CN105785326A - Non-forward looking array radar clutter spectrum registration optimization method - Google Patents

Abstract

The invention discloses a non-forward looking array radar clutter spectrum registration optimization method, which is mainly characterized by obtaining L range gate clutter data received by a non-forward looking array radar; calculating covariance matrixes corresponding to the L range gate clutter data respectively; selecting the covariance matrix corresponding to the kth range gate as the covariance matrix corresponding to a range gate to be detected, and with the covariance matrix corresponding to the range gate to be detected being as a reference, calculating transformation matrix Tl' of the l'th range gate clutter data Xl', and carrying out compensation processing on the l'th range gate clutter data Xl' by utilizing the transformation matrix Tl' of the Xl' to obtain l'th range gate clutter data Yl' after compensation; selecting the kth range gate clutter data Xk as reference clutter, and carrying out calculating according to the l'th range gate clutter data Yl' after compensation to obtain the covariance matrix Rk corresponding to the kth range gate; and carrying out calculation according to the covariance matrix Rk corresponding to the kth range gate after compensation to obtain a non-forward looking array radar clutter suppression improvement factor obtained after space-time adaptive filtering processing.

Description

A kind of optimization method of non-working side battle array radar clutter spectrum registration

Technical field

The invention belongs to Radar Technology field, compose the optimization method of registration particularly to a kind of non-working side battle array radar clutter, it is adaptable to solve the radar clutter spectrum precision limitation problem utilizing clutter registration method to obtain.

Background technology

Space-time adaptive processes (space-timeadaptiveprocessing, STAP) technology is significant for the target detection at a slow speed under strong clutter background, and it is closely related with its clutter covariance matrix estimated accuracy to carry out clutter recognition performance corresponding during target detection at a slow speed.Traditional space-time adaptive processes (STAP) technology can obtain good clutter recognition purpose for positive side-looking battle array radar；And for non-working side battle array radar, owing to non-working side battle array radar clutter frequency spectrum has distance dependencies, make to directly utilize the covariance matrix closing on range gate data estimation non-working side battle array radar clutter not mate with actual value, thus causing that utilizing space-time adaptive to process (STAP) technology suppresses the hydraulic performance decline of non-working side battle array radar clutter.

Therefore, the suppression radar clutter purpose that obtain, it is necessary to space-time adaptive is processed (STAP) technology and carries out corresponding pretreatment；General roadmap is that the data heterogeneity that the distance dependencies to non-working side battle array radar clutter frequency spectrum produces compensates the variation characteristic alleviating radar clutter with distance, and using range gate clutter to be detected as sample for reference, Doppler domain is carried out or angle Doppler domain compensates respectively through treating each auxiliary range gate that detecting distance door closes on, make to be distributed during the power sky closing on range gate clutter to reach unanimity with unit to be detected, thus eliminating the distance dependencies of radar clutter.

Existing compensation method mainly has Doppler effect correction (Dopplercompensation, DC) method, angle Doppler effect correction (angle-Dopplercompensation, ADC) method, self adaptation angle Doppler effect correction (adaptiveangle-Dopplercompensation, A2DC) method and based on clutter spectrum registration (Registration-basedcompensation, RBC) method；These methods reduce the radar clutter distance dependencies under airborne radar non-working side battle array configuration to a certain extent so that be obtained in that than the radar clutter rejection directly handled well when using space-time adaptive to process (STAP) technology；But, DC method has carried out main-lobe clutter spectrum center compensation only in Doppler domain；ADC method carries out clutter spectrum center compensation at angle Doppler domain, and needs the systematic parameter utilizing inertial navigation to provide；Although A2DC method does not need to know configuration parameter, and radar clutter rejection is also fine in the ideal case, but the covariance matrix extremely unstable of each distance unit radar clutter in reality, cause that radar clutter rejection declines a lot；RBC method is capable of the full remuneration of radar clutter, and good radar clutter rejection can be obtained in the ideal case, but traditional RBC method utilizes time domain to smooth sub-snap carries out the extraction of radar clutter spectrum peak so that the estimated accuracy of radar clutter spectrum peak and the robustness of RBC method all can to using space-time adaptive process (STAP) technical performance to impact；The iteration self-adapting algorithm (Iterativeadaptivealgorithm, IAA) proposed for 2010, also can obtain accurate clutter spectrum in single snap situation and estimate.

