CN101482610A - Self-adapting special interference restraint technology for phased array radar - Google Patents

Abstract

The invention discloses a self-adapting special interference suppressing technology aim at phased array radar. The conventional phased array radar can suppress the noise to inhibit the interference by self-adapting, firstly the interference data received by the radar receiver can be learned then to form a zero point on the interference direction. The zero point depth and width generated by the conventional processing method all can be influenced by the array error and are not suitable for the special interferences, such as the dense cheating interference, the movement interference, the fast transforming clearance type interference or the composite interference and so on complicated interference forms. Adopting the ultra-lower secondary lobe aerial is the most effective anti-interference method, but under the present technology and technological level, the over-high secondary lobe requirement for the phased array aerial is impractical. The invention can implement the estimations of the interference source number and the orientation by the space spectrum estimating technology firstly, and then can construct the interfering data covariance matrix by using the analog signals, then can obtain the self-adapting minor lobe cancellation weight vector calculated by the self-adapting method, thereby forming the wide zero point and deep zero defect self-adapting directional diagram to inhibit the complicated interferences. The technology of the invention can be used for the signal processing system of the phased array radar, the implementing is simple, and the invention has wide practical application prospects.

Description

Self-adapting special interference restraint technology for phased array radar

Technical field

The present invention relates to a kind of self-adapting special interference restraint technology for phased array radar in the signal Processing field, be applicable to the signal processing system of phased-array radar, as sky-wave OTH radar signal processing system, Large Phased Array Radar signal processing system, airborne radar early warning radar signal processing system and airborne battle-field search radar signal processing system etc.

Background technology

Phased-array radar is for the search volume target, needs comprehensive scanning, and at present in the antagonistic process of radar, facing maximum is that briquettability is disturbed, and the interference mode that adopts is that narrow band noise is disturbed usually.For multichannel phased array system radar, this interference suppresses than being easier to, and can adopt conventional adaptive approach to suppress usually.

But there are following three shortcomings in conventional adaptive approach: the one, can only tackle oppressive noise, can't suppress complicated disturbance, clearance-type as the cheating interference of intensive, motion artifacts, fast conversion disturbs or composite interference, reason is that conventional adaptive processing method can not learn these complicated disturbance fully, thereby causes conventional adaptive algorithm performance severe exacerbation even inefficacy; The 2nd, adaptive disturbance disappears mutually needs the study interfering data, usually the method that adopts is to learn stand-down at radar, but for the radar of PD system, there is not stand-down in it, so this moment, study just was easy to learn target, thereby caused disappearing mutually of target; The 3rd, self-adaptation disappears mutually needs the drain space degree of freedom, and promptly offseting selective interference needs a space array element, if when having broadband interference or intensive interference, when spatial degrees of freedom when disturbing degree of freedom, adaptive algorithm can lose efficacy.

Adopting ultralow secondary lobe antenna to reduce radar is the method that an anti-strong isolated clutter the most effective or duplicity are disturbed to being positioned at antenna main lobe with the susceptibility of external signal, but the reduction that must recognize antenna side lobe is the raising with manufacturing cost, and the broadening of main lobe is a cost.Under current techniques and technological level, it is unpractical that antenna is proposed too high secondary lobe requirement, and particularly large-scale phased array antenna also can't be accomplished ultralow secondary lobe at present.In many cases, adopting the lower signal processing mode of cost is a kind of more feasible method, is significant so tackle complicated disturbance such as the clearance-type interference of cheating interference, the motion artifacts of this intensive, fast conversion or composite interference with the Adaptive Suppression technology.

Summary of the invention

Purpose of the present invention proposes at the weak point in the above-mentioned background technology just.The present invention realizes the estimation of interference parameter by the Estimation of Spatial Spectrum technology, thus the correlation parameter information that obtains disturbing.The data covariance matrix of utilizing the simulating signal structure to disturb then calculates adaptive weight vector by adaptive approach again, thereby forms adaptive direction figure.The major lobe of directional diagram is a sense, and the interference radiating way that the center zero direction obtains for the Estimation of Spatial Spectrum technology notices that be a wide zero point zero point of structure.The degree of depth at wide zero point and direction can change according to the information self-adapting of estimating, it just can adapt to complicated disturbance such as the clearance-type interference of the cheating interference that suppresses intensive, motion artifacts, conversion soon or composite interference like this.In addition, because data covariance matrix is oneself structure, so can directly deposit its inverse matrix in signal processor, only need multiply each other with its taking-up with according to the steering vector of information structuring when weights calculate get final product, thereby realization is to the inhibition of radar secondary lobe complicated disturbance.The invention has the advantages that to can be used for phased-array radar, and it is little to have an operand, is convenient to realize and characteristics such as popularization.

