CN104076337A - Airborne radar clutter suppression method based on array element amplitude and phase error correction - Google Patents

Abstract

The invention belongs to the technical field of radar cluster suppression, and particularly relates to an airborne radar clutter suppression method based on array element amplitude and phase error correction. The airborne radar clutter suppression method based on array element amplitude and phase error correction includes the following steps of firstly, transmitting a pulse signal through a non-side-looking array antenna of airborne radar, receiving corresponding echo data, and estimating the amplitude and phase error vector of each sub-array according to four-dimensional echo data, wherein the corresponding echo data received by the airborne radar are the four-dimensional echo data; secondly, obtaining data obtained after clutter suppression is conducted on the first distance unit of the mth sub-array according to the amplitude and phase error vector of the mth sub-array, forming the data obtained after clutter suppression is conducted on the mth sub-array through the data obtained after clutter suppression is conducted on the first distance unit...the Lth distance unit of the mth sub-array, and forming data obtained after clutter suppression is conducted on the airborne radar through data obtained after clutter suppression is conducted on the first sub-array...the Mth sub-array.

Description

Airborne radar clutter suppression method based on array element amplitude and phase error correction

Technical field

The invention belongs to radar clutter and suppress technical field, the particularly airborne radar clutter suppression method based on array element amplitude and phase error correction, is exactly a kind ofly to utilize pitching to the method for degree of freedom filtering short range clutter specifically.This method is mainly used in solving the non-stationary problem of airborne radar non-working side battle array short range clutter in array element amplitude phase error situation that exists, the clutter of filtering short range effectively, thereby the distance stationarity that strengthens clutter, is improved follow-up space-time two-dimensional self-adaptive processing performance greatly.

Background technology

While looking work under airborne radar, for positive side-looking array antenna, clutter Doppler frequency is expanded with the variation of distance hardly, and now clutter meets stationarity in distance.Therefore in space-time adaptive processing procedure, can from neighbor distance unit, obtain enough training samples for estimate covariance matrix, this clutter suppression method has obtained application widely at radar signal processing field.Yet in order to realize radar, the omnibearing scanning of the realization of goal from all directions is covered, conventionally need a plurality of antenna associated working, when the antenna axis direction of placing and carrier aircraft flying speed angular separation are when non-vanishing, we claim that this antenna is non-working side array antenna.For non-working side array antenna, the range Doppler spectral line of clutter presents case of bending, and the Doppler frequency of clutter is with apart from acute variation.This Range-dependent depends mainly on fuzzy distance the 0th time, or perhaps short range clutter.

Because non-working side battle array short range clutter presents very strong non-stationaryly, and non-stationary degree strengthens along with the increase of antenna axis and carrier aircraft velocity reversal angle.Yet in the phased array antenna decorum, ultralow secondary lobe antenna technology is difficult to realize under existing Radar Technology condition, so the performance number of short range clutter is conventionally very strong.For short range clutter region, meet independent identically distributed range unit number in this case just very limited and no longer meet Brennan criterion about the requirement of number of training object, cause STAP (Space ?the Time Adaptive Processing space-time adaptive process) handling property of clutter to decline, so the application of STAP technology is restricted.The distance non-stationary property causing in order to alleviate non-positive side battle array, first Doppler effect correction has caused radar worker's attention, the basic thought of the method is that the clutter of different distance is carried out to different Doppler effect corrections, and the main clutter centre frequency of different distance is reached unanimity.Specific implementation only need to be multiplied by corresponding linear phase to the burst length data of each distance, process simple in structure, effective, but it compensates main clutter center, to the compensation of sidelobe clutter, not very good, and be only applicable to not have the situation of range ambiguity.Angle Doppler effect correction afterwards, registration compensation, integration update method etc. all have certain abirritation to short range clutter non-stationary to a certain extent, but when the pulse repetition rate of radar is higher, cause in the situation of range ambiguity, their application is just restricted.

To existing the airborne non-working side battle array radar short range clutter in range ambiguity situation to suppress, can utilize the pitching degree of freedom of array antenna and by itself and orientation, time domain combined self-adaptive processing while forming three-dimensional space.This method is clutter reduction effectively, but inevitably causes the excessive problem of calculated amount, and simultaneously because degree of freedom in system is very high, so the independent same distribution number of samples needing also will increase, this is unappeasable often in practice.Even traditional space-time two-dimensional self-adaptive processing need to be carried out dimension-reduction treatment, can how three-dimensional adaptive processing method is carried out to dimension-reduction treatment be its key that be applied.For a kind of method of first utilizing pitching degree of freedom to carry out pitching pre-filtering of the problems referred to above, be suggested, be referred to as pitching pre-filtering method.This pitching pre-filtering method can be weighted the three-dimensional clutter data that receive that row submatrix is synthetic is converted to 2-D data under the prerequisite of not doing self-adaptive processing, and in transfer process, short range clutter is curbed.Specific implementation process is: the angle of pitch place corresponding at short range clutter adopts fixed constraint zero setting, keeps the gain of main beam direction clutter to remain unchanged simultaneously.The method proposes in desirable clutter situation, is only applicable to not exist in the situation of array element amplitude phase error, that is to say that the inhibition method of its short range clutter when there is radar system error is unsane.

Proposing rational algorithm for error correction becomes the key that solves short range Clutter suppression algorithm robustness, and traditional array error bearing calibration comprises active bearing calibration and automatic correcting method.Active correction need to assist the direction of arrival of information source accurately known, once there is deviation in the direction of auxiliary information source, will bring larger error, automatic correcting method can not need the accurately direction of known auxiliary information source, but estimate by completing online actual direction of arrival, correction accuracy is higher, but its calculated amount is often very large, is difficult to Project Realization.The present invention is because radar parameter priori is known, and direction of arrival can be accurately known by calculating, and therefore adopts the method for active correction to carry out error correction.