Summary of the invention

Have problems for above prior art, it is an object of the invention to propose the optimization method of a kind of non-working side battle array radar clutter spectrum registration, the optimization method of this kind of non-working side battle array radar clutter spectrum registration is traditional based on clutter spectrum registration (Registration-basedcompensation for using, RBC) method carries out precision restricted problem during radar clutter Power estimation, non-working side battle array radar clutter is composed to utilize iteration self-adapting algorithm (IAA) to estimate, non-working side battle array radar clutter Power estimation precision is improve with this, after making the spectrum pretreatment of RBC method registration non-working side battle array radar clutter, its clutter is evenly, and then improve the clutter performance using space-time adaptive treatment technology to suppress non-working side battle array radar.

For reaching above-mentioned technical purpose, the present invention adopts the following technical scheme that and is achieved.

The optimization method of a kind of non-working side battle array radar clutter spectrum registration, comprises the following steps:

Step 1, obtains L the range gate clutter data that non-working side battle array radar receives, is designated as X₁,X₂,X₃…X_L；Wherein, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises；

Step 2, calculates each self-corresponding covariance matrix of L range gate clutter data that non-working side battle array radar receives respectively, is designated as

Step 3, chooses the covariance matrix that kth range gate is correspondingAs range gate correspondence covariance matrix to be detected, and with described range gate correspondence covariance matrix to be detected for reference, calculate the l' range gate clutter data X_l'Transformation matrix T_l'；Wherein, k ∈ 1,2 ..., L}, l' ∈ 1,2 ..., L} and l' ≠ k；

Step 4, utilizes the l' range gate clutter data X_l'Transformation matrix T_l', to the l' range gate clutter data X_l'Carry out registration process, obtain the l' range gate clutter data Y after registration_l'；

Step 5, chooses kth range gate clutter data X_kAs reference spurs, and according to the l' range gate clutter data Y after registration_l', calculate the covariance matrix R that the kth range gate after obtaining registration is corresponding_k；

Step 6, according to the covariance matrix R that the kth range gate after registration is corresponding_k, calculate the non-working side battle array radar clutter after space-time adaptive Filtering Processing and suppress improvement factor IF (f_dt), the non-working side battle array radar clutter after described space-time adaptive Filtering Processing suppresses improvement factor IF (f_dt) for weighing the registration effect of non-working side battle array radar clutter spectrum；Wherein, k ∈ 1,2 ..., L}, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

Beneficial effects of the present invention: the inventive method utilizes iteration self-adapting algorithm (IAA in single snap situation, Iterativeadaptivealgorithm) non-working side battle array radar clutter spectrum is estimated, non-working side battle array radar clutter Power estimation precision can be improved, and the covariance matrix according to the relation reconstruct non-working side battle array radar clutter between non-working side battle array radar clutter Spectral structure characteristic and non-working side battle array radar clutter spectrum covariance matrix, make the pretreated non-working side battle array radar clutter of registration evenly, and then improve the clutter performance of space-time adaptive treatment technology suppression non-working side battle array radar.

Accompanying drawing explanation

Below in conjunction with the drawings and specific embodiments, the present invention is described in further detail.

Fig. 1 is the general flow chart of the inventive method；

Fig. 2 (a) does not compensate, for the lower No. 80 distance unit of positive side-looking battle array radar, the clutter spectrum schematic diagram that clutter spectrum obtains,

The clutter spectrum schematic diagram that Fig. 2 (b) uses RBC method to obtain for the lower No. 80 distance unit of positive side-looking battle array radar,

The clutter spectrum schematic diagram that Fig. 2 (c) uses IAA-RBC method to obtain for the lower No. 80 distance unit of positive side-looking battle array radar,

Fig. 2 (d) is the lower No. 80 distance real clutter spectrum schematic diagrams of unit of positive side-looking battle array radar；

Fig. 3 is IAA-RBC method clutter recognition improvement factor and traditional RBC method clutter recognition improvement factor comparison diagram.

Detailed description of the invention

With reference to Fig. 1, for the general flow chart of the inventive method；The optimization method of a kind of non-working side battle array radar clutter spectrum registration, comprises the following steps:

Step 1, obtains L the range gate clutter data that non-working side battle array radar receives, is designated as X₁,X₂,X₃…X_L；Wherein, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

Specifically, L the range gate clutter data X that non-working side battle array radar receives is obtained₁,X₂,X₃…X_L, described L range gate clutter data X₁,X₂,X₃…X_LRespectively column vector is tieed up in MN × 1；Wherein, M represents the total number of pulse that non-working side battle array radar receives, and N represents the reception total number of passage antenna of non-working side battle array radar.