In order to realize above-mentioned goal of the invention, the invention provides a kind of self-adapting special interference restraint technology for phased array radar, comprise following technical step:

(1) utilizes the intrinsic digital receiver of phased-array radar that all array element data are received, and it is sent into signal processing system;

(2) corresponding data that takes out each array element passage forms the data covariance matrix of phased-array radar, and computing formula is as follows

$R_1=\frac{X_1{X}_1^H}{L_1}$

Wherein, X ₁Be the data matrix that each array element of phased array receives, its dimension is M * L ₁, M is an array number, L ₁Be fast umber of beats, the covariance matrix R that obtains ₁Dimension be M * M;

(3) utilize the estimation of Wave arrival direction estimating method realization, at first the data covariance matrix is carried out feature decomposition the interference source angle

R ₁＝UΛU ^H

Λ=diag[λ wherein ₁, λ ₂..., λ _M] the diagonal angle square formation formed for eigenwert, U=[e ₁, e ₂..., e _M] be the eigenmatrix of forming by proper vector, the eigenwert is here arranged from big to small, i.e. λ ₁λ ₂... λ _Nλ _N+1... λ _M, adopt AIC or MDL method to utilize eigenwert to judge big eigenwert number, suppose that the interference source number is N, then eigenwert satisfies

λ ₁>λ ₂>…>λ _N>>λ _N+1>…>λ _M

Judge after the interference source number, then with the eigenmatrix separated into two parts, i.e. the signal subspace E that forms by big eigenwert characteristic of correspondence vector _S=[e ₁, e ₂..., e _N] and the noise subspace E that eigenvector just formed by little eigenwert _N=[e _N+1, e _N+2..., e _M].Utilize the MUSIC method to realize the angle of interference source is estimated that estimation formulas is as follows

$P\left(\mathrm{\θ}\right)=\frac{1}{a^H\left(\mathrm{\θ}\right)E_N{E}_N^{H}a\left(\mathrm{\θ}\right)}$

Utilize P in the following formula (θ) can realize disturbing the estimation θ of angle _p, p=1,2 ..., N, the angle estimation approach adopts search procedure or polynomial expression rooting.

(4) according to estimated parameter reconstruct extension matrix

Constructive formula is as follows

${\left[T({\mathrm{\θ}}_p,{\mathrm{\σ}}_p^2)\right]}_{k,l}=\exp \{-\frac{1}{2}{\mathrm{\σ}}_p^2{[(k-l)\mathrm{\π}\cos {\mathrm{\θ}}_p/180]}^2\}$

The effect of matrix T is the effect that incident direction is disturbed in expansion, promptly enlarges adaptive direction figure and falls into width at zero of 0:00 direction, and zero width that falls into is by σ _pDecision, can select zero sunken width usually is the half-power point beam angle.

(5) according to the covariance matrix of the extension matrix reconstruct interfering data of estimated parameter and reconstruct, the formula of reconstruct is as follows

Wherein, symbol.Expression Hadamard is long-pending, r _pBe p power that disturbs, σ ²Be noise power, I representation unit matrix, the even linear array steering vector of spacing half-wavelength $a\left({\mathrm{\θ}}_p\right)={[1,e^{-\mathrm{j\π}\sin {\mathrm{\θ}}_p},\·\·\·,e^{-\mathrm{j\π}(M-1)\sin {\mathrm{\θ}}_p}]}^T.$ Noise power elects 1 usually as, and jamming power is selected usually greater than 30dB.