Summary of the invention

The object of the invention is to for non-working side battle array short range clutter non-stationaryly and make the problem of traditional space-time two-dimensional self-adaptive processing hydraulic performance decline, propose the airborne radar clutter suppression method based on array element amplitude and phase error correction.First the present invention proofreaies and correct array element amplitude phase error, then the clutter data after proofreading and correct is carried out to pitching filtering processing, thereby the rejection that makes STAP process clutter has obtained improving greatly.

For realizing above-mentioned technical purpose, the present invention adopts following technical scheme to be achieved.

Airborne radar clutter suppression method based on array element amplitude and phase error correction comprises the following steps: step 1, on described airborne radar, be provided with non-working side array antenna, described non-working side array antenna is comprised of M row submatrix, each row submatrix is the even linear array being comprised of N array element, M and N are natural number, the array element distance of each row submatrix is d, and the spacing between any two adjacent row submatrixs is d;

Utilize the non-working side array antenna transponder pulse signal of airborne radar, and receive corresponding echo data, it is four-dimensional echo data that airborne radar receives corresponding echo data; According to described four-dimensional echo data, estimate the amplitude phase error vector of each row submatrix;

Step 2, according to the amplitude phase error vector of m row submatrix, show that l range unit clutter of m row submatrix suppresses rear data m gets 1 to M, and l gets 1 to L, the umber of pulse of the non-working side array antenna transmitting that P is airborne radar, and L represents the range unit number of airborne radar; After utilizing the 1st range unit clutter inhibition of m row submatrix, L rear data of range unit clutter inhibition of data to the m row submatrix form m the rear data of row submatrix clutter inhibition data after utilizing the 1st row submatrix clutter to suppress data after suppressing to M row submatrix clutter form airborne radar clutter and suppress rear data

Feature of the present invention and further improvement are:

The concrete sub-step of described step 1 is:

The range ambiguity number of times of (1.1) l the corresponding clutters of range unit is N _rlinferior, the i1 time range ambiguity clutter with respect to the angle of pitch of non-working side array antenna is ? pitching steering vector be expressed as:

Wherein, the transposition of T representing matrix or vector, λ is the wavelength that airborne radar transmits;

Utilize extremely the pitching guiding matrix B that forms m row submatrix _m:

(1.2) the pitching guiding matrix B to m row submatrix _mcarry out feature decomposition, in the pitching guiding matrix B of m row submatrix _mall eigenwerts in, choose K maximum eigenwert, K<N _rl; The K that utilization is chosen a maximum eigenwert characteristic of correspondence vector forms the array manifold matrix A of m row submatrix _m,

A _m＝[a _m1,a _m2,···,a _mK]；

Wherein, a _m1to a _mKrepresent respectively K the maximum eigenwert characteristic of correspondence vector of choosing;

(1.3) airborne radar receives corresponding echo data and is expressed as four-dimensional echo data X _{n * M * P * L}, at described four-dimensional echo data X _{n * M * P * L}in, extract the 2-D data X corresponding with l range unit of m row submatrix _{n * P}(m, l), then draws 2-D data X _{n * P}the pitching covariance matrix R of (m, l) _ml:

R _ml＝(X _N×P(m,l)X ^H _N×P(m,l))/P

Wherein, the conjugate transpose of H representing matrix;

(1.4) to R _mlcarry out feature decomposition, at R _mlall eigenwerts in choose N-K minimum eigenwert, utilize N-K the minimum eigenwert characteristic of correspondence vector formation noise subspace U choosing _n;

Structure cost function J _m, cost function J _mexpression formula be:

$J_m=\underset{k^{\&prime;}=1}{\overset{K}{\mathrm{\&Sigma;}}}{\mathrm{\&delta;}}_m^H{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}}^H{U}_N{U}_N^H{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}}{\mathrm{\&delta;}}_m$

Wherein, the conjugate transpose of H representing matrix, k' is that natural number and k' get 1 to K, α _mk'=diag (a _mk'), a _mk'represent k' the eigenwert characteristic of correspondence vector of choosing in sub-step (1.3), diag (a _mk') represent with a _mk'in the diagonal matrix that forms as the elements in a main diagonal of each element;

Order ${\mathrm{\&Omega;}}_m=\underset{k^{\&prime;}=1}{\overset{K}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}}^H{U}_N{U}_N^H{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}},$ Cost function J _mbe expressed as:

$J_m={\mathrm{\&delta;}}_m^H{\mathrm{\&Omega;}}_m{\mathrm{\&delta;}}_m$

Then set up following quadratic constraints optimization problem:

$\left\{\begin{array}{c}{\mathrm{\&delta;}}_m=\underset{{\mathrm{\&delta;}}_m}{\min }{\mathrm{\&delta;}}_m^H{\mathrm{\&Omega;}}_m{\mathrm{\&delta;}}_m\\ s.t.{\mathrm{\&delta;}}_m^{H}f=1\end{array}\right.$

Wherein, δ _mamplitude phase error vector for m row submatrix to be solved; F is the column vector of N * 1 dimension, and first element of f is 1, and all the other elements are 0;

By solving described quadratic constraints optimization problem, estimate the amplitude phase error vector delta of m row submatrix _m.In sub-step (1.4), use method of Lagrange multipliers to solve described quadratic constraints optimization problem, described in the amplitude phase error vector delta of m row submatrix solving _mfor:

${\mathrm{\&delta;}}_m=\frac{{\mathrm{\&Omega;}}_m^{-1}f}{f^T{\mathrm{\&Omega;}}_m^{-1}f}$

Wherein, the transposition of T representing matrix or vector, Shang Biao ?1 representing matrix contrary.