Step 2, calculates each self-corresponding covariance matrix of L range gate clutter data that non-working side battle array radar receives respectively, is designated as

Specifically, l ∈ 1,2 ..., L}, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises, and calculates the l covariance matrix corresponding to range gateSub-step be:

The 2.1 normalization spatial frequencys interval [-1 setting non-working side battle array radar respectively, 1] and the normalization Doppler frequency interval [-1,1] of non-working side battle array radar, and respectively to described normalization spatial frequency interval [-1,1] uniform sampling is carried out, it is thus achieved that K_sIndividual sampled point；Described normalization Doppler frequency interval [-1,1] is carried out uniform sampling, it is thus achieved that K_tIndividual sampled point；Use f_s,nRepresent the normalization spatial frequency of the n-th sampled point on the non-working side battle array radar normalization spatial frequency interval set, use f_d,mRepresent the normalization Doppler frequency of m-th sampled point on the non-working side battle array radar normalization Doppler frequency interval set；Wherein, n ∈ 1,2 ..., K_s, m ∈ 1,2 ..., K_t; then non-working side battle array radar normalization Doppler frequency interval that the is non-working side battle array radar normalization spatial frequency of setting is interval and that set is built into plane when non-working side battle array radar initializes empty; calculating obtains when non-working side battle array radar initializes empty any one sampled point in plane, and (m, n) the clutter power initial value at place isIts expression formula is:

${\hat{P}}_{m,n}^{\left(0\right)}={\left|\frac{v^H(f_{d,m},f_{s,n})X_l}{v^H(f_{d,m},f_{s,n})v(f_{d,m},f_{s,n})}\right|}^2,m\∈\{1,2,\mathrm{...},K_t\},n\∈\{1,2,\mathrm{...},K_s\}$

Wherein, v (f_d,m,f_s,n) represent when non-working side battle array radar initializes empty any one sampled point in plane (m, steering vector during n) place empty, andf_s,nRepresent the normalization spatial frequency of the n-th sampled point, f on the non-working side battle array radar normalization spatial frequency interval set_d,mRepresent the normalization Doppler frequency of m-th sampled point on the non-working side battle array radar normalization Doppler frequency interval set,Represent that kronecker amasss, ()^HRepresent conjugate transpose, ()^TRepresent transposition, | | represent the operation that takes absolute value, K_sRepresent the sampled point number that the non-working side battle array radar normalization spatial frequency interval set comprises, K_tRepresent the sampled point number that the non-working side battle array radar normalization Doppler frequency interval set comprises.

2.2 initialize: β represents default iterations, and i represents iterations, and initial value is 1.

2.3 calculate successively and obtain when the non-working side battle array radar after ith iteration initializes empty any one sampled point (m, n) covariance matrix at place in planeAny one sampled point (m, n) the clutter power value at place in plane when initializing empty with the non-working side battle array radar after ith iteration

Wherein, u ∈ 1,2 ..., K_t, v ∈ 1,2 ..., K_s, n ∈ 1,2 ..., K_s, m ∈ 1,2 ..., K_t, u ≠ m, v ≠ n, K_sRepresent the sampled point number that the non-working side battle array radar normalization spatial frequency interval set comprises, K_tRepresent the sampled point number that the non-working side battle array radar normalization Doppler frequency interval set comprises.

Specifically, the non-working side battle array radar clutter power distribution matrix after ith iterationExpression formula be:

${\tilde{P}}^{\left(i\right)}=diag\left(\left[\begin{array}{cccccc}{\hat{P}}_{1,1}^{(i-1)}& {\hat{P}}_{2,1}^{(i-1)}& \mathrm{...}& {\hat{P}}_{m,n}^{(i-1)}& \mathrm{...}& {\hat{P}}_{K_t,K_s}^{(i-1)}\end{array}\right]\right)$

Wherein, diag () represents diagonal matrix, and namely diagonal element isDiagonal matrix,Represent when the non-working side battle array radar after the i-th-1 time iteration initializes empty the clutter power value of any one sample point in plane.

For any one sampled point (m, n) the clutter power value at place in plane when the non-working side battle array radar after calculating ith iteration initializes emptyFirst any one sampled point (m, n) covariance matrix at place in plane are calculated when the non-working side battle array radar after ith iteration initializes emptyIts expression formula is:

${\hat{R}}_{m,n}^{\left(i\right)}=\underset{u=1,u\≠m}{\overset{K_t}{\Σ}}\underset{v=1,v\≠n}{\overset{K_s}{\Σ}}{\hat{P}}_{u,v}^{\left(i\right)}v(f_{d,u},f_{s,v})v^H(f_{d,u},f_{s,v}),$