(6) utilize the data covariance matrix R of reconstruct, ask self-adaptation power, formula is as follows

W＝R ^-1a(θ _q)/[a ^H(θ _q)R ^-1a(θ _q)]

Wherein, the steering vector of main lobe sensing $a\left({\mathrm{\θ}}_q\right)={[1,e^{-\mathrm{j\π}\sin {\mathrm{\θ}}_q},\·\·\·,e^{-\mathrm{j\π}(M-1)\sin {\mathrm{\θ}}_q}]}^T.$

(7) all array received data are carried out adaptive weighted processing, it is as follows to handle formula

Y＝W ^HX

Wherein, X is that all array elements of array receive data, and Y is the output data vector of adaptive array, has suppressed the complicated disturbance in space in the output data of this moment.

Wherein, direction of arrival in the step (3) is estimated also to adopt minimum modulus algorithm (MNM), least variance method (MVM), minimum entropy algorithm (MEM), maximum likelihood (ML), weighting subspace fitting (WSF), invariable rotary subspace (ESPRIT) etc., uses the MUSIC method among the embodiment angle is estimated.Reconstruct extension matrix in the step (4), the method that can adopt constant zero to fall into is reconstructed, and this moment is desirable

${\left[\stackrel{\‾}{T}\right]}_{k,l}=\exp \{-\frac{1}{2}{\mathrm{\σ}}_{\max }^2{[(k-l)\mathrm{\π}/180]}^2\}$

In the formula

For

A upper limit, replace T with T, this moment matrix T do not change with interference parameter.As adopt the reconstruct extension matrix of following formula, then the reconstruct data covariance matrix of step (5) can be reduced to

The invention has the advantages that:

(1) since the required covariance matrix of self-adaptation according to the priori signal configuration, so its zero point depth and width can preestablish, thereby make adaptive algorithm can suppress complicated disturbance, disturb or composite interference etc. as the clearance-type of the cheating interference of intensive, motion artifacts, fast conversion, and the structure interference covariance matrix avoided the influence of array error in the learning process, so the inhibition ability of the robustness of algorithm and complicated disturbance is stronger.

(2) the present invention adopts and estimates that earlier interference parameter carries out adaptive method again, the parameter information that so just can make full use of interference comes the data covariance matrix of reconstruct interference, so just can solve the data problem concerning study in the adaptive array, can avoid occurring the phenomenon of signal cancellation, so the performance of algorithm is more stable.

(3) for a definite radar, the covariance matrix of structure and reception data independence, so just can calculated in advance good whole spatial domain different come to the time self-adaptation power, when adopt the direction of arrival algorithm for estimating obtain signal come to the time, just directly from memory, access relevant weight vector, need not to calculate again,, be convenient to Project Realization so it is very low to calculate the operand of adaptive weight at this moment.

(4) the inventive method can be used for transforming the signal processing system of existing Large Phased Array Radar, does not need additionally to increase treatment channel and equipment, only needs that original wave beam formation processing unit is replaced with the direction of arrival estimation unit and gets final product.So, do not need to change the structure of radar receiving system, have application value.

Description of drawings

Fig. 1 is the block diagram of embodiments of the invention.

With reference to Fig. 1, embodiments of the invention are by array digital receiver 1, data pick-up 2, direction of arrival estimates 3, reconstruct extension matrix 4, reconstruct data covariance matrix 5, adaptive weight calculate 6 and adaptive beam form 7 and form.The signal that the array digital receiver receives the space among the embodiment wherein, disturb and message pick-up such as noise and storing in the signal processing system, 2 of data pick-ups extracting part divided data from memory is sent into direction of arrival estimation unit 3, the direction that the direction of arrival estimation unit adopts the MUSIC algorithm to carry out interference source is estimated and the interference source number is estimated, and estimated parameter delivered to reconstruct extension matrix unit 4, unit 4 is according to direction parameter and two parametric configuration extension matrixs of width at zero point, and it is sent to reconstruct data covariance matrix unit 5, direction estimated information in conjunction with interference source forms the data covariance matrix of disturbing again, it is sent into adaptive weight computing unit 6, just can calculate and satisfy the self-adaptation power that the zero sunken degree of depth and zero falls into width, adaptive beam forms unit 7 and utilizes self-adaptation power that unit 6 sends here and original array stored data to carry out adaptive beam to form and just can suppress complicated disturbance, just can export the result of interference inhibition at last.