In step 2, show that l range unit clutter of m row submatrix suppresses rear data process comprise following sub-step:

First draw the short range clutter steering vector a that l range unit of m row submatrix is corresponding _1mlthe journey clutter steering vector a corresponding with l range unit of m row submatrix _2ml;

Wherein, the transposition of T representing matrix or vector, λ is the wavelength that airborne radar transmits, represent the angle of pitch corresponding to range ambiguity the 0th time, represent main beam pointing when airborne radar transmits;

Then according to following formula, draw the short range clutter steering vector with error compensation that l range unit of m row submatrix is corresponding the long-range clutter steering vector with error compensation corresponding with l range unit of m row submatrix

Wherein, represent that Hadamard is long-pending; δ _mit is the amplitude phase error vector of m row submatrix;

According to following formula construction, go out pitching filtering weight vector corresponding to l range unit of m row submatrix

$\left\{\begin{array}{c}{\tilde{w}}_{\mathrm{ml}}^H{\tilde{a}}_{1\mathrm{ml}}=0\\ {\tilde{w}}_{\mathrm{ml}}^H{\tilde{a}}_{2\mathrm{ml}}=1\end{array}\right.$

Wherein, the conjugate transpose of H representing matrix;

At described four-dimensional echo data X _{n * M * P * L}in, extract the 2-D data X corresponding with l range unit of m row submatrix _{n * P}(m, l), according to pitching filtering weight vector corresponding to l range unit of m row submatrix show that l range unit clutter of m row submatrix suppresses rear data ${\tilde{X}}_{N\&times;P}(m,l)$ For: ${\tilde{X}}_{N\&times;P}(m,l)={\tilde{w}}_{\mathrm{ml}}^H{X}_{N\&times;P}(m,l).$

Beneficial effect of the present invention is:

1) the present invention is directed to the non-stationary problem of airborne radar non-working side battle array short range clutter distance, make full use of pitching degree of freedom and carry out pitching filtering processing, for each range unit, with the pitching weights of fixing, with change of distance short range clutter place formation faster recess, keeping the gain of long-range clutter constant simultaneously, therefore when short range clutter obtains fine inhibition, main-beam clutter but has very large residue;

2) for the situation that has array element amplitude phase error, the recess that pitching filter weights forms departs from short range clutter position, and after pitching filtering, short range clutter has residue greatly.The present invention utilizes and estimates that the row submatrix amplitude phase error vector obtaining compensates traditional pitching filtering weight vector, then utilizes this weight vector in pitching dimension, to carry out matched filtering to the corresponding clutter data of this row submatrix.Experiment shows to adopt the array element amplitude and phase error correction algorithm that the present invention proposes to proofread and correct, and then carry out pitching filtering and process short range clutter and substantially obtained effectively suppressing, and it takes full advantage of radar parameter prior imformation, and calculated amount is less, is convenient to engineering application.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of the airborne radar clutter suppression method based on array element amplitude and phase error correction of the present invention;

Fig. 2 a is the range error of the 5th each array element of row submatrix estimated according to the present invention in emulation experiment and the comparison diagram of the actual margin error of the 5th each array element of row submatrix;

Fig. 2 b is the phase error of the 5th each array element of row submatrix estimated according to the present invention in emulation experiment and the comparison diagram of the substantial phase error of the 5th each array element of row submatrix;

Fig. 2 c is the amplitude square error root curve synoptic diagram of the 5th each array element of row submatrix estimating according to the present invention in emulation experiment;

The phase place square error root curve synoptic diagram of the 5th each array element of row submatrix estimating according to the present invention in Fig. 2 d emulation experiment;

Fig. 3 a is that in emulation experiment, the 5th row submatrix carried out pitch orientation figure before array amplitude and phase error correction, carried out the pitch orientation figure after array amplitude and phase error correction, static pitch orientation figure and short range clutter pitch orientation figure;

Fig. 3 b is the local enlarged diagram of Fig. 3 a;

Fig. 4 is the range Doppler figure that has the original echo data of array element amplitude phase error in emulation experiment;

Fig. 5 a does not carry out in emulation experiment in array amplitude and phase error correction situation through the filtered echo data range Doppler of pitching figure;

Fig. 5 b does not carry out the echo data range Doppler figure after pitching filtering and EFA processing in array amplitude and phase error correction situation in emulation experiment;

Fig. 6 a carries out in emulation experiment in array amplitude and phase error correction situation through the filtered echo data range Doppler of pitching figure;

Fig. 6 b carries out the echo data range Doppler figure after pitching filtering and EFA processing in array amplitude and phase error correction situation in emulation experiment.

Embodiment

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

With reference to Fig. 1, it is the process flow diagram of the airborne radar clutter suppression method based on array element amplitude and phase error correction of the present invention.On airborne radar of the present invention, be provided with non-working side array antenna, non-working side array antenna is the rectangular surfaces array antenna of N * M dimension, N and M are natural number, N represents the array element line number (orientation is to element number of array) of non-working side array antenna, and M represents the array element columns (pitching is to element number of array) of non-working side array antenna.That is to say, non-working side array antenna is comprised of M row submatrix, each row submatrix is the even linear array being comprised of N array element, the array element distance of each row submatrix is d, spacing between any two adjacent row submatrixs is d, and non-working side array antenna orientation is d to array element distance and the pitching of non-working side array antenna to array element distance.

First clutter Doppler frequency and clutter are analyzed to the relation of the oblique distance of carrier aircraft, ignored the impact of earth curvature on result of calculation, clutter Doppler frequency f _dfor:

$f_d=\frac{2V}{\mathrm{\&lambda;}}(\cos \mathrm{\&psi;}\cos \mathrm{\&alpha;}-\sqrt{1-{\left(\frac{H^{\&prime;}}{R_l}\right)}^2-{\cos }^2\mathrm{\&psi;}}\sin \mathrm{\&alpha;})$

Wherein, the flying speed that V is carrier aircraft, λ is the wavelength that airborne radar transmits, H' and R _lrepresent that respectively carrier aircraft height and clutter are to the oblique distance of carrier aircraft, α is crab angle (antenna axis and heading), because non-working side array antenna is arranged on airborne radar, so crab angle α is definite value; ψ represents cone angle (angle of main beam pointing and antenna axial direction) on the same ripple position corresponding with clutter, and ψ is changeless value.Therefore, clutter Doppler frequency f _dr _lfunction.Work as H'/R _lwhen larger, clutter is less to the oblique distance of antenna array, the angle of pitch (clutter is to antenna) is larger, now f _dwith oblique distance R _lchange more violent; Otherwise, f _dwith oblique distance R _lchange milder.In non-working side battle array situation, the performance of this specific character of clutter in forward sight battle array be the most obvious, and in the embodiment of the present invention, non-working side battle array is forward sight battle array.

