CN106970358A - The optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum - Google Patents

Abstract

The invention belongs to Radar Technology field, a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum is disclosed, including：Obtain each range gate clutter data that non-working side battle array radar is received, determine the Spatial Doppler frequency location where each training sample unit maximum power point, determine clutter spectrum center, reference is used as using the Doppler frequency and spatial frequency of range cell to be detected, calculate the corresponding correction matrix of clutter data of each training sample unit, calculate the clutter covariance matrix of the range cell to be detected after correcting, so as to obtain filtering weight vector, clutter data to non-working side battle array radar carries out clutter recognition, so that its clutter is more uniform after non-working side battle array radar clutter spectrum processing.

Description

The optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum

Technical field

Match somebody with somebody the invention belongs to the angle Doppler in Radar Technology field, more particularly to a kind of non-working side battle array radar clutter spectrum Accurate optimization method, it is adaptable to when known radar configuration parameter has error, is carried out ideal to radar clutter spectrum Compensation.

Background technology

Airborne radar has the functions such as air alert, scouting, control and guided weapon, is played in China's national defense construction Important function.Airborne radar it is lower regard work when, had a very wide distribution due to ground clutter, intensity it is big, moving-target is covered completely And None- identified.How ground clutter is effectively suppressed, be the important content of airborne radar signal transacting.Space-time adaptive Processing (space-time adaptive processing, STAP) method is to suppress ground clutter, compensating platform motion to cause Clutter spectrum widening, the effective means of detection ground target at a slow speed.Clutter covariance matrix (Clutter Covariance Matrix, CCM) be space-time adaptive processing method clutter recognition performance deciding factor.In actual applications, list to be detected The clutter covariance matrix of member is to utilize and unit independent same distribution (Independent Identically to be detected Distributed, IID) close on training sample unit, inverted (Sample Matrix by covariance of sampling Inversion, SMI) method estimation obtains, to realize adaptively clutter recognition, in order that the output letter of space-time filter Miscellaneous noise ratio loss is less than 3dB, for estimating that the independent identically distributed number of training of clutter covariance matrix need to be more than system certainly By twice spent.For non-working side battle array, ground clutter Doppler frequency has serious distance dependencies, especially in short range bar Under part, distance dependencies are more notable so that clutter distribution is unsatisfactory for independent same distribution condition, and then can not be come with training sample Accurately estimate the clutter covariance matrix of unit to be detected, make the clutter recognition hydraulic performance decline of space-time adaptive processing method.

At present, the method for compensation distance dependencies has a lot, including Doppler's warpage (Doppler Warping, DW) method, Angle Doppler effect correction (Angle Doppler Compensation, ADC) method, adaptive angle Doppler compensation (Adaptive Angle Doppler Compensation, A2DC) method and based on registration compensate (Registration Based Compensation, RBC) etc. method.Wherein, DW methods be substantially by caused by the different angles of pitch training sample with Clutter Doppler frequency differences between detecting distance unit are compensated, so that the clutter in training sample is approximately put down Surely.ADC methods and A2DC methods will also in addition to being compensated to the Doppler differences between training sample and range cell to be detected Spatial frequency variance is compensated caused by the different angles of pitch.RBC methods carry out radar clutter spectrum using the smooth sub- snap of time domain The extraction of peak value, can realize the full remuneration of radar clutter, and can obtain good radar clutter in the ideal case Rejection.

However, in above-mentioned compensation method, DW methods only carry out main-lobe clutter spectrum center compensation, obtained benefit in Doppler domain It is not especially desirable to repay effect.ADC methods by each training sample unit and are treated by the compensation in space angle domain and Doppler domain The noise performance registration of sample unit is detected, clutter transition matrix is calculated, so that clutter reduction distance is relied on to a certain extent Property, for main-lobe clutter, the compensation effect of this method becomes apparent.But, in actual applications, the clutter of each training sample Transition matrix is required for flight configuration parameter (such as angle of pitch, azimuth, yaw angle and the speed provided according to carrier aircraft platform Deng), calculated and obtained using clutter space-time coupled relation, thus the error of arbitrary disposition parameter all can be to the clutter distance of this method Dependence compensation performance affects.Although A2DC methods are not required to know configuration parameter, and radar is miscellaneous in the ideal case Ripple rejection is also fine, but the covariance matrix of each range cell radar clutter is extremely unstable in practice, causes radar Clutter recognition hydraulic performance decline is a lot.RBC methods carry out the extraction of radar clutter spectrum peak using the smooth sub- snap of time domain so that radar The estimated accuracy of clutter spectrum peak and the robustness of RBC methods all can be to being impacted using STAP technical performances.

The content of the invention

In view of the above-mentioned problems of the prior art, it is an object of the invention to provide a kind of non-working side battle array radar clutter The optimization method of the angle Doppler registration of spectrum so that its clutter is more uniform after non-working side battle array radar clutter spectrum processing.

The present invention proposes a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum.This method Estimate that non-working side battle array radar clutter is composed by adaptive iteration method (IAA, Iterative Adaptive Approach), The clutter power approximation of each training sample is obtained, it is determined that the Doppler frequency position where each training sample maximum power point Put and spatial frequency position, determine its clutter spectrum center, in conjunction with angle Doppler effect correction (ADC, Angle Doppler Compensation) method carries out the two dimension compensation of spatial domain and Doppler domain, obtains and unit clutter statistical characteristicses phase to be detected Consistent training sample data so that its clutter is more uniform after non-working side battle array radar clutter spectrum processing, and then improves sky When self-adaptive processing (STAP) technology non-working side battle array radar application in clutter performance.

