initial weights vector
Away from also smaller, thus the output performance of Beam-former is influenced it is smaller, but when desired signal power is larger, Beam-former
Desired signal is inhibited as interference signal easily, and then algorithm output performance is caused to decline.
For this problem, a series of steady beamforming algorithms are suggested, and main representative algorithm is interference noise square
Battle array restructing algorithm.
" Beamforming Method of the interference noise covariance matrix reconstruct based on subspace " (number of patent application
201510680829.5, applying date 2015.10.19, applicant University of Electronic Science and Technology), it describes a kind of dry based on subspace
The Beamforming Method for disturbing noise covariance matrix reconstruct, replaces sample covariance square with the interference noise covariance matrix of reconstruct
Battle array, while a kind of new steering vector algorithm for estimating is proposed to estimate the steering vector of desired signal, improve Wave beam forming calculation
Method has good robustness to interference signal steering vector error, but this method does not consider the complexity that algorithm engineering is realized
Problem.
" the Adaptive beamformer method reconstructed based on relevant calculation and covariance matrix " (number of patent application
201510548956.X, applying date 2015.08.31, applicant Acoustical Inst., Chinese Academy of Sciences), it describes a kind of based on phase
The Adaptive beamformer method calculated with covariance matrix reconstruct is closed, is solved in sampling number of snapshots containing desired signal
In the case of, the deformation of adaptive direction figure main lobe and secondary lobe raising of SMI algorithm, desired signal cancellation, output Signal to Interference plus Noise Ratio are serious
The problem of decline, but this method only for be to receive in signal in the biggish situation of desired signal power, to desired signal
There is no effect when power is faint, complexity can increase instead.
A kind of " Adaptive beamformer method based on covariance matrix reconstruct " (number of patent application
201611055429.6, applying date 2016.11.25, applicant's line University of Electronic Science and Technology), it describes a kind of based on covariance square
Battle array reconstruct Adaptive beamformer method, this method by reconstruct interference plus noise covariance matrix, eliminate for calculate from
Desired signal components in the sample covariance matrix of Beam-former weight vector are adapted to, so that the performance of Beam-former is improved,
But this method be still for desired signal power it is big in the case where, not consider small-power desired signal under algorithm performance
Problem.
Paper " Robust Adaptive Beamforming Based on Interference Covariance
Matrix Reconstruction and Steering Vector Estimation " (author: Gu Y etc., periodical IEEE
Transactions on Signal Processing, date issued in July, 2012) it proposes to reconstruct using Estimation of Spatial Spectrum
The method of interference noise covariance matrix, but this method is needed to the entire spatial domain in addition to angular regions where desired signal
Accumulation calculating is carried out, calculation amount is very big.
Paper " Robust MVDR beamforming based on covariance matrix
Reconstruction " (author: Pengcheng M, periodical Science China Information Sciences are delivered
In April, 2013 on date) interference noise matrix, but the party reconstructed using sample covariance matrix removal desired signal compositions, method
Method can have large error in steering vector angle mismatching.
Paper " Robust adaptive beamforming via a novel subspace method for
Interference covariance matrix reconstruction " (author: Yuan X, periodical Signal
Processing, date issued in July, 2016) propose only need to calculate interference signal where angular regions reconstruct covariance square
Battle array, substantially reduces the complexity of algorithm.
Paper " Robust adaptive beamforming based on interference covariance
Matrix sparse reconstruction. " (author: Gu Y, periodical Signal Processing, date issued 2013
September) reconstruction of interference noise covariance matrix is reconstructed using sparse constraint to improve algorithm in sampled data containing expectation letter
Performance in the case of number situation and model mismatch, and its output performance is similar to optimal performance, but algorithm under low signal-to-noise ratio
The promotion of performance is much smaller than the raising of algorithm complexity, and few documents are related to the problem of covariance matrix effectively reconstructs,
Therefore aiming at the problem that effectively reconstruct covariance matrix, there is important research significance.
Summary of the invention
The object of the present invention is to provide a kind of Beamforming Methods based on effective reconstruct covariance matrix, solve existing at present
The performance boost in high input signal-to-noise ratio of the Beam-former of some interference noise covariance matrix reconstruct is higher, but low
Performance boost less complexity the problem of greatly increasing in the case of input signal-to-noise ratio.