Wherein, u ∈ 1,2 ..., K_t, v ∈ 1,2 ..., K_s, n ∈ 1,2 ..., K_s, m ∈ 1,2 ..., K_t, u ≠ m, v ≠ n, K_sRepresent the sampled point number that the non-working side battle array radar normalization spatial frequency interval set comprises, K_tRepresent the sampled point number that the non-working side battle array radar normalization Doppler frequency interval set comprises,Represent when the non-working side battle array radar after ith iteration initializes empty any one sampled point (u, v) the clutter power value at place, v (f in plane_d,u,f_s,v) represent plane up-sampling point when non-working side battle array radar initializes empty (u, steering vector during v) place empty, f_d,uRepresent the normalization Doppler frequency of u sampled point, f on the non-working side battle array radar normalization Doppler frequency interval set_s,vRepresent the normalization spatial frequency of v sampled point, () on the non-working side battle array radar normalization spatial frequency interval set^HRepresent conjugate transpose.

Then, any one sampled point (m, n) covariance matrix at place in plane when initializing empty according to the non-working side battle array radar after ith iterationCalculate when the non-working side battle array radar after ith iteration initializes empty any one sampled point (m, n) the clutter power value at place in planeIts expression formula is:

${\hat{P}}_{m,n}^{\left(i\right)}=|\frac{v^H(f_{d,m},f_{s,n}){\left({\hat{R}}_{m,n}^{\left(i\right)}\right)}^{-1}{X}_1}{v^H(f_{d,m},f_{s,n}){\left({\hat{R}}_{m,n}^{\left(i\right)}\right)}^{-1}v(f_{d,m},f_{s,n})}{|}^2$

Wherein, v (f_d,m,f_s,n) represent when non-working side battle array radar initializes empty any one sampled point in plane (m, steering vector during n) place empty,Represent when the non-working side battle array radar after ith iteration initializes empty any one sampled point (m, n) covariance matrix at place, f in plane_s,nRepresent the normalization spatial frequency of the n-th sampled point, f on the non-working side battle array radar normalization spatial frequency interval set_d,mThe normalization Doppler frequency of m-th sampled point on the non-working side battle array radar normalization Doppler frequency interval that expression sets, n ∈ 1,2 ..., K_s, m ∈ 1,2 ..., K_t, K_sRepresent the sampled point number that the non-working side battle array radar normalization spatial frequency interval set comprises, K_tRepresent the sampled point number that the non-working side battle array radar normalization Doppler frequency interval set comprises.

If 2.4Or i < β, makes i add 1, return sub-step 2.3；

IfOr i >=β, then iteration terminates, and now obtains steering vector v (f during plane up-sampling point (1,1) place when non-working side battle array radar initializes empty empty_d,1,f_s,1Plane up-sampling point (K when)-non-working side battle array radar initializes empty_t,K_s) place empty time steering vectorAnd calculating obtains the l covariance matrix corresponding to range gateWherein, WithRespectively K_t×K_sDimension matrix.

Specifically, ifOr i < β, makes i add 1, return sub-step 2.3；

IfOr i >=β, then iteration terminates, and now obtains steering vector v (f during plane up-sampling point (1,1) place when non-working side battle array radar initializes empty empty_d,1,f_s,1Plane up-sampling point (K when)-non-working side battle array radar initializes empty_t,K_s) place empty time steering vectorAnd calculating obtains the l covariance matrix corresponding to range gateWherein, set according to the convergence rate of the power spectrum l1 norm difference of adjacent twice iteration, WithRespectively K_t×K_sDimension matrix, l ∈ 1,2 ..., L}, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

The covariance matrix that the l range gate is correspondingExpression formula is:

$\hat{R}=\&lsqb;v(f_{d,1},f_{s,1}),v(f_{d,2},f_{s,1}),...,v(f_{d,m},f_{s,n}),...,v(f_{d,K_t},f_{s,K_s})\&rsqb;{\hat{P}}^{\left(final\right)}{\&lsqb;v(f_{d,1},f_{s,1}),v(f_{d,2},f_{s,1}),...,v(f_{d,m},f_{s,n}),...,v(f_{d,K_t},f_{s,K_s})\&rsqb;}^H$

Wherein, v (f_d,m,f_s,n) represent when non-working side battle array radar initializes empty in plane any one sampled point to (m, steering vector during n) place empty,The non-working side battle array radar clutter power distribution matrix obtained when representing iteration stopping；ε represents the constant gone to zero determined, β represents default iterations.

