CN109901148A - Broadband signal DOA estimation method based on covariance matrix rarefaction representation - Google Patents
Broadband signal DOA estimation method based on covariance matrix rarefaction representation Download PDFInfo
- Publication number
- CN109901148A CN109901148A CN201910218265.1A CN201910218265A CN109901148A CN 109901148 A CN109901148 A CN 109901148A CN 201910218265 A CN201910218265 A CN 201910218265A CN 109901148 A CN109901148 A CN 109901148A
- Authority
- CN
- China
- Prior art keywords
- frequency
- signal
- matrix
- angle
- vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000011159 matrix material Substances 0.000 title claims abstract description 84
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000010276 construction Methods 0.000 claims description 5
- 230000003595 spectral effect Effects 0.000 claims description 4
- 238000001228 spectrum Methods 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 238000001514 detection method Methods 0.000 abstract description 5
- 238000005516 engineering process Methods 0.000 abstract description 3
- 238000002474 experimental method Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 3
- 238000000354 decomposition reaction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 238000000342 Monte Carlo simulation Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000001427 coherent effect Effects 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Abstract
The invention belongs to Radar Signal Processing Technology field, the broadband signal DOA estimation method based on covariance matrix rarefaction representation is disclosed, for estimating broadband signal DOA.This method comprises: step 1, establishes broadband signal frequency-domain model；Step 2, it is based on the broadband signal frequency-domain model, the covariance matrix at reference frequency is obtained by focus processing；Step 3, rarefaction representation is carried out using the covariance matrix at the reference frequency, constructed about sparse vector b0Objective function；Step 4, the objective function is solved, sparse vector b is obtained0, and then determine according to the corresponding position of the nonzero element of the sparse vector incident angle of broadband incident signal.The present invention can accurately estimate the direction of arrival of signal, detection probability with higher and resolving accuracy without estimating information source number in the case where low signal-to-noise ratio, low snap, low angle interval.
Description
Technical field
The present invention relates to Radar Signal Processing Technology fields, more particularly to the letter of the broadband based on covariance matrix rarefaction representation
Number direction of arrival (Direction Of Arrival, DOA) estimation method.
Background technique
Array signal processing is an important research method of Radar Signal Processing, and Mutual coupling is array signal
The hot issue of processing.Incoming signal can regard sparse as on airspace, by the sparsity in incoming signal airspace and dilute
The thought that dredging indicates combines, thus it is applied to DOA estimation with its reasonability.
2005, Malioutov et al. proposed most classic sparse restructing algorithm-L1- on the basis of narrow band signal
SVD (Singular Value Decomposition) algorithm, algorithm are decomposed the received data of array element by singular value decomposition
For main signal section and secondary noise section, signal section is indicated using redundant dictionary, using second order Based On The Conic Model into
Row solves the DOA for estimating signal, and algorithm complexity substantially reduces.But this algorithm, there is also some defects, which exists
In treatment process, need known source information, when information source number is unknown or Sources number estimation mistake, will cause target missing or
The phenomenon that pseudo- target；In addition to this, algorithm has ignored noise subspace, so that estimating performance decline in low signal-to-noise ratio.2011
Year Yin and Chen proposes the L1-SRACV algorithm based on array covariance matrix rarefaction representation, the algorithm for narrow band signal
The covariance matrix obtained using array received data auto-correlation passes through the Estimation of Mean for the small characteristic value that Eigenvalues Decomposition obtains
Noise power, obtains estimation error of the covarianee matrix, and this way may not need the DOA of Sources number estimation signal.Practical application
Middle width strip signal largely exists, and the above two classes algorithm is not directly applicable in broadband signal, and in low signal-to-noise ratio, low fast
It claps, often degradation even fails performance in the case where low angle interval.
Summary of the invention
In view of this, the present invention provide the broadband signal DOA estimation method based on covariance matrix rarefaction representation, using turn
It changes matrix and dimensionality reduction is carried out to each column of covariance matrix, eliminate the influence of noise power.This way is not necessarily to estimate information source number,
Noise power, and can accurately estimate in the case where low signal-to-noise ratio, low snap, low angle interval the direction of arrival of signal,
Detection probability with higher and resolving accuracy.