In order to achieve the above objectives, the present invention provides a kind of Beamforming Method based on effective reconstruct covariance matrix,
Itself the following steps are included:
Step 1: using the even linear array of M array element composition, there is J+1 far field narrow band signal to be incident on array in space,
Assuming that as a reference point with No. 1 array element, signal incident direction angle is θ, other array elements are with the time delays between No. 1 array element
τm, and meet τm=(m-1) dsin θ/c, then can establish array received signal model；
Step 2: covariance matrix characteristic value distribution situation being obtained by feature decomposition method, is obtained in array received signal
Big characteristic value number；
Step 3: signal corresponding to big characteristic value is estimated using MUSIC algorithm, by setting power spectrum threshold value: critical door
ξ estimates angle corresponding to high-power signal；
Step 4: the angular regions Θ of known desired, and have Θ=[θmin,θmax], and assume desired signal incoming wave
Direction is θd, θd∈Θ；The angle direction estimated is set as θi(i=1,2 ..., P), wherein P is the signal source estimated
Number, by judging θiReconstruct association is determined with the presence or absence of being located within the scope of desired signal angular regions Θ in (i=1,2 ..., P)
The critical value of variance matrix: judge factor-beta；
Step 5: corresponding covariance matrix in the case of the different judgement factors is found out respectively, is specifically included:
Step 5.1: in low signal-to-noise ratio, covariance matrix being modified using noise average method, to disappear
The influence that sample covariance matrix is generated except limited number of snapshots；
Step 5.2: in high s/n ratio, the method reconstructed using interference noise covariance matrix, to improve algorithm
Performance；
Step 6: finally seeking corresponding array weight vector under different situations.
The above-mentioned Beamforming Method based on effective reconstruct covariance matrix, wherein in step 1, the array antenna
The establishment step of receipt signal model includes:
Step 1.1: array received signal model is write as:
In formula, J represents interference number, s0(t) desired signal, θ are represented0Desired signal is represented to sj(t) (j=1 ...,
J j-th of interference signal, θ) are representedj(j=1 ..., J) represents interference signal and comes to a (θ0) it is guiding corresponding to desired signal
Vector；a(θj) (j=1 ..., J) be steering vector corresponding to interference signal, n (t) is white Gaussian noise, and is defined
Step 1.2: it is assumed that irrelevant between desired signal and interference signal, then the true covariance matrix of array received data
It is write as:
RX=E [X (t) XH(t)]=Rs+Rj+n
In formula, Rs、Rj+nIt is desired signal, covariance matrix corresponding to interference and noise signal, () respectivelyHRepresent matrix
Transposition symbol；
Step 1.3: the true covariance matrix of array received data is replaced using sample covariance matrix:
R=X (t) XH(t)/K
In formula, K is sampling number of snapshots.
The above-mentioned Beamforming Method based on effective reconstruct covariance matrix, wherein in step 2, seek big characteristic value
Method can be described as:
In formula, λi(i=1,2 ..., M) is covariance matrix characteristic value, and meets λ1≥λ2≥…≥λM, Λ=diag
{λ1,…,λM, ei(i=1,2 ..., M) is characterized the corresponding feature vector of value, U=[e1,e2,…,eM]。
The above-mentioned Beamforming Method based on effective reconstruct covariance matrix, wherein in step 3, MUSIC algorithm is base
In minimizing what output power was realized, indicate are as follows:
In formula, θ represents spatial domain angle, UN(θ) is the corresponding subspace of noise signal, and a (θ) is space angle region pair
The steering vector answered, PMUSICAs output power corresponding to weighted vector.
The above-mentioned Beamforming Method based on effective reconstruct covariance matrix, wherein in step 4, it is known that desired signal
Angular regions Θ, and have Θ=[θmin,θmax], and assume that desired signal arrival bearing is θd, θd∈Θ；What setting estimated
Angle direction is θi(i=1,2 ..., P), wherein P is the signal source number estimated, by judging θi(i=1,2 ..., P) be
No includes desired signal to determine judgement factor-beta, is indicated are as follows:
|θi-θd|≤δ
Wherein, δ is critical factor, meets 0≤δ≤θmax-θmin|, when meeting above formula there are i (i=1,2 ..., P), say
Bright desired signal can be estimated, i.e. input SNR is larger, need to reconstruct interference noise covariance matrix at this time；And when upper
When formula is all invalid to any i (i=1,2 ..., P), illustrate that desired signal is not estimated, desired signal is to sampling at this time
The influence of covariance matrix is smaller, can ignore, and there is no need to reconstruct interference noise matrix.
Judge that factor-beta can indicate are as follows:
When judging factor-beta=1, expression needs to reconstruct covariance matrix, when judge factor-beta=0, indicate to input SNR compared with
Small, desired signal can not be estimated, and the influence to algorithm performance is smaller, not need reconstruct covariance matrix.