Draw snap vector C when ideally clutter is empty, when ideally clutter is empty, snap vector C is:

Wherein, N _rfor range ambiguity number of times, N _cbe the independent noise source number of a rang ring, a _ipbe the clutter complex magnitude of the i time range ambiguity p independent noise source, i gets 1 to N _r, p gets 1 to N _c, be the clutter spatial domain pitching steering vector of the i time range ambiguity, orientation, the spatial domain steering vector of the i time range ambiguity p independent noise source, the time domain steering vector of the i time range ambiguity p independent noise source, the flying speed that V is carrier aircraft.

with expression formula be respectively:

Wherein, be that the i time each independent noise source of range ambiguity is with respect to the angle of pitch of non-working side array antenna, θ _ipthat the i time range ambiguity p independent noise source is with respect to the position angle of non-working side array antenna, α is axially (mode vertical with the front of non-working side array antenna) and carrier aircraft heading angle of non-working side array antenna, d is the array element distance of each row submatrix, and d is the spacing between any two adjacent row submatrixs.λ is the wavelength that airborne radar transmits, f _rfor pulse repetition rate, N represents the array element line number of non-working side array antenna, and M represents the array element columns of non-working side array antenna, the umber of pulse of the non-working side array antenna transmitting that P is airborne radar.

Set up the array element amplitude phase error matrix of non-working side array antenna, the expression formula of snap vector C when empty according to the array element amplitude phase error matrix of non-working side array antenna and clutter, draws the echo data model of non-working side array antenna.

Specifically, the array element amplitude phase error matrix Γ of model non-working side array antenna, the array element amplitude phase error matrix Γ of non-working side array antenna is expressed as:

Wherein, ρ _nmthe range error that represents the array element of the capable m row of non-working side array antenna n, φ _nmthe phase error that represents the array element of the capable m row of non-working side array antenna n, n gets 1 to N, and m gets 1 to M.Order in above-mentioned array element amplitude phase error model, conventionally using first array element as with reference to array element, δ _mthe amplitude phase error vector that represents m row submatrix, δ _mexpansion form be:

${\mathrm{\&delta;}}_m={[{\mathrm{\&rho;}}_{1m}{e}^{{\mathrm{j\&phi;}}_{1m}},{\mathrm{\&rho;}}_{2m}{e}^{{\mathrm{j\&phi;}}_{2m}},...,{\mathrm{\&rho;}}_{\mathrm{Nm}}{e}^{{\mathrm{j\&phi;}}_{\mathrm{Nm}}}]}^T.$

Snap vector while drawing clutter sky when array element amplitude phase error exists snap vector when clutter when array element amplitude phase error exists is empty for:

Wherein, N _rfor range ambiguity number of times, N _cbe the independent noise source number of a rang ring, a _ipbe the clutter complex magnitude of the i time range ambiguity p independent noise source, i gets 1 to N _r, p gets 1 to N _c; a=[11 ... 1] ^t, A is the column vector of P * 1 dimension, the element in A is the transposition of 1, T representing matrix or vector entirely. represent Kronecker long-pending (apposition), represent Hadamard long-pending (inner product).

The echo data model of non-working side array antenna is:

$\tilde{x}=\tilde{c}+n$

Wherein, n is the white Gaussian noise vector of NMP dimension, and it obeys average is zero, and variance is σ ²i _nMPgaussian distribution, I _nMPthe unit matrix that represents NMP dimension.

Airborne radar clutter suppression method based on array element amplitude and phase error correction of the present invention comprises the following steps:

Step 1, utilizes the non-working side array antenna transponder pulse signal of airborne radar, and receives corresponding echo data, and it is four-dimensional echo data that airborne radar receives corresponding echo data; According to described four-dimensional echo data, estimate the amplitude phase error vector of each row submatrix.

Its concrete sub-step is:

Because M different lines submatrix of non-working side array antenna is separate, so the estimation problem of the array element amplitude phase error matrix Γ of non-working side array antenna can be decomposed into the estimation problem of the amplitude phase error vector of M separate row submatrix.

Estimate the amplitude phase error vector delta of the 1st row submatrix ₁amplitude phase error vector delta to M row submatrix _m.Estimate the amplitude phase error vector delta of m row submatrix _mprocess comprise following sub-step:

The range ambiguity number of times of (1.1) l the corresponding clutters of range unit is N _rlinferior, the i1 time range ambiguity clutter (each independent noise source) with respect to the angle of pitch of non-working side array antenna is i1 gets 1 to N _rl, H' represents carrier aircraft height, R' _i1be the oblique distance of the i1 time range ambiguity, ? pitching steering vector be expressed as:

Wherein, the transposition of T representing matrix or vector, d is the array element distance of each row submatrix, and d is the spacing between any two adjacent row submatrixs, and λ is the wavelength that airborne radar transmits, and N represents the array element line number of non-working side array antenna.

Utilize extremely the pitching guiding matrix B that forms m row submatrix _m:

(1.2) due to when i1 gets 2 to N _rltime, for the long-range clutter angle of pitch, extremely very approaching each other, so pitching steering vector extremely there is stronger correlativity.Therefore, the pitching guiding matrix B to m row submatrix _mcarry out feature decomposition, in the pitching guiding matrix B of m row submatrix _mall eigenwerts in, choose K maximum eigenwert (larger than noise characteristic value), K<N _rl, the value of K is determined according to the eigenwert of the white Gaussian noise vector n of NMP dimension, guarantees that each maximum eigenwert of choosing is greater than all eigenwerts of n.