Embodiment

It is as follows to implement principle of the present invention: at first utilize the array received data to carry out the number and the parameter estimation of interference source, utilize these information to form extension matrix and interfering data covariance matrix then, be formed with certain zero at last and fall into the self-adaptation power that the degree of depth and zero falls into width, utilize this self-adaptation power that the array received data are carried out adaptive beam at last and form, thereby realization is to the inhibition of complicated disturbance.

Suppose that phased-array radar has M array element, N interference is M=64 among the embodiment, and N=2, the angle of interference is respectively θ _J1, θ _J2The detailed step of whole invention once is described below in conjunction with drawings and Examples:

(1) by array numeral receiver unit 1 with the data storage of M array element passage receiving in system, this part is identical with original system to size, the sampling precision requirement of storer.

(2) data of needs are extracted from the storer of system by data pick-up unit 2, the data of extraction are X ₁, its dimension is M * L ₁, M is an array number, L ₁Be fast umber of beats, need to satisfy L generally speaking ₁2M, thereby the covariance matrix R of extracted data obtained ₁

$R_1=\frac{X_1{X}_1^H}{L_1}$

(3) M that sends here of 3 pairs of data extracting units 2 of direction of arrival estimation unit * M dimension covariance matrix R ₁Carry out feature decomposition

R ₁＝UΛU ^H

Λ=diag[λ wherein ₁, λ ₂..., λ _M] the diagonal angle square formation formed for eigenwert, U=[e ₁, e ₂..., e _M] be the eigenmatrix of forming by proper vector, the eigenwert is here arranged from big to small, i.e. λ ₁λ ₂... λ _Nλ _N+1... λ _M, adopt AIC or MDL method to utilize eigenwert to judge big eigenwert number, suppose that the interference source number is N=2, then eigenwert satisfies

λ ₁>λ ₂>>λ ₃>…>λ _M

Judge after the interference source number, then with the eigenmatrix separated into two parts, i.e. the signal subspace E that forms by big eigenwert characteristic of correspondence vector _S=[e ₁, e ₂] and the noise subspace E that eigenvector just formed by little eigenwert _N=[e ₃, e ₄..., e _M].Utilize the MUSIC method to realize the angle of interference source is estimated that estimation formulas is as follows

$P\left(\mathrm{\θ}\right)=\frac{1}{a^H\left(\mathrm{\θ}\right)E_N{E}_N^{H}a\left(\mathrm{\θ}\right)}$

Utilize P in the following formula (θ) can realize disturbing the estimation of angle, the angle estimation approach adopts search procedure or polynomial expression rooting, is 2 interference among the embodiment, supposes that the angle that estimates is θ _J1And θ _J2, then need these two angles are sent into reconstruct extension matrix unit 4.

(4) interference source of sending here according to direction of arrival estimation unit 3 is counted N and angle information, and reconstruct extension matrix unit 4 a reconstruct N extension matrix suppose to need to form the zero wide σ of being respectively that disturbs ₁, σ ₂..., σ _N, then the formula of reconstruct extension matrix is as follows

${\left[T({\mathrm{\θ}}_p,{\mathrm{\σ}}_p^2)\right]}_{k,l}=\exp \{-\frac{1}{2}{\mathrm{\σ}}_p^2{[(k-l)\mathrm{\π}\cos {\mathrm{\θ}}_p/180]}^2\}$

2 interference sources are arranged among the embodiment, and then two of reconstruct extension matrixs are

${\left[T_1({\mathrm{\θ}}_1,{\mathrm{\σ}}_1^2)\right]}_{k,l}=\exp \{-\frac{1}{2}{\mathrm{\σ}}_1^2{[(k-l)\mathrm{\π}\cos {\mathrm{\θ}}_1/180]}^2\}$

${\left[T_2({\mathrm{\θ}}_2,{\mathrm{\σ}}_2^2)\right]}_{k,l}=\exp \{-\frac{1}{2}{\mathrm{\σ}}_2^2{[(k-l)\mathrm{\π}\cos {\mathrm{\θ}}_2/180]}^2\}$

The extension matrix of reconstruct is delivered in the reconstruct data covariance matrix unit 5 then.