Step 3, chooses the covariance matrix that kth range gate is correspondingAs range gate correspondence covariance matrix to be detected, and with described range gate correspondence covariance matrix to be detected for reference, utilize minimum 2-normCalculate the l' range gate clutter data X_l'Transformation matrix T_l'；Wherein, k ∈ 1,2 ..., L}, l' ∈ 1,2 ..., L} and l' ≠ k, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

Specifically, the l' range gate clutter data X_l'Transformation matrix T_l', its expression formula is:

$T_{l^{\&prime;}}=V_k{\mathrm{\&Lambda;}}_k^{1/2}{\mathrm{\&Lambda;}}_{l^{\&prime;}}^{-1/2}{V}_{l^{\&prime;}}^H$

Wherein, V_kRepresentCharacteristic vector, Λ_kRepresentEigenvalue matrix, V_l'RepresentCharacteristic vector, Λ_l'RepresentEigenvalue matrix,Represent the covariance matrix that the l range gate is corresponding,Represent the covariance matrix that the l' range gate is corresponding,The covariance matrix that expression kth range gate is corresponding, l ∈ 1,2 ..., L}, k ∈ 1,2 ..., L}, l' ∈ 1,2 ..., L} and l' ≠ k, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

Step 4, utilizes the l' range gate clutter data X_l'Transformation matrix T_l', to the l' range gate clutter data X_l'Carry out registration process, obtain the l' range gate clutter data Y after registration_l'；Wherein, l' ∈ 1,2 ..., L}, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

Specifically, the l' range gate clutter data Y after described registration_l', its expression formula is: Y_l'=T_l'X_l'；Wherein, T_l'Represent the l' range gate clutter data X_l'Transformation matrix, l' ∈ 1,2 ..., L}, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

Step 5, chooses kth range gate clutter data X_kAs reference spurs, and according to the l' range gate clutter data Y after registration_l', calculate the covariance matrix R that the kth range gate data after obtaining registration are corresponding_k；Wherein, k ∈ 1,2 ..., L}, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

Specifically, the covariance matrix R that kth range gate after described registration is corresponding_k, its expression formula is:

$R_k=(\underset{l^{\&prime;}=1,l^{\&prime;}\&NotEqual;k}{\overset{L}{\&Sigma;}}{Y}_{l^{\&prime;}}{Y}_{l^{\&prime;}}^H+X_k{X}_k^H)/L$

Wherein, Y_l'Represent the l' range gate clutter data after registration, X_kExpression kth range gate clutter data, k ∈ 1,2 ..., L}, l' ∈ 1,2 ..., L} and l' ≠ k, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises, ()^HRepresent conjugate transpose.

Calculate the covariance matrix R that the kth range gate after obtaining registration is corresponding_k, it is possible to carry out power training and process for space-time adaptive radar clutter suppression.

Step 6, according to the covariance matrix R that the kth range gate after registration is corresponding_k, calculate the non-working side battle array radar clutter after space-time adaptive Filtering Processing and suppress improvement factor IF (f_dt), the non-working side battle array radar clutter after described space-time adaptive Filtering Processing suppresses improvement factor IF (f_dt) for weighing the registration effect of non-working side battle array radar clutter spectrum；Wherein, k ∈ 1,2 ..., L}, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

Specifically, according to the covariance matrix R that kth range gate after described registration is corresponding_k, calculate the non-working side battle array radar clutter after space-time adaptive Filtering Processing and suppress improvement factor IF (f_dt), the non-working side battle array radar clutter after described space-time adaptive Filtering Processing suppresses improvement factor IF (f_dt) for weighing the registration effect of non-working side battle array radar clutter spectrum.

Non-working side battle array clutter recognition improvement factor IF (f after described space-time adaptive Filtering Processing_dt), its expression formula is:

$IF\left(f_{dt}\right)=\frac{P_C+P_N}{v{(f_{dt},f_{st})}^H{R}_k^{-1}v(f_{dt},f_{st})}$

Wherein, v (f_dt,f_st) represent steering vector during the moving-target sky comprised in the clutter that non-working side battle array radar receives, f_dtRepresent the moving-target normalization Doppler frequency comprised in the clutter that non-working side battle array radar receives, f_stRepresent the moving-target normalization spatial frequency comprised in the clutter that non-working side battle array radar receives, P_CRepresent the non-working side battle array radar clutter input power arranged, P_NRepresent the non-working side battle array radar noise input power arranged.

The effect of the present invention can be further illustrated by following emulation experiment.