The K that utilization is chosen a maximum eigenwert characteristic of correspondence vector forms the array manifold matrix A of m row submatrix _m,

A _m＝[a _m1,a _m2,···,a _mK]。

Wherein, a _m1to a _mKrepresent respectively K the maximum eigenwert characteristic of correspondence vector of choosing.

(1.3) utilize the non-working side array antenna transponder pulse signal of airborne radar, and receive corresponding echo data, it is four-dimensional echo data X that airborne radar receives corresponding echo data _{n * M * P * L}, wherein, N represents the array element line number of non-working side array antenna, M represents the array element columns of non-working side array antenna, and the umber of pulse of the non-working side array antenna transmitting that P is airborne radar, L represents the range unit number of airborne radar.

At four-dimensional echo data X _{n * M * P * L}in, extract the 2-D data X corresponding with l range unit of m row submatrix _{n * P}(m, l), then draws 2-D data X _{n * P}the pitching covariance matrix R of (m, l) _ml:

R _ml＝(X _N×P(m,l)X ^H _N×P(m,l))/P

Wherein, the conjugate transpose of H representing matrix.

(1.4), because the same row submatrix different distance of non-working side array antenna unit amplitude phase error is identical, can utilize amplitude phase error vector that l range unit of m row submatrix is corresponding as the error width phase vector of m row submatrix.Now can estimate according to following steps the amplitude phase error vector delta of m row submatrix _m:

To R _mlcarry out feature decomposition, at R _mlall eigenwerts in choose N-K minimum eigenwert, utilize N-K the minimum eigenwert characteristic of correspondence vector formation noise subspace U choosing _n.

Structure cost function J _m, cost function J _mexpression formula be:

$J_m=\underset{k^{\&prime;}=1}{\overset{K}{\mathrm{\&Sigma;}}}{\mathrm{\&delta;}}_m^H{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}}^H{U}_N{U}_N^H{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}}{\mathrm{\&delta;}}_m$

Wherein, the conjugate transpose of H representing matrix, k' is that natural number and k' get 1 to K, α _mk'=diag (a _mk'), a _mk'represent k' the eigenwert characteristic of correspondence vector of choosing in sub-step (1.3), diag () represents to ask diagonal matrix, diag (a _mk') represent with a _mk'in the diagonal matrix (a that forms as the elements in a main diagonal of each element _mk'in each element by its arrangement of elements, be set in sequence in α _mkprincipal diagonal on).

Order ${\mathrm{\&Omega;}}_m=\underset{k^{\&prime;}=1}{\overset{K}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}}^H{U}_N{U}_N^H{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}},$ Cost function J _mbe expressed as:

$J_m={\mathrm{\&delta;}}_m^H{\mathrm{\&Omega;}}_m{\mathrm{\&delta;}}_m$

Then set up following quadratic constraints optimization problem:

$\left\{\begin{array}{c}{\mathrm{\&delta;}}_m=\underset{{\mathrm{\&delta;}}_m}{\min }{\mathrm{\&delta;}}_m^H{\mathrm{\&Omega;}}_m{\mathrm{\&delta;}}_m\\ s.t.{\mathrm{\&delta;}}_m^{H}f=1\end{array}\right.$

Wherein, for constraint condition, the conjugate transpose of H representing matrix, f=[1,0,, 0] ^t, f is the column vector of N * 1 dimension, and first element of f is 1, and all the other elements are 0.When the 1st range ambiguity clutter with respect to the angle of pitch of non-working side array antenna is to N _rinferior range ambiguity clutter with respect to the angle of pitch of non-working side array antenna is accurate when known, use method of Lagrange multipliers to solve the amplitude phase error vector delta of m row submatrix _m, δ _mfor:

${\mathrm{\&delta;}}_m=\frac{{\mathrm{\&Omega;}}_m^{-1}f}{f^T{\mathrm{\&Omega;}}_m^{-1}f}$

Wherein, the transposition of T representing matrix or vector, Shang Biao ?1 representing matrix contrary.

Step 2, according to the amplitude phase error vector of each row submatrix, carries out matched filtering to the echo data of respective column submatrix in pitching dimension, show that respective column submatrix clutter suppresses rear data.After utilizing the 1st row submatrix clutter to suppress rear data to the M row submatrix clutter inhibition, data composition airborne radar clutter suppresses rear data after m row submatrix clutter suppresses, data are expressed as m gets 1 to M, and airborne radar clutter suppresses rear data for: ${\tilde{X}}_{N\&times;M\&times;P\&times;L}=[{\tilde{X}}_{N\&times;P\&times;L}\left(1\right),...,{\tilde{X}}_{N\&times;P\&times;L}\left(M\right)].$

Data after m row submatrix clutter suppresses after being suppressed by the 1st range unit clutter of m row submatrix, after L range unit clutter inhibition of data to the m row submatrix, data form, ${\tilde{X}}_{N\&times;P\&times;L}\left(m\right)$ For:

${\tilde{X}}_{N\&times;P\&times;L}\left(m\right)={[\tilde{X}}_{N\&times;P}(m,1),...,{\tilde{X}}_{N\&times;P}(m,l),...,{\tilde{X}}_{N\&times;P}(m,L)].$

Wherein, represent that l range unit clutter of m row submatrix suppresses rear data, l gets 1 to L.

Show that l range unit clutter of m row submatrix suppresses rear data process comprise following sub-step:

First draw the short range clutter steering vector a that l range unit of m row submatrix is corresponding _1mljourney clutter steering vector (main beam steering vector) a corresponding with l range unit of m row submatrix _2ml.Owing to removing the 0th angle of pitch sine value corresponding to range ambiguity outward, extremely it is very approaching each other, extremely all be in close proximity to 0.Angle of pitch sine value corresponding to main beam when so the present invention uses airborne radar to transmit substitute extremely

So, the short range clutter steering vector a that l range unit of m row submatrix is corresponding _1mllong-range clutter steering vector (main beam steering vector) a corresponding with l range unit of m row submatrix _2mlbe respectively:

Wherein, the transposition of T representing matrix or vector, d is the array element distance of each row submatrix, and d is the spacing between any two adjacent row submatrixs, and λ is the wavelength that airborne radar transmits, and N represents the array element line number of non-working side array antenna, represent the angle of pitch corresponding to range ambiguity the 0th time, represent main beam pointing when airborne radar transmits.