In order to achieve the above objectives, the embodiment of the present invention adopts the following technical scheme that
A kind of broadband signal DOA estimation method based on covariance matrix rarefaction representation is provided, wherein receiving array is M
The even linear array of omnidirectional's array element composition, and array element spacing is d, the receiving array is received from P broadband of space far-field incidence
Incoming signal, incident angle are denoted as θ respectively1,θ2,…,θi,…,θP, the range of incident frequencies is [fl,fh], flIndicate incidence most
Low frequency, fhIndicate incident highest frequency, θiIndicate the incident angle of i-th of broadband incident signal, 1≤i≤P；
The described method includes:
Step 1, broadband signal frequency-domain model is established；
Step 2, it is based on the broadband signal frequency-domain model, the covariance square at reference frequency is obtained by focus processing
Battle array；
Step 3, rarefaction representation is carried out using the covariance matrix at the reference frequency, constructed about sparse vector b0's
Objective function
Wherein, B=[b1,…,bm,…,bM], bmThe m of representing matrix B is arranged, m=1,2 ... M, bmBe Q × 1 coefficient to
Amount, Two norms of representing matrix B q row,W indicates weighting matrix, W-1/2For the On Square-Rooting Matrices of weighting matrix W inverse matrix；G is that block is diagonal
Matrix,Gm=[e1,e2,…em-1,em+1,…,eM]T, GmFor the matrix of (M-1) × M, em
For the vector of M × 1, emM-th of element be 1, remaining element is 0；The vectorization of vec () representing matrix；Table
Show redundant dictionary；η indicates decision threshold；RYIndicate the covariance matrix at reference frequency；
Step 4, the objective function is solved, sparse vector b is obtained0, and then according to the nonzero element of the sparse vector
Corresponding position determines the incident angle of the P broadband incident signal.
Based on the broadband signal DOA estimation method provided by the invention based on covariance matrix rarefaction representation, make an uproar in analysis
Between acoustical power and covariance matrix on the basis of relationship, construction block diagonal matrix is to traditional based on covariance matrix
Sparse representation model improves, and eliminates the inaccurate influence to algorithm performance of noise power estimation by dimensionality reduction, therefore be not required to
It wants known or estimating noise power can estimate broadband signal direction of arrival, and between low signal-to-noise ratio, low snap, low angle
The direction of arrival of signal, detection probability with higher and resolving accuracy can be accurately estimated in the case where.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of estimation side broadband signal DOA based on covariance matrix rarefaction representation provided in an embodiment of the present invention
The flow diagram of method；
Fig. 2 is DOA estimated result figure of the method provided in an embodiment of the present invention under coherent signal；
Fig. 3 is that the DOA of method provided in an embodiment of the present invention at different conditions estimates performance schematic diagram；Wherein, Fig. 3
It (a) is that DOA of the method provided in an embodiment of the present invention under different signal-to-noise ratio estimates that performance schematic diagram, Fig. 3 (b) are the present invention
The method that embodiment provides estimates performance schematic diagram in the DOA of different number of snapshots now, and Fig. 3 (c) is the embodiment of the present invention
DOA of the method for offer under different angle intervals estimates performance schematic diagram.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Firstly, the application scenarios of the embodiment of the present invention are described as follows:
Receiving array receives the P broadband incident signal from space far-field incidence, and incident angle is denoted as θ respectively1,θ2,…,
θi,…,θP, the range of incident frequencies is [fl,fh], flIndicate incident low-limit frequency, fhIndicate incident highest frequency, θiIndicate the
The incident angle of i broadband incident signal, 1≤i≤P.
Next, the scheme to the embodiment of the present invention is specifically described:
Flow diagram shown in Figure 1, the width provided in an embodiment of the present invention based on covariance matrix rarefaction representation
Band signal DOA estimation method, comprising the following steps:
Step 1, broadband signal frequency-domain model is established.