To reach above-mentioned purpose, the present invention, which is adopted the following technical scheme that, to be achieved.

A kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum, methods described includes following step Suddenly：

Step 1, the clutter data for the L range cell that non-working side battle array radar is received is obtained, and sets k-th of distance Unit is range cell to be detected, then other L-1 range cell in addition to range cell to be detected are training sample unit；L Represent the range cell total number that the clutter data that non-working side battle array radar is received is included；

Step 2, the normalization spatial frequency interval [- 1,1] and the non-working side of the non-working side battle array radar are set The normalization Doppler frequency of battle array radar is interval [- 1,1]；And non-working side battle array radar normalization spatial frequency it is interval and The normalization Doppler frequency interval of the non-working side battle array radar constitutes the space-time two-dimensional plane of non-working side battle array radar；

Step 3, l=1, l ∈ { 1,2 ..., L }, l ≠ k, l is made to represent l-th of training sample unit；

Step 4, to the clutter data of l-th of training sample unit non-working side battle array radar space-time two-dimensional plane It is upper to be sampled, obtain K_s×K_tIndividual sampled point；K_sRepresent the clutter data of l-th of training sample unit in non-working side battle array thunder The sampled point number in the normalization spatial frequency interval reached, K_tRepresent the clutter data of l-th of training sample unit in non-positive side Depending on the sampled point number that the normalization Doppler frequency of battle array radar is interval；

Step 5, for l-th of training sample unit clutter data non-working side battle array radar space-time two-dimensional plane on (p, q) individual sampled point, obtain (p, q) individual sample point space-time steering vector, p ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}；So as to obtain space-time two-dimensional plane of the clutter data in non-working side battle array radar of l-th of training sample unit Upper all K_s×K_tThe space-time steering vector of individual sample point；

Step 6, according to the clutter data of l-th of training sample unit, the space-time of (p, q) individual sample point is oriented to arrow Amount, calculates the initial power of (p, q) individual sample pointP ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}；And according to every The initial power of individual sample point determines the clutter power spectrum P of the clutter data of l-th of training sample unit⁽⁰⁾；

And set iterations n initial value to be 1；

Step 7, according to the power of (p, q) individual sample pointAnd the space-time of (p, q) individual sample point is oriented to arrow The covariance matrix of amount construction (p, q) individual sample pointP ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}；

Step 8, according to the clutter data of l-th of training sample unit, the space-time steering vector of (p, q) individual sample point And the covariance matrix of (p, q) individual sample pointCalculate the power of (p, q) individual sample point after nth iterationP ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}；And according to the power of each sample point after nth iterationIt is determined that Clutter power spectrum P after the clutter data nth iteration of l-th of training sample unit⁽ⁿ⁾；

Step 9, ifAnd 0_＜N ＜ num, then make n value plus 1, repeats step 7 and step Rapid 8；Num represents the maximum of the iterations of setting,Expression is askedNorm；

IfOr n ＞ num, then stop iteration, and by each sampled point after nth iteration The power at placeIt is used as the final power of each sample point in the clutter data of l-th of training sample unit；

Step 10, the maximum of the final power of each sample point in the clutter data of l-th of training sample unit is determined Value, obtain the corresponding sample point of the maximum normalization Doppler frequency and normalization spatial frequency, using its as The corresponding normalization Doppler frequency of power spectrum and normalization spatial frequency of the clutter data of l-th of training sample unit；

Step 11, the corresponding normalization Doppler frequency of clutter power spectrum of the clutter data of range cell to be detected is determined With normalization spatial frequency；According to the corresponding normalization Doppler frequency of the power spectrum of the clutter data of l-th of training sample unit Rate and normalization spatial frequency, the corresponding normalization Doppler frequency of clutter power spectrum of the clutter data of range cell to be detected With normalization spatial frequency, the corresponding amendment Doppler frequency of power spectrum of the clutter data of l-th of training sample unit is calculated With amendment spatial frequency；

Step 12, according to the corresponding amendment Doppler frequency of power spectrum of the clutter data of l-th of training sample unit Rate and amendment spatial frequency, calculate the power spectrum correspondence Doppler domain transition matrix of the clutter data of l-th of training sample unit With spatial domain transition matrix, and then according to the power spectrum correspondence Doppler domain conversion square of the clutter data of l-th of training sample unit Battle array and spatial domain transition matrix, calculate the power spectrum correspondence correction matrix for the clutter data for obtaining l-th of training sample unit；

According to the power of the clutter data of l-th training sample unit and the clutter data of l-th of training sample unit Spectrum correspondence correction matrix, obtains the clutter data by revised l-th of training sample unit；

Step 13, make l value plus 1, and repeat step 4 to step 12, so as to obtain L-1 by revised miscellaneous Wave number evidence, so that according to the L-1 association that the clutter data for obtaining range cell to be detected is calculated by revised clutter data Variance matrix；

The space-time steering vector of range cell to be detected is obtained, so as to be oriented to according to the space-time of the range cell to be detected The covariance matrix of the clutter data of vector and range cell to be detected, tries to achieve filtering weight vector, according to it is described filtering power to Amount carries out angle Doppler registration to the clutter data of the L range cell respectively.

Beneficial effects of the present invention：The inventive method utilizes IAA methods estimation non-working side battle array in the case of single snap Radar clutter is composed, and is obtained the clutter power approximation of each training sample, is determined its clutter spectrum center, enter in conjunction with ADC methods Row spatial domain and the two dimension compensation of Doppler domain so that its clutter is more uniform after non-working side battle array radar clutter spectrum processing, and then Improve clutter performance of space-time adaptive processing (STAP) technology in non-working side battle array radar.