Then according to following formula, draw the short range clutter steering vector with error compensation that l range unit of m row submatrix is corresponding the long-range clutter steering vector with error compensation corresponding with l range unit of m row submatrix

According to following formula construction, go out pitching filtering weight vector corresponding to l range unit of m row submatrix

$\left\{\begin{array}{c}{\tilde{w}}_{\mathrm{ml}}^H{\tilde{a}}_{1\mathrm{ml}}=0\\ {\tilde{w}}_{\mathrm{ml}}^H{\tilde{a}}_{2\mathrm{ml}}=1\end{array}\right.$

Wherein, the conjugate transpose of H representing matrix, represent generalized inverse computing.

Drawing pitching filtering weight vector corresponding to l range unit of m row submatrix afterwards, at four-dimensional echo data X _{n * M * P * L}in, extract the 2-D data X corresponding with l range unit of m row submatrix _{n * P}(m, l), utilizes pitching filtering weight vector corresponding to l range unit of m row submatrix to the 2-D data X extracting _{n * P}(m, l) carries out matched filtering in pitching dimension, show that l range unit clutter of m row submatrix suppresses rear data for:

${\tilde{X}}_{N\&times;P}(m,l)={\tilde{w}}_{\mathrm{ml}}^H{X}_{N\&times;P}(m,l)$

The above-mentioned pitching filtering method of error correction that exists is actually a row submatrix weighting building-up process, utilize the present invention to estimate that the row submatrix amplitude phase error vector obtaining compensates pitching filtering weight vector, then utilize the weight vector after compensation in pitching dimension, to carry out matched filtering to the corresponding clutter data of this row submatrix.Compare with the pitching filtering Processing Algorithm of not carrying out amplitude and phase error correction, due to process error correction, this algorithm can make formed recess aim at more accurately short range clutter, thereby effectively suppress short range clutter, and retrain weight vector simultaneously, the gain of long-range clutter is remained unchanged, so main-beam clutter have very large residue.Neighbor distance unit clutter stationarity has been strengthened, that is to say pending range unit is carried out when follow-up EFA (Extended Factored Algorithm) processes obtaining from neighbor distance unit abundant independent same distribution sample estimate covariance matrix, this algorithm of simulation results show can reach good short range clutter inhibition, and after EFA processing, long-range clutter has obtained suppressing well.

Effect of the present invention can further illustrate by following emulation experiment:

1, emulated data obtains and experiment condition

1) in this experiment, simulation parameter is as shown in table 1:

Table 1 emulation experiment parameter list

Parameter name	Numerical value	Parameter name numerical value
			Antenna array N * M	16×8	Earth radius R e/km6378
Pulse number P	64	Carrier aircraft speed v/(ms -1)125
			Main beam and front angle	90	Carrier aircraft height H/m7000
Front and velocity angle α/(°)	-90	Light velocity C/ (ms -1)3×10 8
			Distance samples frequency f s/MHz	1	Pulse repetition rate 4000
Miscellaneous noise ratio CNR/ (dB)	80	Wavelength X/m0.25

2) airborne radar of the present invention adopts the rectangular surfaces array antenna of 16 * 8 dimensions, take the 5th row submatrix of front as example is to the actual web phase error vector of this row submatrix with estimate that the amplitude phase error vector of this row submatrix obtain contrasts, and the pitch orientation figure (the pitch orientation figure of the 5th the row submatrix that traditional space-time two-dimensional adaptive processing method draws) of the 5th row submatrix of array element amplitude and phase error correction will do not carried out, carry out the pitch orientation figure (drawing by the present invention) of the 5th row submatrix after array element amplitude and phase error correction, and contrast between the static pitch orientation figure of the 5th row submatrix.

3) for the original echo data that have array element amplitude phase error, carry out pitching filtering processing (traditional space-time two-dimensional adaptive processing method), then carry out EFA processing, and by the range Doppler figure contrast of its range Doppler figure and original echo data.

4) error calibration method that exists the original echo data of array element amplitude phase error first to utilize the present invention to propose is proofreaied and correct to (the inventive method), then the echo data after pitching filtering being processed carries out EFA processing, by the range Doppler figure contrast of its range Doppler figure and original echo data.

2, emulated data result and analysis

In emulation experiment, antenna array adds 10% random magnitude error and the random phase errors of 6 °.Because 8 row submatrix situations of front are similar, wherein the 5th the row submatrix of only take is herein example explanation array element amplitude phase error estimated result.With reference to Fig. 2 a, be the range error of the 5th each array element of row submatrix estimated according to the present invention in emulation experiment and the comparison diagram of the actual margin error of the 5th each array element of row submatrix.In Fig. 2 a, transverse axis represents the array element sequence number of the 5th row submatrix, and the longitudinal axis represents the range error of the 5th each array element of row submatrix, and unit is dB.With reference to Fig. 2 b, be the phase error of the 5th each array element of row submatrix estimated according to the present invention in emulation experiment and the comparison diagram of the substantial phase error of the 5th each array element of row submatrix.In Fig. 2 b, transverse axis represents the array element sequence number of the 5th row submatrix, and the longitudinal axis represents the phase error of the 5th each array element of row submatrix, and unit is degree.From Fig. 2 a and Fig. 2 b, can find out, the amplitude phase error curve of the 5th each array element of row submatrix estimating by the present invention is substantially identical with the actual web phase graph of errors of the 5th each array element of row submatrix.