Specifically, step 1 may include following sub-step:
(1a) determines the signal form that m-th of omnidirectional's array element receives:
Wherein, si(t) i-th of broadband signal is indicated,The delay that m-th of array element is reached for i-th of signal, for uniform
Linear arraynmIt (t) is the white Gaussian noise of m-th of array element；
(1b) carries out Fourier transformation to the signal that m-th of omnidirectional array element receives, and obtains frequency-domain model:
Wherein, J indicates the sub-band number that broadband signal divides on frequency domain, fjIndicate the frequency of j-th of sub-band, am
(fj,θi) expression incident angle be θi, frequency fjWhen array steering vector, Nm(fj) indicate white Gaussian noise nm(t) Fu
In leaf transformation；
(1c) is write the corresponding frequency-domain model of M omnidirectional's array element as matrix form, obtains frequency fjThe broadband signal frequency at place
Domain model: X (fj)=A (θ, fj)S(fj)+N(fj)；
Wherein: S (fj)=[S1(fj),S2(fj),…,SP(fj)]T, N (fj)=[N1(fj),N2(fj),…,NM(fj)]T, A
(θ,fj)=[a (θ1,fj),…,a(θi,fj),…,a(θP,fj)], θ indicates incident angle set,
Step 2, it is based on the broadband signal frequency-domain model, the covariance square at reference frequency is obtained by focus processing
Battle array.
Wherein, step 2 can specifically include following sub-step:
(2a) utilizes Beamforming Method, determines the pre-estimation angle of each incoming signal, obtains pre-estimation angle set β；
The centre frequency of (2b) specific broadband signals incident frequencies as reference frequency, and then according to the reference frequency,
The broadband signal frequency-domain model and the pre-estimation angle, construction obtain focussing matrix T (fj)=V (fj)U(fj)；
Wherein, U (fj) and V (fj) it is matrix A (β, f respectivelyj)AH(β,f0) by SVD decompose after left singular vector and
Right singular vector, f0For reference frequency；
(2c) utilizes focussing matrix, the frequency domain data on each narrow-band component is focused in the reference frequency, in turn
The covariance matrix at the reference frequency is obtained by auto-correlation processing:
In a kind of concrete implementation mode, step (2a) be can specifically include:
The covariance matrix under each frequency point is calculated in (2a1), wherein in j-th of frequency point fjUnder covariance matrix be
Rcbf=X (fj)XH(fj)。
(2a2) takes an angle every 0.1 ° in the airspace of [- 90 °, 90 °], empty in the obtained angled lower progress of institute
Between compose search, determine wherein the corresponding angle of spectral peak, the angle are pre-estimation angle；
Wherein, spatial spectrum searches for formula are as follows: What θ ' expression was got in the airspace of [- 90 °, 90 °] appoints
One angle.
The set that (2a3) whole pre-estimation angle is constituted is to determine pre-estimation angle set β:
Wherein, K0Indicating the number of the pre-estimation angle obtained using conventional beamformer method, BW indicates beam angle,
α indicates the coefficient of adjustment beam angle BW, βkIndicate k-th of the pre-estimation angle estimated, 1≤k≤K0。
Step 3, rarefaction representation is carried out using the covariance matrix at the reference frequency, constructed about sparse vector b0's
Objective function
Wherein, B=[b1,…,bm,…,bM], bmThe m of representing matrix B is arranged, m=1,2 ... M, bmBe Q × 1 coefficient to
Amount, Two norms of representing matrix B q row,W indicates weighting matrix, W-1/2For the On Square-Rooting Matrices of weighting matrix W inverse matrix；G is that block is diagonal
Matrix,Gm=[e1,e2,…em-1,em+1,…,eM]T, GmFor the matrix of (M-1) × M,
emFor the vector of M × 1, emM-th of element be 1, remaining element is 0；The vectorization of vec () representing matrix；Table
Show redundant dictionary；η indicates decision threshold；RYIndicate the covariance matrix at reference frequency.
The theory deduction process that objective function in step 3 is given below is as follows:
For ideal agonic covariance matrix R, it is assumed that setIt covers all possible
Signal incident direction, Q are angular divisions numbers, wherein Q > > M, construct redundant dictionary with thisWhereinAccording to linear theory, each column of R can be by redundancy
Dictionary linear expression:
Wherein, σ2For noise power, bmIt is the coefficient vector of Q × 1, error term emIndicate the vector of M × 1, bmAnd emAll it is
Unknown, ideal bmVector should be in addition to the element non-zero of corresponding sense, and other elements are zero, have and incidence angle
Spend identical sparse spectrum.According to bmSparsity can solve the unique of underdetermined equation (formula 1) using sparse representation theory
Solution, to estimate target source orientation.