(1) emulation experiment data explanation

In order to suppress non-working side battle array radar clutter performance to contrast with traditional RBC method, method of the present invention emulation adopts even linear array；Selecting suitable repetition, be left out range ambiguity problem, the miscellaneous noise ratio of non-working side battle array radar array element level is 30dB, and No. 80 distance unit (19km) is processed.Traditional RBC method is respectively adopted time domain and spatial domain sub-aperture smoothing technique obtains multiple samples, the spatially and temporally sub-aperture respectively N in method of the present invention emulation_s=8 and M_t=6, the loss that sub-aperture is brought is 10*log10 (MN/M_tN_s), non-working side battle array radar simulation parameter is as shown in table 1.

Table 1

(2) simulation result and analysis

The simulation result of the present invention is shown in Fig. 2 (a)～Fig. 2 (d) and Fig. 3；Wherein, Fig. 2 (a) does not compensate, for the lower No. 80 distance unit of positive side-looking battle array radar, the clutter spectrum schematic diagram that clutter spectrum obtains, the clutter spectrum schematic diagram that Fig. 2 (b) uses RBC method to obtain for the lower No. 80 distance unit of positive side-looking battle array radar, the clutter spectrum schematic diagram that Fig. 2 (c) uses IAA-RBC method to obtain for the lower No. 80 distance unit of positive side-looking battle array radar, Fig. 2 (d) is the lower No. 80 distance real clutter spectrum schematic diagrams of unit of positive side-looking battle array radar, Fig. 3 is IAA-RBC method clutter recognition improvement factor and traditional based on clutter spectrum registration (Registration-basedcompensation, RBC) method clutter recognition improvement factor comparison diagram.

Owing under non-homogeneous scene, sample number is not enough, uncompensated radar clutter spectrum secondary lobe is high, differentiate rate variance；Use RBC method and use IAA-RBC method can obtain the frequency spectrum of high-resolution, but RBC method reduces dimension owing to time domain is smooth, so that the precision of spectrum estimation is restricted, for instance clutter spectrum broadening etc.；The IAA-RBC method that the inventive method uses reconstructs covariance matrix by alternative manner, is absent from time domain and smooths the time domain loss caused, and the precision of Power estimation is higher.

In sum, emulation experiment demonstrates the correctness of the present invention, validity and reliability.

Obviously, the present invention can be carried out various change and modification without deviating from the spirit and scope of the present invention by those skilled in the art；So, if these amendments of the present invention and modification belong within the scope of the claims in the present invention and equivalent technologies thereof, then the present invention is also intended to comprise these change and modification.

Claims (8)

1. the optimization method of a non-working side battle array radar clutter spectrum registration, it is characterised in that comprise the following steps:

Step 1, obtains L the range gate clutter data that non-working side battle array radar receives, is designated as X₁,X₂,X₃…X_L；Wherein, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises；

Step 2, calculates each self-corresponding covariance matrix of L range gate clutter data that non-working side battle array radar receives respectively, is designated as

Step 3, chooses the covariance matrix that kth range gate is correspondingAs range gate correspondence covariance matrix to be detected, and with described range gate correspondence covariance matrix to be detected for reference, calculate the l' range gate clutter data X_l'Transformation matrix T_l'；Wherein, k ∈ 1,2 ..., L}, l' ∈ 1,2 ..., L} and l' ≠ k；

Step 4, utilizes the l' range gate clutter data X_l'Transformation matrix T_l', to the l' range gate clutter data X_l'Carry out registration process, obtain the l' range gate clutter data Y after registration_l'；

Step 5, chooses kth range gate clutter data X_kAs reference spurs, and according to the l' range gate clutter data Y after registration_l', calculate the covariance matrix R that the kth range gate after obtaining registration is corresponding_k；

Step 6, according to the covariance matrix R that the kth range gate after registration is corresponding_k, calculate the non-working side battle array radar clutter after space-time adaptive Filtering Processing and suppress improvement factor IF (f_dt), the non-working side battle array radar clutter after described space-time adaptive Filtering Processing suppresses improvement factor IF (f_dt) for weighing the registration effect of non-working side battle array radar clutter spectrum；Wherein, k ∈ 1,2 ..., L}, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

2. the optimization method of a kind of non-working side battle array radar clutter as claimed in claim 1 spectrum registration, it is characterised in that in step 1, described L range gate clutter data respectively MN × 1 dimension column vector；Wherein, M represents the total number of pulse that non-working side battle array radar receives, and N represents the reception total number of passage antenna of non-working side battle array radar.