In order to eliminate single experiment adventitious circumstances, for array element amplitude phase error, estimate the impact of accuracy, in emulation experiment, supplement carry out 100 Monte ?Carlo experiment, with reference to Fig. 2 c, amplitude square error root curve synoptic diagram for the 5th each array element of row submatrix estimating according to the present invention in emulation experiment, in Fig. 2 c, transverse axis represents the array element sequence number of the 5th row submatrix, and the longitudinal axis represents the amplitude square error root of the 5th each array element of row submatrix, and unit is dB.With reference to Fig. 2 d, the phase place square error root curve synoptic diagram of the 5th each array element of row submatrix estimating according to the present invention in emulation experiment, in Fig. 2 c, transverse axis represents the array element sequence number of the 5th row submatrix, the longitudinal axis represents the phase place square error root of the 5th each array element of row submatrix, and unit is degree.From Fig. 2 c and Fig. 2 d, find out, the width phase square error root of the 5th each array element of row submatrix estimating according to the present invention in emulation experiment is smaller, has further proved the accuracy of amplitude phase error method of estimation of the present invention.

In emulation experiment, utilize the amplitude phase error of the 5th each array element of row submatrix estimating, according to the 5th row submatrix carried out to array amplitude and phase error correction.With reference to Fig. 3 a, for the 5th row submatrix in emulation experiment carried out pitch orientation figure before array amplitude and phase error correction (in corresponding diagram 3a " have array element amplitude phase error), carries out pitch orientation figure after array amplitude and phase error correction the comparison diagram of (in corresponding diagram 3a " after array element amplitude and phase error correction), static pitch orientation figure (in corresponding diagram 3a " static state "), short range clutter pitch orientation figure (in corresponding diagram 3a " the short range clutter angle of pitch ").With reference to Fig. 3 b, it is the local enlarged diagram of Fig. 3 a.In Fig. 3 a and Fig. 3 b, the short range clutter angle of pitch represents with vertical line, and transverse axis represents the angle of pitch, and unit is degree, and the longitudinal axis represents pitch orientation figure, the dB of unit.From Fig. 3 a and Fig. 3 b, find out, before carrying out array amplitude and phase error correction, pitch orientation figure recess departs from short range clutter position, and recess position and short range clutter position have obtained good registration after carrying out array amplitude and phase error correction, the validity of the error calibration method that this is proposed by the invention is described.

In order to verify the benefit to pitching filtering after error correction herein, in emulation experiment, provided the range Doppler figure of the original echo data that have array element amplitude phase error, do not carry out in array amplitude and phase error correction situation the echo data range Doppler figure of (traditional space-time two-dimensional adaptive processing method) after pitching filtering, do not carry out the echo data range Doppler figure after pitching filtering and EFA processing in array amplitude and phase error correction situation, carry out in array amplitude and phase error correction situation the echo data range Doppler figure of (space-time two-dimensional adaptive processing method of the present invention) after pitching filtering, carry out the echo data range Doppler figure after pitching filtering and EFA processing in array amplitude and phase error correction situation.With reference to Fig. 4, for there is the range Doppler figure of the original echo data of array element amplitude phase error in emulation experiment, with reference to Fig. 5 a, for not carrying out in emulation experiment in array amplitude and phase error correction situation through the filtered echo data range Doppler of pitching figure, with reference to Fig. 5 b, for not carrying out the echo data range Doppler figure after pitching filtering and EFA processing in array amplitude and phase error correction situation in emulation experiment, with reference to Fig. 6 a, for carrying out in emulation experiment in array amplitude and phase error correction situation through the filtered echo data range Doppler of pitching figure, with reference to Fig. 6 b, for carrying out the echo data range Doppler figure after pitching filtering and EFA processing in array amplitude and phase error correction situation in emulation experiment.In Fig. 4, Fig. 5 a, Fig. 5 b, Fig. 6 a and Fig. 6 b, transverse axis represents Doppler's unit number, and the longitudinal axis represents range unit, and the gray-scale value of pixel represents the amplitude of echo data, and pixel is whiter, illustrates that the range value of corresponding echo data is larger.

From Fig. 5 a, Fig. 5 b, Fig. 6 a and Fig. 6 b, can find out, not carrying out the filtered range Doppler figure of pitching in array amplitude and phase error correction situation with traditional space-time two-dimensional adaptive processing method compares, the range Doppler figure short range clutter that is undertaken drawing after pitching filtering by method of the present invention has obtained good inhibition, and clutter has obtained enhancing apart from stationarity.

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

Claims (4)

1. the airborne radar clutter suppression method based on array element amplitude and phase error correction, it is characterized in that, on described airborne radar, be provided with non-working side array antenna, described non-working side array antenna is comprised of M row submatrix, each row submatrix is the even linear array being comprised of N array element, M and N are natural number, and the array element distance of each row submatrix is d, and the spacing between any two adjacent row submatrixs is d;

The described airborne radar clutter suppression method based on array element amplitude and phase error correction comprises the following steps:

Step 1, utilizes the non-working side array antenna transponder pulse signal of airborne radar, and receives corresponding echo data, and it is four-dimensional echo data that airborne radar receives corresponding echo data; According to described four-dimensional echo data, estimate the amplitude phase error vector of each row submatrix;

Step 2, according to the amplitude phase error vector of m row submatrix, show that l range unit clutter of m row submatrix suppresses rear data m gets 1 to M, and l gets 1 to L, the umber of pulse of the non-working side array antenna transmitting that P is airborne radar, and L represents the range unit number of airborne radar; After utilizing the 1st range unit clutter inhibition of m row submatrix, L rear data of range unit clutter inhibition of data to the m row submatrix form m the rear data of row submatrix clutter inhibition data after utilizing the 1st row submatrix clutter to suppress data after suppressing to M row submatrix clutter form airborne radar clutter and suppress rear data