As the above analysis, known to traditional DOA estimation algorithm based on covariance matrix rarefaction representation needs
Noise power.In order to avoid estimating noise power, find that noise power only influences r by carrying out analysis to (formula 1)mM
M-th of equation in (formula 1) can be removed by transition matrix, be obtained by a element:
Wherein, Gm=[e1,e2,…em-1,em+1,…,eM]TFor the matrix of M-1 × M, emFor the vector of M × 1, emM
A element is 1, remaining element is 0.Pass through GmConversion, can be to avoid estimating noise power.Enable B=[b1,b2,…,
bm,…,bM], it can be found that each column sparsity structure all having the same, i.e. each b in BmIn nonzero element appear in B
Same a line.Two norms are taken to obtain the vector b of new Q × 1 every a line of matrix B0:
Wherein, m-th of element of B (q, m) representing matrix B q row,Two norms of representing matrix B q row.Thus may be used
The sparsity for obtaining matrix B can be by b0It indicates, as long as finding a vector b0Matrix B will be made to meet constraint well, for
The solution of (formula 2) can be equivalent to find a vector b sparse enough0。
Most direct solution measures b0The method of sparse degree is to calculate b0In sparse number, i.e., solution b0L0Norm, but
It is direct solution l0Norm needs to combine optimizing, mathematically referred to as NP-Hard problem.l1Norm is proved to dilute enough at b °
Dredge when and l0Norm high probability equivalence is equal, while l1Norm is convex function, it is ensured that has optimal solution, therefore l can be used1
Norm solves NP-Hard problem.This, which is based on sparse constraint solution Mutual coupling for (formula 2), to indicate are as follows:
In in practical situations, covariance matrix RYIt is to be obtained by array received data, i.e. RY=R+ △ RY, wherein △ RY
For estimation error of the covarianee matrix, then evaluated error △ RYMeet following distribution:
Wherein, the vectorization of vec () representing matrix, AsN (μ, C) indicate that approximate mean value of obeying is μ, covariance square
Battle array is the multiple normal distribution of C, and N is number of samples,Indicate Kronecker product, G GmThe block pair of (m=1,2 ... M) construction
Angular moment battle array, it may be assumed that
If the R obtained using observationYWhen instead of error-free R, since there are evaluated error △ RYThe then constraint of (formula 4)
No longer set up, available by (formula 5) and knowledge of statistics:
W-1/2△RY~NAs(0,IM×(M-1)) (formula 6)
(formula 6) can indicate are as follows:
In (formula 7), weighting matrix isIndicate evaluated error △ RYAssociation
Variance matrix, W-1/2For the On Square-Rooting Matrices of W inverse matrix, whereinMeet freedom degree (M (M-1))
Chi square distribution.The thresholding η for inhibiting residual error is introduced, is represented byTherefore (formula 2) can be with
It indicates are as follows:
It is the objective function in step 3 shown in (formula 8).Wherein, and the determination of η be with (formula 8) high probability establishment
Premised on, in Matlab software emulation, chi2inv (1- ρ, M can be used2) be calculated, general setting one is sufficiently high
Probability ρ=0.001.
Step 4, the objective function is solved, sparse vector b is obtained0, and then it is corresponding according to the nonzero element of sparse vector
Position determine the incident angle of P broadband incident signal.
Wherein, it should be noted that it will be understood by those skilled in the art that b0It is the column vector of Q × 1, non-zero
The corresponding position of element is the incident direction of signal.
Preferably, in step 4, the objective function is solved, obtains sparse vector b0, it can specifically include:
The form that the objective function Cheng Keyong convex programming packet is solved:And then solved using convex programming packet, it obtains dilute
Dredge vector b0。
Wherein, complete 1 vector of 1 expression Q × 1, γ indicate variable vector, γ=[γ1,γ2,…γQ]T, γqIndicate γ's
Q-th of element, σ2For noise power, g indicates variable element.
The theoretical foundation of above-mentioned conversion is:
Based on the aforementioned analysis to step 3, it is known that,Cause
When this sample number is sufficiently large, evaluated error △ RYIt can be inhibited well.Due to | | b0||1For convex function, quadratic constraints is also
Convex constraint, thus (formula 8) can be converted to above-mentioned form, be solved using convex programming packet (such as Sedumi and CVX).