Brief description of the drawings

In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing There is the accompanying drawing used required in technology description to be briefly described, it should be apparent that, drawings in the following description are only this Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can be with Other accompanying drawings are obtained according to these accompanying drawings.

Fig. 1 is a kind of optimization of the angle Doppler registration of non-working side battle array radar clutter spectrum provided in an embodiment of the present invention The schematic flow sheet of method；

Fig. 2 is clutter spectrum schematic diagram optimal under non-working side battle array radar；

Fig. 3 is not compensated obtained clutter spectrum schematic diagram under non-working side battle array radar；

Fig. 4 is the clutter spectrum schematic diagram that is obtained using existing A2DC methods under non-working side battle array radar；

Fig. 5 is the clutter spectrum schematic diagram that is obtained using the inventive method under non-working side battle array radar；

Fig. 6 is non-working side battle array radar clutter suppression improvement factor contrast in the case of four kinds shown in Fig. 2, Fig. 3, Fig. 4, Fig. 5 Schematic diagram.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiments.It is based on Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made Embodiment, belongs to the scope of protection of the invention.

The embodiment of the present invention provides a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum, such as Shown in Fig. 1, methods described comprises the following steps：

Step 1, the clutter data for the L range cell that non-working side battle array radar is received is obtained, and sets k-th of distance Unit is range cell to be detected, then other L-1 range cell in addition to range cell to be detected are training sample unit；L Represent the range cell total number that the clutter data that non-working side battle array radar is received is included.

General, when including target data in the clutter data for the L range cell that non-working side battle array radar is received, then It regard the range cell where target as range cell to be detected.

In step 1, the clutter data for the L range cell that non-working side battle array radar is received is obtained, is specially：

The linear array for setting the antenna of the non-working side battle array radar to be made up of equally distributed N number of array element, and institute State the M pulse that non-working side battle array radar is launched in a coherent processing inteval；

L times is carried out respectively on each pulse, each array element passage apart from upper sampling, then the clutter number being an actually-received According to being three-dimensional data that N × M × L is；The clutter data that m-th of reception of impulse of n-th of array element of l-th of range cell is arrived It is designated as x_nml, then the clutter data that n-th of array element of l-th of range cell is received is designated as x_nl=[x_n1l, x_n2l..., x_nMl]^T, The clutter data that N number of array element of l-th of range cell is received lines up the column vector X that NM × 1 is tieed up_l, so that obtain l-th away from From the clutter data Xl=[x of unit_1l ^T, x_2l ^T..., x_Nl ^T]^T, wherein, l ∈ { 1,2 ..., L }.

General, the number for choosing training sample unit is more than 2NM.

Step 2, the normalization spatial frequency interval [- 1,1] and the non-working side of the non-working side battle array radar are set The normalization Doppler frequency of battle array radar is interval [- 1,1]；And non-working side battle array radar normalization spatial frequency it is interval and The normalization Doppler frequency interval of the non-working side battle array radar constitutes the space-time two-dimensional plane of non-working side battle array radar.

The clutter data of a certain range cell is the data on array number, a two dimensional surface of umber of pulse structure, is The echo data got of most original；Space-time two-dimensional plane is one built by the angle after normalization, Doppler frequency Individual two dimensional surface, angle and Doppler frequency are calculated according to known configuration parameter.

Step 3, l=1, l ∈ { 1,2 ..., L }, l ≠ k, l is made to represent l-th of training sample unit.

Step 4, to the clutter data of l-th of training sample unit non-working side battle array radar space-time two-dimensional plane It is upper to be sampled, obtain K_s×K_tIndividual sampled point；K_sRepresent the clutter data of l-th of training sample unit in non-working side battle array thunder The sampled point number in the normalization spatial frequency interval reached, K_tRepresent the clutter data of l-th of training sample unit in non-positive side Depending on the sampled point number that the normalization Doppler frequency of battle array radar is interval.

In step 4, K_s=10N, k_t=10M；

Wherein, N represents the element number of array that the antenna of non-working side battle array radar is included, and M represents non-working side battle array radar one The pulse number of transmitting, K in individual coherent processing inteval_sRepresent the clutter data of l-th of training sample unit in non-working side battle array The interval sampled point number of the normalization spatial frequency of radar, K_tRepresent the clutter data of l-th of training sample unit in anon-normal The sampled point number in the normalization Doppler frequency interval of side view battle array radar.

For more accurate reconstruct covariance matrix, the spatial domain of space-time plane is divided into K respectively here_sIndividual grid Point, Doppler domain is divided into K_tIndividual mesh point, then the normalization spatial frequency corresponding to each mesh point can be expressed as f_{S, q}, q ∈ { 1,2 ..., K_s, normalization Doppler frequency can be expressed as f_{D, p}, p ∈ { 1,2 ..., K_t, according to carat Metro lower bound reason By unlimited increase grid distribution number will not be obviously improved the performance of Power estimation, be that choose K here_s=10N, k_t=10M, Now space-time plane can be divided into K=K_sK_tIndividual mesh point, i.e., 100NM mesh point, therefore the clutter data of current training sample It can be represented with the angle doppler data at the Grid Sampling point.The normalization spatial frequency interval [- 1,1] is carried out Uniform sampling, obtains K_sIndividual sampled point；Uniform sampling is carried out to the normalization Doppler frequency interval [- 1,1], K is obtained_tIt is individual Sampled point.