The above-mentioned Beamforming Method based on effective reconstruct covariance matrix, wherein in step 5,
When judging factor-beta=0, the corresponding characteristic value of noise signal is averaged as noise average power, is denoted as:
Wherein,For noise average power.Covariance matrix at this time becomes
When judging factor-beta=1, all possible directions of spatial spectral estimation algorithm estimation space Spectral structure are utilized, are indicated
Are as follows:
Wherein, P (θ) is the space spectrum of space angle.
Using above formula it is estimated that interference noise matrix, is denoted as:
In formula,Be in space remove expected angle domain Θ outside institute it is angled, andInclude institute in entire space
Some angle directions, andIntersection be sky, a (θ) be the corresponding steering vector in space angle region.
In order to reduce the complexity of algorithm, cumulative summation form is converted by the integrated form in above formula, is indicated are as follows:
Wherein, a (θl) be angle be θl(l=1,2 ..., L) corresponding steering vector, angular regionsIt is divided into L parts.
The above-mentioned Beamforming Method based on effective reconstruct covariance matrix, wherein in step 6, find out difference respectively
In the case of corresponding array weight vector:
When judging factor-beta=0, indicate that covariance matrix at this time is found out by step 5.1 without reconstructing covariance matrix,
Corresponding weighted vector indicates are as follows:
In formula, a is desired signal guide vector；
When judging factor-beta=1, expression needs to reconstruct covariance matrix, and covariance matrix at this time is found out by step 5.2,
Corresponding weighted vector indicates are as follows:
Compared with the existing technology, the invention has the following advantages:
(1) compared with traditional Wave beam forming class algorithm, the present invention still has preferable output in high s/n ratio
Performance.
(2) compared with other covariance matrix reconstructing methods, invention increases a threshold decision processes, utilize expectation
Signal, which can be estimated, can be good at balancing the relationship between complexity and output performance as critical condition.
(3) present invention is improved in low input signal-to-noise ratio using the method modified covariance method matrix that noise is average
Output performance of the algorithm under limited number of snapshots, and avoid covariance matrix reconstruct bring complexity and increase；And in height
The output performance of algorithm is substantially increased in the case of input signal-to-noise ratio using covariance matrix reconstructing method.
Detailed description of the invention
Fig. 1 is the even linear array structural schematic diagram that the present invention uses；
Fig. 2 is SNR=-15dB down-sampling covariance matrix characteristic value distribution situation figure；
Fig. 3 is SNR=15dB down-sampling covariance matrix characteristic value distribution situation figure；
Fig. 4 is corresponding angle estimation situation map under SNR=-15dB；
Fig. 5 is corresponding angle estimation situation map under SNR=15dB；
Fig. 6 is to judge relational graph between factor-beta and input SNR；
Fig. 7 is output SINR of the invention under SNR=-20dB with number of snapshots change curve；
Fig. 8 is output SINR of the invention under SNR=20dB with number of snapshots change curve；
Fig. 9 is the change curve of output SINR of the invention with input SNR.
Specific embodiment
Below in conjunction with attached drawing, by specific embodiment, the invention will be further described, these embodiments are merely to illustrate
The present invention is not limiting the scope of the invention.
The present invention is by analysing in depth influence of input Signal to Interference plus Noise Ratio (SNR) to Beam-former output performance, the present invention
It proposes a kind of based on effective robust adaptive beamforming method for rebuilding covariance matrix.This method judges the factor by introducing
The standard that covariance matrix whether is reconstructed as algorithm estimates desired signal angle side first with Power Spectrum Estimation Method
To, desired signal angle can be estimated as covariance matrix rebuild standard, then construct different noises respectively
Than the covariance matrix in the case of (SNR), i.e., without reconstructing covariance matrix at low signal-to-noise ratio (SNR), used at high SNR
Covariance matrix reconstructing method finally seeks array weight vector, while being finally reached guarantee Beam-former output performance
Reduce its computation complexity.
It mainly includes the following aspects that the present invention, which effectively reconstructs covariance matrix method:
1, array antenna received signals model is derived
As shown in fig. 1, the even linear array (uniform linear array, ULA) formed using M array element, array element
Spacing is d, wherein meeting d=1/2 λ, λ is the wavelength of desired signal.It is false when having J+1 far field narrow band signal incidence in space
If as a reference point with No. 1 array element, signal incident direction angle is θ, other array elements are τ with the time delays between No. 1 array elementm,
And meet τm=(m-1) dsin θ/c, then can array received model can be write as:
Wherein, J represents interference number, s0(t) desired signal, θ are represented0Represent desired signal to, sj (t) (j=1 ...,
J j-th of interference signal) is represented, θ j (j=1 ..., J), which represents interference signal, to be come to a (θ0) it is guiding corresponding to desired signal
Vector, a (θj) (j=1 ..., J) be steering vector corresponding to interference signal, n (t) is white Gaussian noise, and is defined
Often assume that it is irrelevant between desired signal and interference signal, then array received data true covariance matrix note
Are as follows:
RX=E [X (t) XH(t)]=Rs+Rj+n (2)
In formula, Rs、Rj+nIt is desired signal, covariance matrix corresponding to interference and noise signal, () respectivelyHRepresent matrix
Transposition symbol.