Based on the broadband signal DOA estimation method provided by the invention based on covariance matrix rarefaction representation, make an uproar in analysis
Between acoustical power and covariance matrix on the basis of relationship, construction block diagonal matrix is to traditional based on covariance matrix
Sparse representation model improves, and eliminates the inaccurate influence to algorithm performance of noise power estimation by dimensionality reduction, therefore be not required to
It wants known or estimating noise power can estimate broadband signal direction of arrival, and between low signal-to-noise ratio, low snap, low angle
The direction of arrival of signal, detection probability with higher and resolving accuracy can be accurately estimated in the case where.
Below by way of the effect of emulation experiment the embodiment of the present invention will be further explained the above method:
Assuming that the wide-band coherent signal of two constant powers is incident on from angle [- 10 °, 20 °] and to be made of 8 omnidirectional's array elements
On even linear array, signal frequency is [100MHz, 300MHz], and array element spacing is the half of signal highest frequency corresponding wavelength, structure
Angle variation range is [- 90 °, 90 °] when building redundant dictionary, is divided into 0.1 °.In each experiment, choosing threshold value is 1 °,
If estimation gained angle and real angle absolute value of the difference is small and thresholding, defines this time experiment and correctly detected to be primary, otherwise
Think detection failure.Define RMSE:WhereinEstimate to tie for the r times experiment
Fruit, θ are incident angle.
Experiment one: the validity of the proposed method of the present invention is investigated in this experiment.Under 1GHz sample frequency, sampling number is
1024, frequency domain snap is 8, and signal-to-noise ratio is emulated when being 10dB, and simulation result is as shown in Figure 2.
Figure it is seen that the mentioned method of the embodiment of the present invention can reach the effect of estimation broadband signal DOA, that is, search
Angle corresponding to the spectral peak that rope goes out is the incident direction of signal, while spectral peak main lobe width is relatively narrow, the absolute value of amplitude compared with
Greatly, there is good resolving accuracy and estimation performance.
Experiment two: the mentioned method DOA of the embodiment of the present invention under different signal-to-noise ratio, snap, angle interval is investigated in this experiment
Estimation performance.When probing into influence of the signal-to-noise ratio to algorithm performance, signal-to-noise ratio variation range is [- 12,0], and change interval is
2dB, carries out 200 Monte Carlo experiments under each signal-to-noise ratio, remaining simulated conditions is consistent with experiment one；Probing into angle
When being spaced the influence to innovatory algorithm performance, one of signal incident direction is -10 °, second signal and first signal
Angle interval is gradually increased out of [4 °, 10 °] range, and change interval is 1 °, and 200 Meng Teka are carried out under each angle interval
Sieve experiment, remaining simulated conditions are consistent with experiment one；When probing into influence of the snap number to innovatory algorithm performance, number of snapshots
Mesh variation range is [100,1000], and change interval 100 carries out 200 Monte Carlo experiments in each number of snapshots now,
Remaining simulated conditions are consistent with experiment one, and simulation result is as shown in Figure 3.
From figure 3, it can be seen that the mentioned method of the embodiment of the present invention is equal at low signal-to-noise ratio, low number of snapshots, low angle interval
Square error is all smaller, and the deviation of real angle is smaller, can achieve the effect of correct estimation DOA, that is, has and preferably estimate
Count performance and estimated accuracy.
Claims (5)
1. a kind of broadband signal DOA estimation method based on covariance matrix rarefaction representation, which is characterized in that receiving array M
The even linear array of a omnidirectional's array element composition, and array element spacing is d, the receiving array is received from the P wide of space far-field incidence
Band incoming signal, incident angle are denoted as θ respectively1,θ2,…,θi,…,θP, the range of incident frequencies is [fl,fh], flIndicate incident
Low-limit frequency, fhIndicate incident highest frequency, θiIndicate the incident angle of i-th of broadband incident signal, 1≤i≤P；
The described method includes:
Step 1, broadband signal frequency-domain model is established；
Step 2, it is based on the broadband signal frequency-domain model, the covariance matrix at reference frequency is obtained by focus processing；
Step 3, rarefaction representation is carried out using the covariance matrix at the reference frequency, constructed about sparse vector b0Target
Function
Wherein, B=[b1,…,bm,…,bM], bmThe m of representing matrix B is arranged, m=1,2 ... M, bmIt is the coefficient vector of Q × 1, Two norms of representing matrix B q row,W indicates weighting matrix, W-1/2For the On Square-Rooting Matrices of weighting matrix W inverse matrix；G is that block is diagonal
Matrix,Gm=[e1,e2,…em-1,em+1,…,eM]T, GmFor the matrix of (M-1) × M,
emFor the vector of M × 1, emM-th of element be 1, remaining element is 0；The vectorization of vec () representing matrix；Table
Show redundant dictionary；η indicates decision threshold；RYIndicate the covariance matrix at reference frequency；
Step 4, the objective function is solved, sparse vector b is obtained0, and then it is corresponding according to the nonzero element of the sparse vector
Position determines the incident angle of the P broadband incident signal.