Step 5, for l-th of training sample unit clutter data non-working side battle array radar space-time two-dimensional plane on (p, q) individual sampled point, obtain (p, q) individual sample point space-time steering vector, p ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}；So as to obtain space-time two-dimensional plane of the clutter data in non-working side battle array radar of l-th of training sample unit Upper all K_s×K_tThe space-time steering vector of individual sample point.

In step 5, the space-time steering vector of (p, q) individual sample point is obtained, is specially：

Obtain steering vector of (p, q) the individual sampled point on the normalization spatial frequency interval of non-working side battle array radar

Obtain steering vector of (p, q) the individual sampled point on the normalization Doppler frequency interval of non-working side battle array radar

So as to the space-time steering vector of (p, q) individual sample point

Wherein, f_{D, p}Represent the normalization of interval upper p-th of the sampled point of normalization Doppler frequency of non-working side battle array radar Doppler frequency, f_{S, q}Represent the normalization space of interval upper q-th of the sampled point of normalization spatial frequency of non-working side battle array radar Frequency；Represent kronecker products, ()^TRepresent transposition.

Step 6, according to the clutter data of l-th of training sample unit, the space-time of (p, q) individual sample point is oriented to arrow Amount, calculates the initial power of (p, q) individual sample pointP ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}；And according to each The initial power of sample point determines the clutter power spectrum P of the clutter data of l-th of training sample unit⁽⁰⁾；

And set iterations n initial value to be 1.

In step 6, the initial power of (p, q) individual sample point is calculated using following formula

Wherein, S (f_{D, p}, f_{S, q}) represent (p, q) individual sample point space-time steering vector, X_lRepresent l-th of training sample The clutter data of this unit, ()^HConjugate transposition is represented, | | represent the operation that takes absolute value；

The clutter power spectrum of the clutter data of l-th of training sample unit is determined according to the initial power of each sample point P⁽⁰⁾ForDiag () represents that diagonal matrix, i.e. diagonal element areDiagonal matrix.

Step 7, according to the power of (p, q) individual sample pointAnd the space-time of (p, q) individual sample point is oriented to arrow The covariance matrix of amount construction (p, q) individual sample pointP ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}。

In step 7, according to the power of (p, q) individual sample pointAnd the space-time of (p, q) individual sample point is oriented to The covariance matrix of the individual sample point of vectorial structure (p, q)It is as follows：

Wherein, S (f_{D, p}, f_{S, q}) represent (p, q) individual sample point space-time steering vector.

Step 8, according to the clutter data of l-th of training sample unit, the space-time steering vector of (p, q) individual sample point And the covariance matrix of (p, q) individual sample pointCalculate the power of (p, q) individual sample point after nth iterationP ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}；And according to the power of each sample point after nth iterationIt is determined that Clutter power spectrum P after the clutter data nth iteration of l-th of training sample unit⁽ⁿ⁾。

In step 8, according to the clutter data of l-th of training sample unit, the space-time of (p, q) individual sample point is oriented to arrow The covariance matrix of amount and (p, q) individual sample pointCalculate the work(of (p, q) individual sample point after nth iteration RateIt is as follows：

Wherein, S (f_{D, p}, f_{S, q}) represent (p, q) individual sample point space-time steering vector, X_lRepresent l-th of training sample The clutter data of this unit, ()^HConjugate transposition is represented, | | represent take absolute value operation, ()^-1Representing matrix is inverted；

According to the power of each sample point after nth iterationDetermine the clutter data of l-th of training sample unit Clutter power spectrum P after n iteration⁽ⁿ⁾It is as follows：

Wherein, diag () represents that diagonal matrix, i.e. diagonal element areDiagonal matrix.

Step 9, ifAnd 0_＜N ＜ num, then make n value plus 1, repeats step 7 and step Rapid 8；Num represents the maximum of the iterations of setting,Expression is askedNorm；

IfOr n ＞ num, then stop iteration, and by each sampled point after nth iteration The power at placeIt is used as the final power of each sample point in the clutter data of l-th of training sample unit.

Step 10, the maximum of the final power of each sample point in the clutter data of l-th of training sample unit is determined Value, obtain the corresponding sample point of the maximum normalization Doppler frequency and normalization spatial frequency, using its as The corresponding normalization Doppler frequency of power spectrum and normalization spatial frequency of the clutter data of l-th of training sample unit.

Step 11, the corresponding normalization Doppler frequency of clutter power spectrum of the clutter data of range cell to be detected is determined With normalization spatial frequency；According to the corresponding normalization Doppler frequency of the power spectrum of the clutter data of l-th of training sample unit Rate and normalization spatial frequency, the corresponding normalization Doppler frequency of clutter power spectrum of the clutter data of range cell to be detected With normalization spatial frequency, the corresponding amendment Doppler frequency of power spectrum of the clutter data of l-th of training sample unit is calculated With amendment spatial frequency.

In step 11：Determine the corresponding normalization Doppler frequency of clutter power spectrum of the clutter data of range cell to be detected Rate f_{D, k}With normalization spatial frequency f_{S, k}：

Wherein, λ is the operation wavelength of non-working side battle array radar,For the array element spacing of even linear array, carrier aircraft is with speed V flies along x-axis, and the angle of the plane vertical with x-y plane and x-axis where even linear array isθ_kRepresent distance to be detected The angle of pitch corresponding to unit clutter data,Represent range cell clutter data to be detected relative to corresponding to even linear array Azimuth, array number is N, and pulse recurrence frequency is f_r；

Calculate the corresponding amendment Doppler frequency Δ f of power spectrum of the clutter data of l-th of training sample unit_{D, l}= f_{D, l}-f_{D, k}With amendment spatial frequency Δ f_{S, l}=f_{S, l}-f_{S, k}；

Wherein, f_{D, l}Represent the corresponding normalization Doppler frequency of power spectrum of the clutter data of l-th of training sample unit Rate, f_{S, l}Represent the corresponding normalization spatial frequency of power spectrum of the clutter data of l-th of training sample unit, f_{D, k}Represent to be checked Survey the corresponding normalization Doppler frequency of clutter power spectrum of the clutter data of range cell, f_{S, k}Represent range cell to be detected Clutter data the corresponding normalization spatial frequency of clutter power spectrum.