The usual true covariance matrix characteristic of array received data hardly results in, generally using sample covariance matrix come generation
It replaces, indicates are as follows:
R=X (t) XH(t)/K (3)
Wherein K is sampling number of snapshots.
2, covariance matrix characteristic value distribution situation is obtained by feature decomposition method, when desired signal power is greater than noise
When power, at this moment big characteristic value is corresponding for interference and desired signal power；And when desired power is lower than noise power, at this moment
Big characteristic value only corresponds to jamming power.It is described with formula are as follows:
In above formula, λi(i=1,2 ..., M) is covariance matrix characteristic value, and meets λ1≥λ2≥…≥λM, Λ=
diag{λ1,…,λM, ei(i=1,2 ..., M) it is characterized value λiCorresponding feature vector, U are feature vector eiThe spy being combined into
Levy space, U=[e1,e2,…,eM], R characteristic value distribution situation, such as Fig. 2 can be clearly found out by drawing eigenvalue graph
With shown in Fig. 3.
3, it by the characteristic value distribution situation of analytical sampling covariance matrix R, can obtain big special in array received signal
Value indicative number.Then angle estimation, such as Fig. 4 are carried out to the corresponding signal of big characteristic value using MUSIC algorithm on this basis again
With shown in Fig. 5, by setting power spectrum threshold value, i.e. critical threshold ξ estimates angle corresponding to signal.MUSIC algorithm is
Based on minimizing what output noise power was realized, indicate are as follows:
Wherein, θ represents spatial domain angle, UN(θ) is the corresponding subspace of noise signal, and a (θ) is space angle region pair
The steering vector answered, PMUSICAs output power corresponding to weighted vector, spectrum peak corresponding angle are to need to estimate to do
It disturbs or desired signal angle.
4, the angle direction being estimated that using above formula where high-power signal.Known desired angular regions
Θ, Θ=[θmin,θmax], it is assumed that desired signal arrival bearing is θd, θd∈Θ.Estimation angle direction is θi(i=1,2 ...,
P), wherein P represents the signal source number estimated, by judging θiWith the presence or absence of being located at desired signal in (i=1,2 ..., P)
The critical value that reconstruct covariance matrix is determined within the scope of angular regions Θ, that is, judge factor-beta.It is formulated are as follows:
|θi-θd|≤δ (6)
Wherein δ be critical factor, meet 0≤δ≤| θmax-θmin|, when meeting above formula there are i (i=1,2 ..., P), say
Bright desired signal can be estimated, i.e. expectation power is larger, and input SNR is larger, need to reconstruct interference noise matrix at this time；
And when above formula is all invalid to any i (i=1,2 ..., P), illustrate that desired signal is not estimated, at this time desired signal
Influence to sample covariance matrix is smaller, can ignore, and there is no need to reconstruct interference noise matrix.
Judge that the factor can be write as following form:
When judging factor-beta=1, expression needs to reconstruct covariance matrix, when judge factor-beta=0, indicates to input at this time
SNR is smaller, and desired signal can not be estimated, and the influence to algorithm performance is smaller, does not need reconstruct covariance matrix.
5, different condition judges that the method that covariance matrix is solved corresponding to factor-beta is different, and corresponding method for solving enters
Under:
(1) when judging factor-beta=0, indicate that output SNR is smaller at this time, desired signal can not be estimated, to calculation
The influence of method performance is smaller, does not need reconstruct covariance matrix.In order to eliminate limited number of snapshots to sample covariance matrix
It influences, the corresponding characteristic value of noise signal is averaged as noise average power, is denoted as:
Wherein,For noise average power.
Covariance matrix at this time becomes:
(2) when judging factor-beta=1, expression needs to reconstruct covariance matrix, uses the matrix based on Estimation of Spatial Spectrum here
Reconstructing method realizes that implementation method is as follows:
First with all possible directions of spatial spectral estimation algorithm estimation space Spectral structure, indicate are as follows:
Wherein, P (θ) is space spectrum.Using above formula it is estimated that interference noise matrix, is denoted as:
Wherein,Be in space remove expected angle domain Θ outside institute it is angled, andInclude institute in entire space
Some angle directions, andIntersection be sky.