2. the method according to claim 1, wherein step 1 specifically includes:
(1a) determines the signal form that m-th of omnidirectional's array element receives:
Wherein, si(t) i-th of broadband signal is indicated,The delay that m-th of array element is reached for i-th of signal, for even linear arraynmIt (t) is the white Gaussian noise of m-th of array element；
(1b) carries out Fourier transformation to the signal that m-th of omnidirectional array element receives, and obtains frequency-domain model:
Wherein, J indicates the sub-band number that broadband signal divides on frequency domain, fjIndicate the frequency of j-th of sub-band, am(fj,
θi) expression incident angle be θi, frequency fjWhen array steering vector, Nm(fj) indicate white Gaussian noise nm(t) Fourier
Transformation；
(1c) is write the corresponding frequency-domain model of M omnidirectional's array element as matrix form, obtains frequency fjThe broadband signal frequency domain mould at place
Type: X (fj)=A (θ, fj)S(fj)+N(fj)；
Wherein: S (fj)=[S1(fj),S2(fj),…,SP(fj)]T, N (fj)=[N1(fj),N2(fj),…,NM(fj)]T, A (θ,
fj)=[a (θ1,fj),…,a(θi,fj),…,a(θP,fj)], θ indicates incident angle set,
3. according to the method described in claim 2, it is characterized in that, step 2 specifically includes:
(2a) utilizes Beamforming Method, determines the pre-estimation angle of each incoming signal, obtains pre-estimation angle set β；
The centre frequency of (2b) specific broadband signals incident frequencies is as reference frequency, and then according to the reference frequency, described
Broadband signal frequency-domain model and the pre-estimation angle, construction obtain focussing matrix T (fj)=V (fj)U(fj)；
Wherein, U (fj) and V (fj) it is matrix A (β, f respectivelyj)AH(β,f0) by SVD decompose after left singular vector and right surprise
Different vector, f0For reference frequency；
(2c) utilizes focussing matrix, and the frequency domain data on each narrow-band component is focused in the reference frequency, and then is passed through
Auto-correlation processing obtains the covariance matrix at the reference frequency:
4. according to the method described in claim 3, it is characterized in that, the specific steps of step (2a) include:
The covariance matrix under each frequency point is calculated in (2a1), wherein in j-th of frequency point fjUnder covariance matrix be Rcbf=
X(fj)XH(fj)；
(2a2) takes an angle every 0.1 ° in the airspace of [- 90 °, 90 °], in the obtained angled lower carry out spatial spectrum of institute
Search determines that wherein the corresponding angle of spectral peak, the angle are pre-estimation angle；
Wherein, spatial spectrum searches for formula are as follows:
The set that (2a3) whole pre-estimation angle is constituted is to determine pre-estimation angle set β:
Wherein, K0Indicate the number of the pre-estimation angle obtained using conventional beamformer method, BW indicates that beam angle, α indicate
Adjust the coefficient of beam angle BW, βkIndicate k-th of the pre-estimation angle estimated, 1≤k≤K0。
5. the method according to claim 1, wherein in step 4, the solution objective function is obtained dilute
Dredge vector b0, it specifically includes:
The form that the objective function Cheng Keyong convex programming packet is solved:And then solved using convex programming packet, it obtains dilute
Dredge vector b0；
Wherein, complete 1 vector of 1 expression Q × 1, γ indicate variable vector, γ=[γ1,γ2,…γQ]T, γqIndicate the q of γ
A element, σ2For noise power, g indicates variable element.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910218265.1A CN109901148A (en) | 2019-03-21 | 2019-03-21 | Broadband signal DOA estimation method based on covariance matrix rarefaction representation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910218265.1A CN109901148A (en) | 2019-03-21 | 2019-03-21 | Broadband signal DOA estimation method based on covariance matrix rarefaction representation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109901148A true CN109901148A (en) | 2019-06-18 |
Family
ID=66953552