Step 12, according to the corresponding amendment Doppler frequency of power spectrum of the clutter data of l-th of training sample unit Rate and amendment spatial frequency, calculate the power spectrum correspondence Doppler domain transition matrix of the clutter data of l-th of training sample unit With spatial domain transition matrix, and then according to the power spectrum correspondence Doppler domain conversion square of the clutter data of l-th of training sample unit Battle array and spatial domain transition matrix, calculate the power spectrum correspondence correction matrix for the clutter data for obtaining l-th of training sample unit；

According to the power of the clutter data of l-th training sample unit and the clutter data of l-th of training sample unit Spectrum correspondence correction matrix, obtains the clutter data by revised l-th of training sample unit.

In step 12, the power spectrum correspondence Doppler domain transition matrix of the clutter data of l-th of training sample unit is calculated T_{D, l}With spatial domain transition matrix T_{S, l}：

Calculate the power spectrum correspondence correction matrix T for the clutter data for obtaining l-th of training sample unit_IAA-ADC：

According to the power of the clutter data of l-th training sample unit and the clutter data of l-th of training sample unit Spectrum correspondence correction matrix, obtains the clutter data by revised l-th of training sample unit

Wherein, Δ f_{D, l}Represent the corresponding amendment Doppler frequency of power spectrum of the clutter data of l-th of training sample unit Rate, Δ f_{S, l}The corresponding amendment spatial frequency of power spectrum of the clutter data of l-th of training sample unit is represented,Represent Kronecker is accumulated, and N represents the element number of array that the antenna of non-working side battle array radar is included, and M represents non-working side battle array radar at one The pulse number of transmitting in coherent processing inteval.

Step 13, make l value plus 1, and repeat step 4 to step 12, so as to obtain L-1 by revised miscellaneous Wave number evidence, so that according to the L-1 association that the clutter data for obtaining range cell to be detected is calculated by revised clutter data Variance matrix；

The space-time steering vector of range cell to be detected is obtained, so as to be oriented to according to the space-time of the range cell to be detected The covariance matrix of the clutter data of vector and range cell to be detected, tries to achieve filtering weight vector, according to it is described filtering power to Amount carries out angle Doppler registration to the clutter data of the L range cell respectively.

In step 13, the clutter number for obtaining range cell to be detected is calculated by revised clutter data according to L-1 According to covariance matrix

Wherein,Represent the clutter data by revised l-th of training sample unit；

Obtain the space-time steering vector S (f of range cell to be detected_{D, k}, f_{S, k}), according to the sky of the range cell to be detected When steering vector S (f_{D, k}, f_{S, k}) and range cell to be detected clutter data covariance matrixTry to achieve filtering weight vectorI.e.

Wherein,For normaliztion constant.

Finally, the clutter data of L range cell non-working side battle array radar received is weighed by the filtering respectively Vector carries out clutter recognition, obtains uniform clutter spectrum.

You need to add is that, calculate non-working side battle array radar clutter and suppress improvement factor IF (f_dt), for weighing non-positive side The angle Doppler effect correction effect composed depending on battle array radar clutter.

Specifically, its expression formula is：

Wherein, P_CRepresent the non-working side battle array radar clutter input power set, P_NRepresent the non-working side battle array radar set Noise inputs power.

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

(1) emulation experiment data explanation

When being estimated to the rejection of non-working side battle array radar clutter, in order to many with traditional adaptive angle General Le compensation method is contrasted, and the inventive method emulation uses even linear array；The appropriate repetition of selection, does not consider range ambiguity Problem, the miscellaneous noise ratio of non-working side battle array radar array element level is 60dB, and No. 60 range cell is handled.This part passes through imitative Sampling covariance inversion technique very not compensated, after angle Doppler Compensation Method is compensated and the inventive method is compensated is estimated Clutter power spectrum, and contrasted with optimal clutter power spectrum.The clutter power spectrum of distinct methods uses minimum variance Undistorted response (Minimum Variance Distortionless Response, MVDR) spectrum.Non-working side battle array radar is imitated True parameter is as shown in table 1.

Table 1

(2) simulation result and analysis

The simulation result of the present invention is as shown in Fig. 2~Fig. 6；Wherein, Fig. 2 is non-working side battle array lower No. 60 range cells of radar Optimal clutter spectrum schematic diagram, Fig. 3 is the not compensated obtained clutter spectrum signal of non-working side battle array lower No. 60 range cells of radar Figure, Fig. 4 is the clutter spectrum schematic diagram that non-working side battle array lower No. 60 range cells of radar are obtained using A2DC methods, and Fig. 5 is non-positive side The clutter spectrum schematic diagram obtained depending on lower No. 60 range cells of battle array radar using IAA-ADC；Fig. 6 is uncompensated, A2DC methods, IAA- ADC method clutter recognition improvement factors and optimal clutter recognition improvement factor comparison diagram.