(5) extension matrix sent here according to the angle information of estimating and unit 4 of reconstruct data covariance matrix unit 5 is reconstructed, and formula is as follows

Symbol wherein.Expression Hadamard is long-pending, r _pBe p power that disturbs, σ ²Be noise power, I representation unit matrix, the even linear array steering vector of spacing half-wavelength $a\left({\mathrm{\θ}}_p\right)={[1,e^{-\mathrm{j\π}\sin {\mathrm{\θ}}_p},\·\·\·,e^{-\mathrm{j\π}(M-1)\sin {\mathrm{\θ}}_p}]}^T.$ Noise power elects 1 usually as, and jamming power is selected usually greater than 30dB.

The matrix of reconstruct is among the embodiment

Unit 5 was delivered to unit 6 with data covariance matrix and is carried out the adaptive weight computing after reconstruct was finished.

(6) adaptive weight computing unit 6 utilizes the interfering data covariance matrix that unit 5 is sent here, asks self-adaptation power, and formula is as follows

W＝R ^-1a(θ _q)/[a ^H(θ _q)R ^-1a(θ _q)]

Wherein, the steering vector of main lobe sensing $a\left({\mathrm{\θ}}_q\right)={[1,e^{-\mathrm{j\π}\sin {\mathrm{\θ}}_q},\·\·\·,e^{-\mathrm{j\π}(M-1)\sin {\mathrm{\θ}}_q}]}^T.$ The weight vector that calculates among the embodiment is 64 * 1 n dimensional vector ns.Calculated after the weights, unit 7 is sent weight vector in unit 6.

(7) adaptive beam forms unit 7 extracts array element from system storage all data X, and the weight vector that data and unit 6 are sent here is weighted processing then, and formula is as follows

Y＝W ^HX

Wherein, X is that all array elements of array receive data, and Y is the output data vector of adaptive array, has suppressed the complicated disturbance in space in the output data of this moment.X is 64 * L dimension matrix among the embodiment, and Y is 1 * L n dimensional vector n, and L is the range gate number that radar receives.

In addition, direction of arrival in the step (3) is estimated also to adopt minimum modulus algorithm (MNM), least variance method (MVM), minimum entropy algorithm (MEM), maximum likelihood (ML), weighting subspace fitting (WSF), invariable rotary subspace (ESPRIT) etc., uses the MUSIC method among the embodiment angle is estimated.

Reconstruct extension matrix in the step (4), the method that can adopt constant zero to fall into is reconstructed, at this moment in the retrievable

For

A upper limit, among the embodiment ${\mathrm{\σ}}_{\max }^2=\max \{{\mathrm{\σ}}_1^2,{\mathrm{\σ}}_2^2\},$ Then the matrix of reconstruct has only 1

${\left[\stackrel{\‾}{T}\right]}_{k,l}=\exp \{-\frac{1}{2}{\mathrm{\σ}}_{\max }^2{[(k-l)\mathrm{\π}/180]}^2\}$

Note utilizing the extension matrix of following formula structure and angle irrelevant, only relevant with the zero width that falls into.

Then the reconstruct data covariance matrix of step (5) can be reduced to

The matrix of reconstruct is among the embodiment

Though described embodiments of the present invention in conjunction with the accompanying drawings, those of ordinary skills can make various distortion or modification within the scope of the appended claims.

Claims (4)

1. self-adapting special interference restraint technology for phased array radar comprises following technical step:

(1) utilizes the intrinsic digital receiver of phased-array radar that all array element data are received, and it is sent into signal processing system;

(2) corresponding data that takes out each array element passage forms the data covariance matrix of phased-array radar, and computing formula is as follows

$R_1=\frac{X_1{X}_1^H}{L_1}$

Wherein, X ₁Be the data matrix that each array element of phased array receives, its dimension is M * L ₁, M is an array number, L ₁Be fast umber of beats, the covariance matrix R that obtains ₁Dimension be M * M;

(3) utilize the estimation of Wave arrival direction estimating method realization, at first the data covariance matrix is carried out feature decomposition the interference source angle

R ₁＝UΛH

Λ=diag[λ wherein ₁, λ ₂..., λ _M] the diagonal angle square formation formed for eigenwert, U=[e ₁, e ₂..., e _M] be the eigenmatrix of forming by proper vector, the eigenwert is here arranged from big to small, i.e. λ ₁λ ₂... λ _Nλ _N+1... λ _M, adopt AIC or MDL method to utilize eigenwert to judge big eigenwert number, suppose that the interference source number is N, then eigenwert satisfies