In order to reduce the complexity of algorithm, cumulative summation form generally is converted by the integrated form in above formula, is indicated are as follows:
Wherein, a (θl) be angle be θl(l=1,2 ..., L) corresponding steering vector, angular regionsIt is divided into L parts.
θl(l=1,2 ..., L) it is angle value corresponding to every portion.
6, corresponding covariance matrix is different under different condition, and corresponding array weight vector is sought also not identical.
(1) when judging factor-beta=0, indicate that covariance matrix at this time is by (1) in step 5 without reconstructing covariance matrix
It finds out, corresponding weighted vector can be expressed as
Wherein, a is desired signal guide vector.
(2) when judging factor-beta=1, expression needs to reconstruct covariance matrix, and covariance matrix at this time is by (2) in step 5
It finds out, corresponding weighted vector can be expressed as
Effect of the invention can be illustrated by following emulation:
(1) simulated conditions and content:
1, the output performance of the Beam-former based on effective reconstruct covariance matrix
The setting of emulation experiment condition, the even linear array (ULA) of 16 array elements, array element spacing use half wavelength, it is assumed that expectation
Sense sets 0 °, and estimation desired signal direction sets 3 °, and SNR is set as 10dB, and two interference signals are come to being 50 °
With -40 °, dry make an uproar is set as 30dB than (INR), it is assumed that the desired signal direction that MUSIC algorithm estimates be θ1, interference signal side
To for θ2And θ3, noise is zero mean Gaussian white noise, and desired signal angular regions set Θ1=[θ1-5°,θ1+ 5 °], interference letter
Number angular regions set Θ2=[θ2-5°,θ2+ 5 °] and Θ3=[θ3-5°,θ3+ 5 °], simulation times are set as 100 Monte-
Carlo。
Here algorithm complexity problem is analyzed, at low input SNR, using the method for reconstruct covariance matrix to calculation
The promotion of method performance is little, but the computational complexity of algorithm is but by original O (M3) increase to O (M2S), wherein S > > M, this
In S indicate angular regionsImpartial divide points.And algorithm proposed by the present invention can be good at avoiding the problem, by right
The judgement of input signal signal-to-noise ratio decides whether reconstruct covariance matrix, under low input SNR using the average side of noise
Method carries out covariance matrix amendment, and it is lower to improve output performance and complexity of the algorithm under limited number of snapshots；And in height
It inputs under SNR, the performance of algorithm is largely improved using the method for reconstruct covariance matrix.Therefore proposed by the present invention
Method based on effective reconstruct covariance matrix has certain advantage in balance complexity and algorithm output performance.
Fig. 6 is input SNR and judges the relation curve between the factor.When judging factor-beta for 1, desired signal energy is indicated
It is enough estimated, needs to reconstruct interference noise matrix at this time；When judging factor-beta for 0, indicate that desired signal cannot be estimated
Out, reconstruct interference noise matrix is not needed at this time.As seen from the figure, it is whether covariance matrix is reconstructed when SNR is -8dB
Critical point need to reconstruct interference noise matrix when SNR is greater than -8dB, on the contrary it is then on the contrary.
Fig. 7 and Fig. 8 is that SINR is exported under SNR=-20dB and SNR=20dB with snap number variation relation.Fig. 7 is SNR
When=- 20dB, SINR is with snap number change curve for algorithm output, as seen from the figure, at low SNR, side proposed by the present invention
The output performance of method is 1dB higher or so than the output performance for reconstructing covariance matrix, and main cause is, at low SNR, reconstructs association side
The performance of poor matrix algorithm is not greatly improved, but since the present invention is in view of limited number of snapshots are to algorithm output performance
Influence and it is corrected by the average method of noise, so present invention algorithm performance under limited number of snapshots has
Certain advantage and complexity is significantly reduced.Algorithm output performance is with snap number when Fig. 8 is SNR=20dB
Variation diagram, as seen from the figure, at high SNR, the output performance of the method proposed by the present invention based on effective reconstruct covariance matrix
Isospace Power estimation algorithm is overlapped, this is because the method use of reconstruct covariance matrix is consistent, and is calculated much higher than MVDR
The output performance of method.Therefore algorithm proposed by the present invention is more significant in balance complexity and advantage on output performance.
Fig. 9 is the input SNR and output SINR variation relation figure of several algorithms.As seen from the figure, algorithm proposed by the present invention
For other restructing algorithms, in high SNR, performance is about the same, and in low SNR, performance is slightly lower than the association side after reconstructing
Poor matrix algorithm performance, but algorithm complexity is substantially reduced.