2. the airborne radar clutter suppression method based on array element amplitude and phase error correction as claimed in claim 1, is characterized in that, the concrete sub-step of described step 1 is:

The range ambiguity number of times of (1.1) l the corresponding clutters of range unit is N _rlinferior, the i1 time range ambiguity clutter with respect to the angle of pitch of non-working side array antenna is ? pitching steering vector be expressed as:

Wherein, the transposition of T representing matrix or vector, λ is the wavelength that airborne radar transmits;

Utilize extremely the pitching guiding matrix B that forms m row submatrix _m:

(1.2) the pitching guiding matrix B to m row submatrix _mcarry out feature decomposition, in the pitching guiding matrix B of m row submatrix _mall eigenwerts in, choose K maximum eigenwert, K<N _rl; The K that utilization is chosen a maximum eigenwert characteristic of correspondence vector forms the array manifold matrix A of m row submatrix _m,

A _m＝[a _m1,a _m2,···,a _mK]；

Wherein, a _m1to a _mKrepresent respectively K the maximum eigenwert characteristic of correspondence vector of choosing;

(1.3) airborne radar receives corresponding echo data and is expressed as four-dimensional echo data X _{n * M * P * L}, at described four-dimensional echo data X _{n * M * P * L}in, extract the 2-D data X corresponding with l range unit of m row submatrix _{n * P}(m, l), then draws 2-D data X _{n * P}the pitching covariance matrix R of (m, l) _ml:

R _ml＝(X _N×P(m,l)X ^H _N×P(m,l))/P

Wherein, the conjugate transpose of H representing matrix;

(1.4) to R _mlcarry out feature decomposition, at R _mlall eigenwerts in choose N-K minimum eigenwert, utilize N-K the minimum eigenwert characteristic of correspondence vector formation noise subspace U choosing _n;

Structure cost function J _m, cost function J _mexpression formula be:

$J_m=\underset{k^{\&prime;}=1}{\overset{K}{\mathrm{\&Sigma;}}}{\mathrm{\&delta;}}_m^H{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}}^H{U}_N{U}_N^H{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}}{\mathrm{\&delta;}}_m$

Wherein, the conjugate transpose of H representing matrix, k' is that natural number and k' get 1 to K, α _mk'=diag (a _mk'), a _mk'represent k' the eigenwert characteristic of correspondence vector of choosing in sub-step (1.3), diag (a _mk') represent with a _mk'in the diagonal matrix that forms as the elements in a main diagonal of each element;

Order ${\mathrm{\&Omega;}}_m=\underset{k^{\&prime;}=1}{\overset{K}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}}^H{U}_N{U}_N^H{\mathrm{\&alpha;}}_{{\mathrm{mk}}^{\&prime;}},$ Cost function J _mbe expressed as:

$J_m={\mathrm{\&delta;}}_m^H{\mathrm{\&Omega;}}_m{\mathrm{\&delta;}}_m$

Then set up following quadratic constraints optimization problem:

$\left\{\begin{array}{c}{\mathrm{\&delta;}}_m=\underset{{\mathrm{\&delta;}}_m}{\min }{\mathrm{\&delta;}}_m^H{\mathrm{\&Omega;}}_m{\mathrm{\&delta;}}_m\\ s.t.{\mathrm{\&delta;}}_m^{H}f=1\end{array}\right.$

Wherein, δ _mamplitude phase error vector for m row submatrix to be solved; F is the column vector of N * 1 dimension, and first element of f is 1, and all the other elements are 0;

By solving described quadratic constraints optimization problem, estimate the amplitude phase error vector delta of m row submatrix _m.

3. the airborne radar clutter suppression method based on array element amplitude and phase error correction as claimed in claim 2, it is characterized in that, in sub-step (1.4), use method of Lagrange multipliers to solve described quadratic constraints optimization problem, described in the amplitude phase error vector delta of m row submatrix solving _mfor:

${\mathrm{\&delta;}}_m=\frac{{\mathrm{\&Omega;}}_m^{-1}f}{f^T{\mathrm{\&Omega;}}_m^{-1}f}$

Wherein, the transposition of T representing matrix or vector, Shang Biao ?1 representing matrix contrary.

4. the airborne radar clutter suppression method based on array element amplitude and phase error correction as claimed in claim 1, is characterized in that, in step 2, show that l range unit clutter of m row submatrix suppresses rear data process comprise following sub-step:

First draw the short range clutter steering vector a that l range unit of m row submatrix is corresponding _1mlthe journey clutter steering vector a corresponding with l range unit of m row submatrix _2ml;

Wherein, the transposition of T representing matrix or vector, λ is the wavelength that airborne radar transmits, represent the angle of pitch corresponding to range ambiguity the 0th time, represent main beam pointing when airborne radar transmits;

Then according to following formula, draw the short range clutter steering vector with error compensation that l range unit of m row submatrix is corresponding the long-range clutter steering vector with error compensation corresponding with l range unit of m row submatrix

Wherein, represent that Hadamard is long-pending; δ _mit is the amplitude phase error vector of m row submatrix;

According to following formula construction, go out pitching filtering weight vector corresponding to l range unit of m row submatrix

$\left\{\begin{array}{c}{\tilde{w}}_{\mathrm{ml}}^H{\tilde{a}}_{1\mathrm{ml}}=0\\ {\tilde{w}}_{\mathrm{ml}}^H{\tilde{a}}_{2\mathrm{ml}}=1\end{array}\right.$

Wherein, the conjugate transpose of H representing matrix;

At described four-dimensional echo data X _{n * M * P * L}in, extract the 2-D data X corresponding with l range unit of m row submatrix _{n * P}(m, l), according to pitching filtering weight vector corresponding to l range unit of m row submatrix show that l range unit clutter of m row submatrix suppresses rear data ${\tilde{X}}_{N\&times;P}(m,l)$ For: ${\tilde{X}}_{N\&times;P}(m,l)={\tilde{w}}_{\mathrm{ml}}^H{X}_{N\&times;P}(m,l).$