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910218265.1A Pending CN109901148A (en) | 2019-03-21 | 2019-03-21 | Broadband signal DOA estimation method based on covariance matrix rarefaction representation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109901148A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110579737A (en) * | 2019-07-17 | 2019-12-17 | 电子科技大学 | Sparse array-based MIMO radar broadband DOA calculation method in clutter environment |
| CN111159888A (en) * | 2019-12-28 | 2020-05-15 | 上海师范大学 | Covariance matrix sparse iteration time delay estimation method based on cross-correlation function |
| CN112285639A (en) * | 2020-09-30 | 2021-01-29 | 中国船舶重工集团公司七五0试验场 | Broadband signal orientation estimation method based on cross-shaped sound pressure array |
| CN112363109A (en) * | 2020-11-12 | 2021-02-12 | 西安科技大学 | Weighted direction of arrival estimation method based on broadband strong and weak adjacent signals |
| CN112511209A (en) * | 2021-02-09 | 2021-03-16 | 网络通信与安全紫金山实验室 | Beam control method, system and controller of array antenna based on frequency diversity |
| CN112649786A (en) * | 2020-11-30 | 2021-04-13 | 天津光电通信技术有限公司 | Broadband signal super-resolution direction finding method and device based on virtual array elements |
| CN113406560A (en) * | 2021-05-20 | 2021-09-17 | 华南理工大学 | Angle and frequency parameter estimation method of incoherent distributed broadband source |
| CN113625222A (en) * | 2021-08-27 | 2021-11-09 | 东南大学 | Broadband signal high-resolution direction estimation method based on time delay difference norm weighting |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040113839A1 (en) * | 2002-12-17 | 2004-06-17 | Vaccaro Thomas R. | Wideband signal detection and tracking system |
| CN103954950A (en) * | 2014-04-25 | 2014-07-30 | 西安电子科技大学 | Direction-of-arrival estimation method based on sample covariance matrix sparsity |
| CN106526529A (en) * | 2016-09-19 | 2017-03-22 | 天津大学 | Sparse representation-based direction-of-arrival estimation method in mismatched condition of steering vectors |
| CN108089146A (en) * | 2017-11-20 | 2018-05-29 | 西安电子科技大学 | A kind of high-resolution broadband Wave arrival direction estimating method to estimating angle error robust |
-
2019
- 2019-03-21 CN CN201910218265.1A patent/CN109901148A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040113839A1 (en) * | 2002-12-17 | 2004-06-17 | Vaccaro Thomas R. | Wideband signal detection and tracking system |
| CN103954950A (en) * | 2014-04-25 | 2014-07-30 | 西安电子科技大学 | Direction-of-arrival estimation method based on sample covariance matrix sparsity |
| CN106526529A (en) * | 2016-09-19 | 2017-03-22 | 天津大学 | Sparse representation-based direction-of-arrival estimation method in mismatched condition of steering vectors |
| CN108089146A (en) * | 2017-11-20 | 2018-05-29 | 西安电子科技大学 | A kind of high-resolution broadband Wave arrival direction estimating method to estimating angle error robust |
Non-Patent Citations (3)
| Title |
|---|
| HSIENSEN HUNG ET AL.: "Focussing Matrices for Coherent Signal-Subspace Processing", 《IEEE TRANSACTIONS ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING》 * |
| SANTHOSH KUMAR D. ET AL.: "DOA Estimation of IR-UWB Signals using Coherent Signal Processing", 《2014 IEEE 10TH INTERNATIONAL COLLOQUIUM ON SIGNAL PROCESSING AND ITS APPLICATIONS》 * |
| 施思: "宽带信号高分辨DOA估计算法及降维处理研究", 《中国优秀博硕士学位论文全文数据库(硕士) 信息科技辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110579737A (en) * | 2019-07-17 | 2019-12-17 | 电子科技大学 | Sparse array-based MIMO radar broadband DOA calculation method in clutter environment |
| CN111159888A (en) * | 2019-12-28 | 2020-05-15 | 上海师范大学 | Covariance matrix sparse iteration time delay estimation method based on cross-correlation function |
| CN111159888B (en) * | 2019-12-28 | 2023-06-02 | 上海师范大学 | Covariance matrix sparse iteration time delay estimation method based on cross-correlation function |
| CN112285639A (en) * | 2020-09-30 | 2021-01-29 | 中国船舶重工集团公司七五0试验场 | Broadband signal orientation estimation method based on cross-shaped sound pressure array |