It is discussed in detail although the contents of the present invention have passed through above preferred embodiment, but it should be appreciated that above-mentioned
Description is not considered as limitation of the present invention.After those skilled in the art have read above content, for of the invention
A variety of modifications and substitutions all will be apparent.Therefore, protection scope of the present invention should be limited to the appended claims.
Claims (7)
1. a kind of Beamforming Method based on effective reconstruct covariance matrix, which comprises the following steps:
Step 1: using the even linear array of M array element composition, thering is J+1 far field narrow band signal to be incident on array in space, establish
Array received signal model；
Step 2: covariance matrix characteristic value distribution situation being obtained by feature decomposition method, is obtained big special in array received signal
Value indicative number；
Step 3: estimate signal corresponding to big characteristic value using MUSIC algorithm, pass through setting power spectrum threshold value: critical door ξ come
Estimate angle corresponding to high-power signal；
Step 4: the angular regions Θ of known desired, and have Θ=[θmin,θmax], and assume desired signal arrival bearing
For θd, θd∈Θ；The angle direction estimated is set as θi(i=1,2 ..., P), wherein P is the signal source number estimated, is led to
Cross judgement θiReconstruct covariance square is determined with the presence or absence of being located within the scope of desired signal angular regions Θ in (i=1,2 ..., P)
The critical value of battle array: judge factor-beta；
Step 5: corresponding covariance matrix in the case of the different judgement factors is found out respectively, is specifically included:
Step 5.1: in low signal-to-noise ratio, covariance matrix being modified using noise average method, is had to eliminate
The influence that limit number of snapshots generate sample covariance matrix；
Step 5.2: in high s/n ratio, the method reconstructed using interference noise covariance matrix, to improve the property of algorithm
Energy；
Step 6: finally seeking corresponding array weight vector under different situations.
2. the Beamforming Method as described in claim 1 based on effective reconstruct covariance matrix, which is characterized in that step 1
In, the establishment step of the array antenna received signals model includes:
Step 1.1: array received signal model is write as:
In formula, J represents interference number, s0(t) desired signal, θ are represented0Desired signal is represented to sj(t) (j=1 ..., J) generation
J-th of interference signal of table, θj(j=1 ..., J) represents interference signal and comes to a (θ0) it is steering vector corresponding to desired signal,
a(θj) (j=1 ..., J) be steering vector corresponding to interference signal, n (t) is white Gaussian noise, and is defined
Step 1.2: it is assumed that irrelevant between desired signal and interference signal, then the true covariance matrix of array received data is write
At:
RX=E [X (t) XH(t)]=Rs+Rj+n
In formula, Rs、Rj+nIt is desired signal, covariance matrix corresponding to interference and noise signal, () respectivelyHRepresent matrix transposition
Symbol；
Step 1.3: the true covariance matrix of array received data is replaced using sample covariance matrix:
R=X (t) XH(t)/K
In formula, K is sampling number of snapshots.
3. the Beamforming Method as described in claim 1 based on effective reconstruct covariance matrix, which is characterized in that step 2
In, the method for seeking big characteristic value can be described as:
In formula, λi(i=1,2 ..., M) is covariance matrix characteristic value, and meets λ1≥λ2≥…≥λM, Λ=diag
{λ1,…,λM},ei(i=1,2 ..., M) is characterized the corresponding feature vector of value, U=[e1,e2,…,eM]。
4. the Beamforming Method as described in claim 1 based on effective reconstruct covariance matrix, which is characterized in that step 3
In, MUSIC algorithm is realized based on minimum output power, is indicated are as follows:
In formula, θ represents spatial domain angle, UN(θ) is the corresponding subspace of noise signal, and a (θ) is that space angle region is corresponding
Steering vector, PMUSICAs output power corresponding to weighted vector.