Because sample number is not enough under non-homogeneous scene, uncompensated radar clutter spectrum secondary lobe is high, resolution ratio is poor；Use A2DC Method and the frequency spectrum of high-resolution can be obtained using IAA-ADC methods.But A2DC methods each range cell radar in practice The covariance matrix of clutter is extremely unstable, causes radar clutter rejection to decline a lot；The IAA-ADC methods that the present invention is used are led to Cross alternative manner reconstruct covariance matrix so that more uniform, the performance of Power estimation after non-working side battle array radar clutter spectrum processing Improve.

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

One of ordinary skill in the art will appreciate that：Realizing all or part of step of above method embodiment can pass through Programmed instruction related hardware is completed, and foregoing program can be stored in computer read/write memory medium, and the program exists During execution, the step of execution includes above method embodiment；And foregoing storage medium includes：ROM, RAM, magnetic disc or CD Etc. it is various can be with the medium of store program codes.

The foregoing is only a specific embodiment of the invention, but protection scope of the present invention is not limited thereto, any Those familiar with the art the invention discloses technical scope in, change or replacement can be readily occurred in, should all be contained Cover within protection scope of the present invention.Therefore, protection scope of the present invention should be based on the protection scope of the described claims.

Claims (10)

1. a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum, it is characterised in that methods described Comprise the following steps：

Step 1, the clutter data for the L range cell that non-working side battle array radar is received is obtained, and sets k-th of range cell For range cell to be detected, other L-1 range cell in addition to range cell to be detected are set as training sample unit；L tables Show the range cell total number that the clutter data that non-working side battle array radar is received is included；

Step 2, the normalizing of the normalization spatial frequency interval [- 1,1] of setting non-working side battle array radar and non-working side battle array radar Change Doppler frequency interval [- 1,1]；And the interval and described non-working side of normalization spatial frequency of the non-working side battle array radar The normalization Doppler frequency interval of battle array radar constitutes the space-time two-dimensional plane of non-working side battle array radar；

Step 3, l=1, l ∈ { 1,2 ..., L }, l ≠ k, l is made to represent l-th of training sample unit；

Step 4, the clutter data to l-th of training sample unit is enterprising in the space-time two-dimensional plane of non-working side battle array radar Row sampling, obtains K_s×K_tIndividual sampled point；K_sRepresent the clutter data of l-th of training sample unit in non-working side battle array radar Normalize the interval sampled point number of spatial frequency, K_tRepresent the clutter data of l-th of training sample unit in non-working side battle array The interval sampled point number of the normalization Doppler frequency of radar；

Step 5, for l-th of training sample unit clutter data in the space-time two-dimensional plane of non-working side battle array radar the (p, q) individual sampled point, obtains the space-time steering vector of (p, q) individual sample point, p ∈ { 1,2 ..., K_t, q ∈ 1, 2 ..., K_s}；So as to obtain the clutter data of l-th of training sample unit in the space-time two-dimensional plane of non-working side battle array radar All K_s×K_tThe space-time steering vector of individual sample point；

Step 6, according to the clutter data of l-th of training sample unit, the space-time steering vector of (p, q) individual sample point, meter Calculate the initial power of (p, q) individual sample pointP ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}；And according to each sampling Initial power at point determines the clutter power spectrum P of the clutter data of l-th of training sample unit⁽⁰⁾；

And set iterations n initial value to be 1；

Step 7, according to the power of (p, q) individual sample pointAnd the space-time steering vector structure of (p, q) individual sample point Make the covariance matrix of (p, q) individual sample pointP ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}；

Step 8, according to the clutter data of l-th of training sample unit, the space-time steering vector of (p, q) individual sample point and The covariance matrix of (p, q) individual sample pointCalculate the power of (p, q) individual sample point after nth iterationp ∈ { 1,2 ..., K_t, q ∈ { 1,2 ..., K_s}；And according to the power of each sample point after nth iterationDetermine, it is individual Clutter power spectrum P after the clutter data nth iteration of training sample unit⁽ⁿ⁾；

Step 9, if | | P⁽ⁿ⁾||_l1-||P^(n-1)||_l1＞ ε, and 0 ＜ n ＜ num, then make n value plus 1, repeats step 7 and step Rapid 8；Num represents the maximum of the iterations of setting, | | | |_l1L1 norms are sought in expression；

If | | P⁽ⁿ⁾||_l1-||P^(n-1)||_l1＜ ε, or n ＞ num, then stop iteration, and by each sampled point after nth iteration The power at placeIt is used as the final power of each sample point in the clutter data of l-th of training sample unit；

Step 10, the maximum of the final power of each sample point in the clutter data of l-th of training sample unit is determined, is obtained The normalization Doppler frequency and normalization spatial frequency of the corresponding sample point of the maximum are taken, using it as l-th The corresponding normalization Doppler frequency of power spectrum and normalization spatial frequency of the clutter data of training sample unit；

Step 11, determine the corresponding normalization Doppler frequency of clutter power spectrum of the clutter data of range cell to be detected and return One changes spatial frequency；According to the power spectrum of the clutter data of l-th of training sample unit it is corresponding normalization Doppler frequency and Spatial frequency is normalized, the clutter power spectrum of the clutter data of range cell to be detected is corresponding to be normalized Doppler frequency and return One changes spatial frequency, calculates the corresponding amendment Doppler frequency of power spectrum of the clutter data of l-th of training sample unit and repaiies Positive space frequency；

Step 12, according to the corresponding amendment Doppler frequency of the power spectrum of the clutter data of l-th of training sample unit and Spatial frequency is corrected, the power spectrum correspondence Doppler domain transition matrix and sky of the clutter data of l-th of training sample unit is calculated Domain transition matrix, so power spectrum correspondence Doppler domain transition matrix according to the clutter data of l-th of training sample unit and Spatial domain transition matrix, calculates the power spectrum correspondence correction matrix for the clutter data for obtaining l-th of training sample unit；