| CN112285639B (en) * | 2020-09-30 | 2023-10-27 | 中国船舶重工集团公司七五0试验场 | Wideband signal azimuth estimation method based on cross sound pressure array |
| CN112363109A (en) * | 2020-11-12 | 2021-02-12 | 西安科技大学 | Weighted direction of arrival estimation method based on broadband strong and weak adjacent signals |
| CN112649786A (en) * | 2020-11-30 | 2021-04-13 | 天津光电通信技术有限公司 | Broadband signal super-resolution direction finding method and device based on virtual array elements |
| CN112511209A (en) * | 2021-02-09 | 2021-03-16 | 网络通信与安全紫金山实验室 | Beam control method, system and controller of array antenna based on frequency diversity |
| CN112511209B (en) * | 2021-02-09 | 2021-06-04 | 网络通信与安全紫金山实验室 | Beam control method, system and controller of array antenna based on frequency diversity |
| CN113406560A (en) * | 2021-05-20 | 2021-09-17 | 华南理工大学 | Angle and frequency parameter estimation method of incoherent distributed broadband source |
| CN113625222A (en) * | 2021-08-27 | 2021-11-09 | 东南大学 | Broadband signal high-resolution direction estimation method based on time delay difference norm weighting |
| CN113625222B (en) * | 2021-08-27 | 2022-10-28 | 东南大学 | Broadband signal high-resolution direction estimation method based on time delay difference norm weighting |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109901148A (en) | Broadband signal DOA estimation method based on covariance matrix rarefaction representation | |
| CN110261841B (en) | MIMO radar single measurement vector DOA estimation method based on iterative weighted near-end projection | |
| CN104698433B (en) | DOA Estimation in Coherent Signal method based on single snapshot data | |
| CN105137399B (en) | The radar self-adaption Beamforming Method filtered based on oblique projection | |
| CN104749553B (en) | Direction of arrival angle method of estimation based on rapid sparse Bayesian learning | |
| CN108375763B (en) | Frequency division positioning method applied to multi-sound-source environment | |
| CN110196410B (en) | Array antenna main lobe interference suppression method and system | |
| CN105137409B (en) | The sane space-time adaptive processing method of echo signal mutually constrained based on width | |
| CN105306123A (en) | Robust beamforming method with resistance to array system errors | |
| CN114047473B (en) | Arrival angle and polarization angle measuring method based on polarization sensitive annular array | |
| CN104777450B (en) | A kind of two-stage MUSIC microphone array direction-finding method | |
| CN109600152A (en) | A kind of Adaptive beamformer method based on the transformation of subspace base | |
| CN109597048B (en) | Meter wave radar DOA estimation method based on two-dimensional convolutional neural network | |
| CN106980104A (en) | Signal direction of arrival automatic correcting method for sensor array | |
| CN111580042B (en) | Deep learning direction finding method based on phase optimization | |
| Semper et al. | Grid-free direction-of-arrival estimation with compressed sensing and arbitrary antenna arrays | |
| CN110196417B (en) | Bistatic MIMO radar angle estimation method based on emission energy concentration | |
| CN104156553B (en) | Coherent signal Wave arrival direction estimating method and system without Sources number estimation | |
| CN113835068B (en) | Blind source separation real-time main lobe interference resistance method based on independent component analysis | |
| CN108828586B (en) | Bistatic MIMO radar angle measurement optimization method based on beam domain | |
| CN104298850A (en) | Coherent signal direction finding method and system with unknown signal source number | |
| CN106685507A (en) | Beam forming method based on Constrained Kalman in colored noise environment | |
| CN113805139A (en) | Broadband signal sparse representation direction-of-arrival estimation method based on focusing transformation | |
| Cai et al. | A MUSIC-type DOA estimation method based on sparse arrays for a mixture of circular and non-circular signals | |
| CN114460531A (en) | Uniform linear array MUSIC spatial spectrum estimation method |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190618 |