5. the Beamforming Method as described in claim 1 based on effective reconstruct covariance matrix, which is characterized in that step 4
In, judge that factor-beta is whether the information source estimated in foundation step 3 includes desired signal to define, and indicates are as follows:
In formula, desired signal angular regions Θ, Θ=[θmin,θmax], it is assumed that desired signal arrival bearing is θd, θd∈ Θ, estimation
Angle direction is θi(i=1,2 ..., P), wherein P is the information source number estimated, and δ is critical factor, 0≤δ of satisfaction≤| θmax-
θmin|。
6. the Beamforming Method as described in claim 1 based on effective reconstruct covariance matrix, which is characterized in that step 5
In,
When judging factor-beta=0, the corresponding characteristic value of noise signal is averaged as noise average power, is denoted as:
Wherein,For noise average power, covariance matrix at this time becomes
When judging factor-beta=1, all possible directions of spatial spectral estimation algorithm estimation space Spectral structure are utilized, are indicated are as follows:
Wherein, P (θ) is space spectrum；
Using above formula it is estimated that interference noise matrix, is denoted as:
In formula,Be in space remove expected angle domain Θ outside institute it is angled, andComprising all in entire space
Angle direction, andIntersection be sky, a (θ) be the corresponding steering vector in space angle region；
In order to reduce the complexity of algorithm, cumulative summation form is converted by the integrated form in above formula, is indicated are as follows:
Wherein, a (θl) be angle be θl(l=1,2 ..., L) corresponding steering vector, angular regionsIt is divided into L parts.
7. the Beamforming Method as described in claim 1 based on effective reconstruct covariance matrix, which is characterized in that step 5
In, corresponding array weight vector under different situations is found out respectively:
When judging factor-beta=0, indicate that covariance matrix at this time is found out by step 5.1 without reconstructing covariance matrix, it is corresponding
Weighted vector indicate are as follows:
In formula, a is desired signal guide vector；
When judging factor-beta=1, expression needs to reconstruct covariance matrix, and covariance matrix at this time is found out by step 5.2, corresponding
Weighted vector indicate are as follows:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811480843.0A CN109639333B (en) | 2018-12-05 | 2018-12-05 | Beam forming method based on effective reconstruction covariance matrix |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811480843.0A CN109639333B (en) | 2018-12-05 | 2018-12-05 | Beam forming method based on effective reconstruction covariance matrix |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109639333A true CN109639333A (en) | 2019-04-16 |
| CN109639333B CN109639333B (en) | 2021-12-24 |
Family
ID=66071289
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811480843.0A Active CN109639333B (en) | 2018-12-05 | 2018-12-05 | Beam forming method based on effective reconstruction covariance matrix |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109639333B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110133588A (en) * | 2019-05-14 | 2019-08-16 | 普联技术有限公司 | A kind of antenna positioning method, device and equipment |
| CN110361697A (en) * | 2019-07-09 | 2019-10-22 | 西安电子科技大学 | A kind of robust ada- ptive beamformer method based on covariance matrix mixing reconstruct |
| CN111551892A (en) * | 2020-04-22 | 2020-08-18 | 武汉工程大学 | Steady self-adaptive beam forming method and device |
| CN111722178A (en) * | 2020-05-29 | 2020-09-29 | 东南大学 | Far-field narrow-band signal incoming wave direction estimation method based on numerical solution of directivity model |
| CN112379334A (en) * | 2020-11-02 | 2021-02-19 | 上海无线电设备研究所 | Adaptive beam forming method and device |
| CN114994595A (en) * | 2022-08-03 | 2022-09-02 | 四川太赫兹通信有限公司 | Direction-of-arrival acquisition method, terahertz phased array beam control method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110075601A1 (en) * | 2009-09-28 | 2011-03-31 | Dunmin Zheng | Systems and Methods for Adaptive Interference Cancellation Beamforming |
| CN105204006A (en) * | 2015-10-19 | 2015-12-30 | 电子科技大学 | Beam forming method based on subspace interference-plus-noise covariance matrix reconstruction |
| CN107167778A (en) * | 2017-07-03 | 2017-09-15 | 电子科技大学 | It is a kind of that the robust ada- ptive beamformer method estimated with steering vector is reconstructed based on covariance matrix |
| CN107544052A (en) * | 2017-08-07 | 2018-01-05 | 大连大学 | A kind of second-order statistic reconstruct DOA estimation method based on matrix completion |
-
2018
- 2018-12-05 CN CN201811480843.0A patent/CN109639333B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110075601A1 (en) * | 2009-09-28 | 2011-03-31 | Dunmin Zheng | Systems and Methods for Adaptive Interference Cancellation Beamforming |
| CN105204006A (en) * | 2015-10-19 | 2015-12-30 | 电子科技大学 | Beam forming method based on subspace interference-plus-noise covariance matrix reconstruction |