λ ₁>λ ₂>…>λ _N>>λ _N+1>…>λ _M

Judge after the interference source number, then with the eigenmatrix separated into two parts, i.e. the signal subspace E that forms by big eigenwert characteristic of correspondence vector _s=[e ₁, e ₂..., e _N] and the noise subspace E that eigenvector just formed by little eigenwert _N=[e _N+1, e _N+2..., e _M].Utilize the MUSIC method to realize the angle of interference source is estimated that estimation formulas is as follows

$P\left(\mathrm{\θ}\right)=\frac{1}{a^H\left(\mathrm{\θ}\right)E_N{E}_N^{H}a\left(\mathrm{\θ}\right)}$

Utilize P in the following formula (θ) can realize disturbing the estimation θ of angle _p, p=1,2 ..., N, the angle estimation approach adopts search procedure or polynomial expression rooting.

(4) according to estimated parameter reconstruct extension matrix

Constructive formula is as follows

${\left[T({\mathrm{\θ}}_p,{\mathrm{\σ}}_p^2)\right]}_{k,l}=\exp \{-\frac{1}{2}{\mathrm{\σ}}_p^2{[(k-l)\mathrm{\π}\cos {\mathrm{\θ}}_p/180]}^2\}$

The effect of matrix T is the effect that incident direction is disturbed in expansion, promptly enlarges adaptive direction figure and falls into width at zero of 0:00 direction, and zero width that falls into is by σ _pDecision, can select zero sunken width usually is the half-power point beam angle.

(5) according to the covariance matrix of the extension matrix reconstruct interfering data of estimated parameter and reconstruct, the formula of reconstruct is as follows

Wherein, symbol.Expression Hadamard is long-pending, r _pBe p power that disturbs, σ ²Be noise power, I representation unit matrix, the even linear array steering vector of spacing half-wavelength $a\left({\mathrm{\θ}}_p\right)={[1,e^{-\mathrm{j\π}\sin {\mathrm{\θ}}_p},\·\·\·,e^{-\mathrm{j\π}(M-1)\sin {\mathrm{\θ}}_p}]}^T.$ Noise power elects 1 usually as, and jamming power is selected usually greater than 30dB.

(6) utilize the data covariance matrix R of reconstruct, ask self-adaptation power, formula is as follows

W＝R ^-1a(θ _q)/[a ^H(θ _q)R ^-1a(θ _q)]

Wherein, the steering vector of main lobe sensing $a\left({\mathrm{\θ}}_q\right)={[1,e^{-\mathrm{j\π}\sin {\mathrm{\θ}}_q},\·\·\·,e^{-\mathrm{j\π}(M-1)\sin {\mathrm{\θ}}_q}]}^T.$

(7) all array received data are carried out adaptive weighted processing, it is as follows to handle formula

Y＝W ^HX

Wherein, X is that all array elements of array receive data, and Y is the output data vector of adaptive array, has suppressed the complicated disturbance in space in the output data of this moment.

2. according to the described self-adapting special interference restraint technology for phased array radar of claim 1, it is characterized in that direction of arrival estimates also may adopt minimum modulus algorithm (MNM), least variance method (MVM), minimum entropy algorithm (MEM), maximum likelihood (ML), weighting subspace fitting (WSF), invariable rotary subspace (ESPRIT) in the step (3).

3. according to the described self-adapting special interference restraint technology for phased array radar of claim 1, it is characterized in that, reconstruct extension matrix reconstruct in the step (4), the method that can adopt constant zero to fall into is reconstructed, and this moment is desirable

${\left[\stackrel{\‾}{T}\right]}_{k,l}=\exp \{-\frac{1}{2}{\mathrm{\σ}}_{\max }^2{[(k-l)\mathrm{\π}/180]}^2\}$

In the formula

For

A upper limit, replace T with T, this moment matrix T do not change with interference parameter.

4. according to the described self-adapting special interference restraint technology for phased array radar of claim 1, it is characterized in that, reconstruct data covariance matrix in the step (5), this moment is desirable