According to the power spectrum pair of the clutter data of l-th training sample unit and the clutter data of l-th of training sample unit Correction matrix is answered, the clutter data by revised l-th of training sample unit is obtained；

Step 13, make l value plus 1, and repeat step 4 to step 12, so that obtaining L-1 passes through revised clutter number According to so that according to the L-1 covariance that the clutter data for obtaining range cell to be detected is calculated by revised clutter data Matrix；

The space-time steering vector of range cell to be detected is obtained, so that according to the space-time steering vector of the range cell to be detected And the covariance matrix of the clutter data of range cell to be detected, filtering weight vector is tried to achieve, according to the filtering weight vector point The other clutter data to the L range cell carries out angle Doppler's registration, and the clutter data for obtaining L range cell is carried out Result after angle Doppler registration.

2. a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum according to claim 1, Characterized in that, in step 1, obtaining the clutter data for the L range cell that non-working side battle array radar is received, being specially：

The linear array for setting the antenna of the non-working side battle array radar to be made up of equally distributed N number of array element, and it is described non- The M pulse that positive side view battle array radar is launched in a coherent processing inteval；

L times is carried out respectively on each pulse, each array element passage apart from upper sampling, then the clutter data being an actually-received is The three-dimensional data of N × M × L dimensions；By m-th of reception of impulse of n-th of array element of l-th of range cell to clutter data be designated as x_nml, then the clutter data that n-th of array element of l-th of range cell is received is designated as x_nl=[x_n1l, x_n2l..., x_nMl]^T, by l The clutter data that N number of array element of individual range cell is received lines up the column vector X that NM × 1 is tieed up_l, so that it is single to obtain l-th of distance The clutter data X of member_l=[x_1l ^T, x_2l ^T..., x_Nl ^T]^T, wherein, l ∈ { 1,2 ..., L }, ()^TRepresent transposition.

3. a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum according to claim 1, Characterized in that, in step 4, K_s=10N, K_t=I0M；

Wherein, N represents the element number of array that the antenna of non-working side battle array radar is included, and M represents non-working side battle array radar in a phase The pulse number launched in dry-cure interval, K_sRepresent the clutter data of l-th of training sample unit in non-working side battle array radar The interval sampled point number of normalization spatial frequency, K_tRepresent the clutter data of l-th of training sample unit in non-working side The sampled point number in the normalization Doppler frequency interval of battle array radar.

4. a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum according to claim 1, Characterized in that, in step 5, obtaining the space-time steering vector of (p, q) individual sample point, it is specially：

Obtain steering vector of (p, q) the individual sampled point on the normalization spatial frequency interval of non-working side battle array radar

Obtain steering vector of (p, q) the individual sampled point on the normalization Doppler frequency interval of non-working side battle array radar

So as to the space-time steering vector of (p, q) individual sample point

Wherein, f_{D, p}Represent that how general the normalization of interval upper p-th of the sampled point of normalization Doppler frequency of non-working side battle array radar is Strangle frequency, f_{S, q}Represent the normalization space frequency of interval upper q-th of the sampled point of normalization spatial frequency of non-working side battle array radar Rate；Represent kronecker products, ()^TRepresent transposition.

5. a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum according to claim 1, Characterized in that, in step 6,

The initial power of (p, q) individual sample point is calculated using following formula

${\mathrm{\σ}}_{p,q}^0={\left|\frac{S^H(f_{d,p},f_{s,q})X_l}{S^H(f_{d,p},f_{s,q})S\left(f_{d,p}{f}_{s,q}\right)}\right|}^2$

Wherein, S (f_{D, p}, f_{S, q}) represent (p, q) individual sample point space-time steering vector, X_lRepresent l-th of training sample unit Clutter data, ()^HConjugate transposition is represented, | | represent the operation that takes absolute value；

The clutter power spectrum P of the clutter data of l-th of training sample unit is determined according to the initial power of each sample point⁽⁰⁾,Diag () represents that diagonal matrix, i.e. diagonal element are Diagonal matrix.

6. a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum according to claim 1, Characterized in that, in step 7,

According to the power of (p, q) individual sample pointAnd the space-time steering vector construction the of (p, q) individual sample point The covariance matrix of (p, q) individual sample pointIt is as follows：

$R_{p,q}^n=\underset{p=1}{\overset{K_t}{\mathrm{\Σ}}}\underset{q=1}{\overset{K_s}{\mathrm{\Σ}}}{\mathrm{\σ}}_{p,q}^{n-1}S(f_{d,p},f_{s,q})S^H(f_{d,p},f_{s,q})$

Wherein, S (f_{D, p}, f_{S, q}) represent (p, q) individual sample point space-time steering vector.