| CN107167778A (en) * | 2017-07-03 | 2017-09-15 | 电子科技大学 | It is a kind of that the robust ada- ptive beamformer method estimated with steering vector is reconstructed based on covariance matrix |
| CN107544052A (en) * | 2017-08-07 | 2018-01-05 | 大连大学 | A kind of second-order statistic reconstruct DOA estimation method based on matrix completion |
Non-Patent Citations (2)
| Title |
|---|
| YUJIE GU等: "Robust Adaptive Beamforming Based on Interference Covariance Matrix Reconstruction and Steering Vector Estimation", 《IEEE TRANSACTIONS ON SIGNAL PROCESSING》 * |
| 王昊等: "基于协方差矩阵重构的稳健自适应波束形成算法", 《探测与控制学报》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110133588A (en) * | 2019-05-14 | 2019-08-16 | 普联技术有限公司 | A kind of antenna positioning method, device and equipment |
| CN110361697A (en) * | 2019-07-09 | 2019-10-22 | 西安电子科技大学 | A kind of robust ada- ptive beamformer method based on covariance matrix mixing reconstruct |
| CN110361697B (en) * | 2019-07-09 | 2021-08-13 | 西安电子科技大学 | Robust beam forming method based on covariance matrix hybrid reconstruction |
| CN111551892A (en) * | 2020-04-22 | 2020-08-18 | 武汉工程大学 | Steady self-adaptive beam forming method and device |
| CN111551892B (en) * | 2020-04-22 | 2022-03-15 | 武汉工程大学 | Steady self-adaptive beam forming method and device |
| CN111722178A (en) * | 2020-05-29 | 2020-09-29 | 东南大学 | Far-field narrow-band signal incoming wave direction estimation method based on numerical solution of directivity model |
| CN111722178B (en) * | 2020-05-29 | 2022-07-08 | 东南大学 | Far-field narrow-band signal incoming wave direction estimation method based on numerical solution of directivity model |
| CN112379334A (en) * | 2020-11-02 | 2021-02-19 | 上海无线电设备研究所 | Adaptive beam forming method and device |
| CN112379334B (en) * | 2020-11-02 | 2023-03-14 | 上海无线电设备研究所 | Adaptive beam forming method and device |
| CN114994595A (en) * | 2022-08-03 | 2022-09-02 | 四川太赫兹通信有限公司 | Direction-of-arrival acquisition method, terahertz phased array beam control method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109639333B (en) | 2021-12-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109639333A (en) | A kind of Beamforming Method based on effective reconstruct covariance matrix | |
| Shahbazpanahi et al. | Robust adaptive beamforming for general-rank signal models | |
| US7478041B2 (en) | Speech recognition apparatus, speech recognition apparatus and program thereof | |
| CN107167778A (en) | It is a kind of that the robust ada- ptive beamformer method estimated with steering vector is reconstructed based on covariance matrix | |
| CN110045321B (en) | Robust DOA estimation method based on sparse and low-rank recovery | |
| CN103728594B (en) | Based on the external illuminators-based radar sea clutter disturbance restraining method of hyperchannel NLMS | |
| CN109254261B (en) | Coherent signal null deepening method based on uniform circular array EPUMA | |
| CN103984676A (en) | Rectangular projection adaptive beamforming method based on covariance matrix reconstruction | |
| CN109490850A (en) | Wideband array Adaptive beamformer method under major lobe suppression | |
| CN109407055A (en) | The Beamforming Method utilized based on multipath | |
| CN107153178B (en) | Target detection method for external radiation source radar reference signal with multipath interference | |
| CN107276658A (en) | The Beamforming Method reconstructed under coloured noise based on covariance matrix | |
| CN109597048B (en) | Meter wave radar DOA estimation method based on two-dimensional convolutional neural network | |
| CN110888105A (en) | DOA estimation method based on convolutional neural network and received signal strength | |
| CN109600152A (en) | A kind of Adaptive beamformer method based on the transformation of subspace base | |
| CN105699988B (en) | For the denoising MVDR cheating interference suppressing methods of navigation neceiver | |
| CN112462352B (en) | Line spectrum enhancement method suitable for low signal-to-noise ratio condition | |
| CN110082789B (en) | Space-time domain self-adaptive wide-linear rank-reduction beam forming method based on circular array | |
| CN106599551A (en) | Rapid adaptive beam-forming algorithm applied to array antenna soccer robot | |
| CN113051739A (en) | Robustness self-adaptive processing method based on sparse constraint | |
| CN110389319B (en) | MIMO radar DOA estimation method based on low-altitude multipath condition | |
| CN110161476A (en) | Radar beam forming method based on power iteration generalized Rayleigh quaotient algorithm | |
| Nayir et al. | Hybrid-field channel estimation for massive MIMO systems based on OMP cascaded convolutional autoencoder | |
| Haoyu et al. | Mainlobe interference suppression method based on blocking matrix preprocessing with low sidelobe constraint | |
| Yang et al. | Direction‐of‐arrival estimation method based on least‐squares by reconstructing covariance matrix with automatic diagonal loading |
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 |