3. the optimization method of a kind of non-working side battle array radar clutter as claimed in claim 1 spectrum registration, it is characterised in that in step 2, each self-corresponding covariance matrix of L range gate clutter data that described non-working side battle array radar receives, also include；

Calculate the l covariance matrix corresponding to range gateIts sub-step is:

The 2.1 interval normalization Doppler frequency intervals with non-working side battle array radar of normalization spatial frequency setting non-working side battle array radar respectively, and respectively described normalization spatial frequency interval is carried out uniform sampling, it is thus achieved that K_sIndividual sampled point, carries out uniform sampling to described normalization Doppler frequency interval, it is thus achieved that K_tIndividual sampled point, then non-working side battle array radar normalization Doppler frequency interval that the is non-working side battle array radar normalization spatial frequency of setting is interval and that set is built into plane when non-working side battle array radar initializes empty, calculating obtains when non-working side battle array radar initializes empty any one sampled point (m, n) the clutter power initial value at place in plane

Wherein, n ∈ 1,2 ..., K_s, m ∈ 1,2 ..., K_t, l ∈ 1,2 ..., L}, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises, K_sRepresent the sampled point number that the non-working side battle array radar normalization spatial frequency interval set comprises, K_tRepresent the sampled point number that the non-working side battle array radar normalization Doppler frequency interval set comprises；

2.2 initialize: β represents default iterations, and i represents iterations, and initial value is 1；

2.3 calculate successively and obtain when the non-working side battle array radar after ith iteration initializes empty any one sampled point (m, n) covariance matrix at place in planeAny one sampled point (m, n) the clutter power value at place in plane when initializing empty with the non-working side battle array radar after ith iteration

If 2.4Or i < β, makes i add 1, return sub-step 2.3；

IfOr i >=β, then iteration terminates, and now obtains steering vector v (f during plane up-sampling point (1,1) place when non-working side battle array radar initializes empty empty_d,1,f_s,1Plane up-sampling point (K when)-non-working side battle array radar initializes empty_t,K_s) place empty time steering vectorAnd calculating obtains the l covariance matrix corresponding to range gateWherein, WithRespectively K_t×K_sDimension matrix, ε represents the constant gone to zero determined, β represents default iterations,RepresentNorm.

4. the optimization method of a kind of non-working side battle array radar clutter as claimed in claim 3 spectrum registration, it is characterised in that any one sampled point (m, n) the clutter power initial value at place in plane when described non-working side battle array radar initializes emptyAny one sampled point (m, n) covariance matrix at place in plane when non-working side battle array radar after described ith iteration initializes emptyAny one sampled point (m, n) the clutter power value at place in plane when non-working side battle array radar after described ith iteration initializes emptyThe covariance matrix corresponding with described the l range gateIts expression formula is respectively as follows:

${\hat{P}}_{m,n}^{\left(0\right)}=|\frac{v^H\left(f_{d,m},f_{s,n}\right)X_l}{v^H\left(f_{d,m},f_{s,n}\right)v\left(f_{d,m},f_{s,n}\right)}{|}^2,m\&Element;\left\{1,2,...,K_t\right\},n\&Element;\left\{1,2,...,K_s\right\}$

${\hat{R}}_{m,n}^{\left(i\right)}=\underset{u=1,u\&NotEqual;m}{\overset{K_t}{\&Sigma;}}\underset{v=1,v\&NotEqual;n}{\overset{K_s}{\&Sigma;}}{\hat{P}}_{u,v}^{\left(i\right)}v(f_{d,u},f_{s,v})v^H(f_{d,u},f_{s,v})$

${\hat{P}}_{m,n}^{\left(i\right)}=|\frac{v^H\left(f_{d,m},f_{s,n}\right){\left({\hat{R}}_{m,n}^{\left(i\right)}\right)}^{-1}{X}_l}{v^H\left(f_{d,m},f_{s,n}\right){\left({\hat{R}}_{m,n}^{\left(i\right)}\right)}^{-1}v\left(f_{d,m},f_{s,n}\right)}{|}^2$

${\hat{R}}_l=\&lsqb;v\left(f_{d,1},f_{s,1}\right),v\left(f_{d,2},f_{s,1}\right),\mathrm{..};v\left(f_{d,m},f_{s,n}\right),\mathrm{..};v\left(f_{d,K_t},f_{s,K_s}\right)\&rsqb;{\hat{P}}^{\left(final\right)}{\&lsqb;v\left(f_{d,1},f_{s,1}\right),v\left(f_{d,2},f_{s,1}\right),\mathrm{..};v\left(f_{d,m},f_{s,n}\right),\mathrm{..};v\left(f_{d,K_t},f_{s,K_s}\right)\&rsqb;}^H$