7. a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum according to claim 1, Characterized in that, in step 8,

According to the clutter data of l-th of training sample unit, the space-time steering vector of (p, q) individual sample point and (p, Q) covariance matrix of individual sample pointCalculate the power of (p, q) individual sample point after nth iterationIt is as follows：

${\mathrm{\σ}}_{p,q}^n={\left|\frac{S^H(f_{d,p},f_{s,q}){\left(R_{p,q}^n\right)}^{-1}{X}_l}{S^H(f_{d,p},f_{s,q}){\left(R_{p,q}^n\right)}^{-1}S\left(f_{d,p}{f}_{s,q}\right)}\right|}^2$

Wherein, S (f_{D, p}, f_{S, q}) represent (p, q) individual sample point space-time steering vector, X_lRepresent l-th of training sample unit Clutter data, ()^HConjugate transposition is represented, | | represent take absolute value operation, ()^-1Representing matrix is inverted；

According to the power of each sample point after nth iterationDetermine the clutter data n-th of l-th of training sample unit Clutter power spectrum P after iteration⁽ⁿ⁾It is as follows：

$P^{\left(n\right)}=diag(\[{\mathrm{\σ}}_{1,1}^n,{\mathrm{\σ}}_{2,1}^n,{\mathrm{\σ}}_{3,1}^n,...,{\mathrm{\σ}}_{K_t,K_s}^n\])$

Wherein, diag () represents that diagonal matrix, i.e. diagonal element areDiagonal matrix.

8. a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum according to claim 1, Characterized in that, in step 11：

Determine the corresponding normalization Doppler frequency f of clutter power spectrum of the clutter data of range cell to be detected_{D, k}And normalization Spatial frequency f_{S, k}：

Wherein, λ is the operation wavelength of non-working side battle array radar,For the array element spacing of even linear array, non-working side battle array radar Place carrier aircraft is flown with speed v along x-axis, and plane where even linear array is vertical with x-y plane, and plane where even linear array and The angle of x-axis isθ_kThe angle of pitch corresponding to range cell clutter data to be detected is represented,Represent range cell to be detected Clutter data is relative to the azimuth corresponding to even linear array, f_rFor pulse recurrence frequency；

Calculate the corresponding amendment Doppler frequency Δ f of power spectrum of the clutter data of l-th of training sample unit_{D, l}=f_{D, l-fd, k} With amendment spatial frequency Δ f_{S, l}=f_{S, l}-f_{S, k}；

Wherein, f_{D, l}Represent the corresponding normalization Doppler frequency of power spectrum of the clutter data of l-th of training sample unit, f_{S, l} Represent the corresponding normalization spatial frequency of power spectrum of the clutter data of l-th of training sample unit, f_{D, k}Represent distance to be detected The corresponding normalization Doppler frequency of clutter power spectrum of the clutter data of unit, f_{S, k}Represent the clutter of range cell to be detected The corresponding normalization spatial frequency of clutter power spectrum of data.

9. a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum according to claim 1, Characterized in that, in step 12,

Calculate the power spectrum correspondence Doppler domain transition matrix T of the clutter data of l-th of training sample unit_{D, l}With spatial domain conversion Matrix T_{S, l}：

$T_{d,l}=\left[\begin{array}{cccc}1& 0& ...& 0\\ 0& e^{j2{\mathrm{\π\Δf}}_{d,l}}& ...& ...\\ ...& ...& ...& 0\\ 0& ...& 0& e^{j2\mathrm{\π}(M-1){\mathrm{\Δf}}_{d,l}}\end{array}\right]$

$T_{s,l}=\left[\begin{array}{cccc}1& 0& ...& 0\\ 0& e^{j2{\mathrm{\π\Δf}}_{s,l}}& ...& ...\\ ...& ...& ...& 0\\ 0& ...& 0& e^{j2\mathrm{\π}(N-1){\mathrm{\Δf}}_{s,l}}\end{array}\right]$

Calculate the power spectrum correspondence correction matrix T for the clutter data for obtaining l-th of training sample unit_IAA-ADC：

$T_{IAA-ADC}=\left[\begin{array}{cccc}1& 0& \mathrm{...}& 0\\ 0& e^{j2{\mathrm{\π\Δf}}_{d,l}}& \mathrm{...}& \mathrm{...}\\ \mathrm{...}& \mathrm{...}& \mathrm{...}& 0\\ 0& \mathrm{...}& 0& e^{j2\mathrm{\π}(M-1){\mathrm{\Δf}}_{d,l}}\end{array}\right]\⊗\left[\begin{array}{cccc}1& 0& \mathrm{...}& 0\\ 0& e^{j2{\mathrm{\π\Δf}}_{s,l}}& \mathrm{...}& \mathrm{...}\\ \mathrm{...}& \mathrm{...}& \mathrm{...}& 0\\ 0& \mathrm{...}& 0& e^{j2\mathrm{\π}(N-1){\mathrm{\Δf}}_{s,l}}\end{array}\right]$

According to the power spectrum pair of the clutter data of l-th training sample unit and the clutter data of l-th of training sample unit Correction matrix is answered, the clutter data by revised l-th of training sample unit is obtained

Wherein, Δ f_{D, l}Represent the corresponding amendment Doppler frequency of power spectrum of the clutter data of l-th of training sample unit, Δ f_{S, l}The corresponding amendment spatial frequency of power spectrum of the clutter data of l-th of training sample unit is represented,Represent kronecker Product, N represents the element number of array that the antenna of non-working side battle array radar is included, and M represents non-working side battle array radar in a Coherent processing The pulse number of transmitting in interval.

10. a kind of optimization method of the angle Doppler registration of non-working side battle array radar clutter spectrum according to claim 1, Characterized in that, in step 13,

According to the L-1 covariance matrix that the clutter data for obtaining range cell to be detected is calculated by revised clutter data

Wherein,Represent the clutter data by revised l-th of training sample unit；

Obtain the space-time steering vector S (f of range cell to be detected_{D, k}, f_{S, k}), led according to the space-time of the range cell to be detected To vector S (f_{D, k}, f_{S, k}) and range cell to be detected clutter data covariance matrixTry to achieve range cell to be detected Filtering weight vector

Wherein,For normaliztion constant.