Wherein, v (f_d,m,f_s,n) represent when non-working side battle array radar initializes empty any one sampled point in plane (m, steering vector during n) place empty, f_s,nRepresent the normalization spatial frequency of the n-th sampled point, f on the non-working side battle array radar normalization spatial frequency interval set_d,mRepresent the normalization Doppler frequency of m-th sampled point on the non-working side battle array radar normalization Doppler frequency interval set,Represent that kronecker amasss, ()^HRepresent conjugate transpose, ()^TRepresent transposition, | | represent the operation that takes absolute value,Represent when the non-working side battle array radar after ith iteration initializes empty any one sampled point (u, v) the clutter power value at place, v (f in plane_d,u,f_s,v) represent plane up-sampling point when non-working side battle array radar initializes empty (u, steering vector during v) place empty, f_d,uRepresent the normalization Doppler frequency of u sampled point, f on the non-working side battle array radar normalization Doppler frequency interval set_s,vRepresent the normalization spatial frequency of v sampled point on the non-working side battle array radar normalization spatial frequency interval set,Representing the non-working side battle array radar clutter power distribution matrix that obtains during iteration stopping, ε represents the constant gone to zero determined, β represents default iterations, u ∈ 1,2 ..., K_t, v ∈ 1,2 ..., K_s, n ∈ 1,2 ..., K_s, m ∈ 1,2 ..., K_t, u ≠ m, v ≠ n, K_sRepresent the sampled point number that the non-working side battle array radar normalization spatial frequency interval set comprises, K_tRepresent the sampled point number that the non-working side battle array radar normalization Doppler frequency interval set comprises.

5. the optimization method of a kind of non-working side battle array radar clutter as claimed in claim 1 spectrum registration, it is characterised in that in step 3, described the l' range gate clutter data X_l'Transformation matrix T_l', its expression formula is:

$T_{l^{\&prime;}}=V_k{\mathrm{\&Lambda;}}_k^{1/2}{\mathrm{\&Lambda;}}_{l^{\&prime;}}^{-1/2}{V}_{l^{\&prime;}}^H$

Wherein, V_kRepresentCharacteristic vector, Λ_kRepresentEigenvalue matrix, V_l'RepresentCharacteristic vector, Λ_l'RepresentEigenvalue matrix,Represent the covariance matrix that the l range gate is corresponding,Represent the covariance matrix that the l' range gate is corresponding,The covariance matrix that expression kth range gate is corresponding, l ∈ 1,2 ..., L}, k ∈ 1,2 ..., L}, l' ∈ 1,2 ..., L} and l' ≠ k, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

6. the optimization method of a kind of non-working side battle array radar clutter as claimed in claim 1 spectrum registration, it is characterised in that in step 4, the l' range gate clutter data Y after described registration_l', its expression formula is: Y_l'=T_l'X_l'；Wherein, T_l'Represent the l' range gate clutter data X_l'Transformation matrix, l' ∈ 1,2 ..., L}, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises.

7. the optimization method of a kind of non-working side battle array radar clutter as claimed in claim 1 spectrum registration, it is characterised in that in steps of 5, the covariance matrix R that kth range gate after described registration is corresponding_k, its expression formula is:

$R_k=(\underset{l^{\&prime;}=1,l^{\&prime;}\&NotEqual;k}{\overset{L}{\mathrm{\&Sigma;}}}{Y}_{l^{\&prime;}}{Y}_{l^{\&prime;}}^H+X_k{X}_k^H)/L$

Wherein, Y_l'Represent the l' range gate clutter data after registration, X_kExpression kth range gate clutter data, k ∈ 1,2 ..., L}, l' ∈ 1,2 ..., L} and l' ≠ k, L represents the total number of range gate that the clutter that non-working side battle array radar receives comprises, ()^HRepresent conjugate transpose.

8. the optimization method of a kind of non-working side battle array radar clutter as claimed in claim 1 spectrum registration, it is characterised in that in step 6, the non-working side battle array radar clutter after described space-time adaptive Filtering Processing suppresses improvement factor IF (f_dt), its expression formula is:

$IF\left(f_{dt}\right)=\frac{P_C+P_N}{v{(f_{dt},f_{st})}^H{R}_k^{-1}v(f_{dt},f_{st})}$

Wherein, v (f_dt,f_st) represent steering vector during the moving-target sky comprised in the clutter that non-working side battle array radar receives, f_dtRepresent the moving-target normalization Doppler frequency comprised in the clutter that non-working side battle array radar receives, f_stRepresent the moving-target normalization spatial frequency comprised in the clutter that non-working side battle array radar receives, P_CRepresent the non-working side battle array radar clutter input power arranged, P_NRepresent the non-working side battle array radar noise input power arranged.