A broadband array mainlobe interference rejection method

Abstract

The application discloses a broadband array main lobe interference suppression method, and steps are as follows: S1: Taylor estimation method is adopted to obtain a higher precision covariance matrix of a received signal model; S2: eigenvalue decomposition is carried out on the covariance matrix obtained in S1, and a characteristic projection matrix is constructed; S3: the received signal is preprocessed by the characteristic projection matrix, and a frequency domain output is obtained; S4: a focusing transformation is adopted to convert the broadband signal into a narrowband signal, then the narrowband signal covariance matrix is reconstructed; S5: the optimal weight vector of the narrowband signal is obtained under the premise that the weight vector projection is unchanged within the main lobe range; S6: the optimal weight vector corresponding to each frequency of the broadband is obtained through the focusing inverse transformation, and the broadband signal from which the main lobe interference is removed is weighted; and S7: a new frequency domain output is obtained, then inverse Fourier transformation is carried out on the new frequency domain output, and the broadband main lobe interference suppression adaptive beam forming is completed. By adopting the method, the main lobe interference can be effectively suppressed, and the operation amount is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of radar signal processing, in particular to a wideband array main lobe interference suppression method. BACKGROUND

[0002] With the advent of the information age, communication is widely used in various fields. Among them, the wideband array main lobe interference suppression adaptive beam forming technology is more critical. This technology can make the adaptive beam forming obtain good characteristics and ensure the normal operation of communication. In the process of wideband adaptive beam forming, when there is interference in the main lobe, the peak of the main lobe beam will be shifted, resulting in beam non-shape, main lobe transmission distortion and a series of problems.

[0003] Domestic scholars have conducted a lot of research on main lobe interference suppression. Among them, there are many methods about blocking matrix processing, eigenvalue projection matrix processing and wideband constant beam. These three methods have certain effect on main lobe interference suppression, but at the same time, they also have some defects. For the blocking matrix processing algorithm, this method needs to block the signal, which will cause the system to lose the degree of freedom and make the main lobe of the directional diagram wide. For the eigenvalue projection matrix processing algorithm, the signal weighting is also processed by the eigenvalue projection matrix, which has little effect on the degree of freedom, but it will cause the main lobe to shift and other problems. For the wideband constant beam, this method increases the burden of operation due to the introduction of the second-order cone programming to obtain the weight coefficient. SUMMARY

[0004] The purpose of the present application is to provide a wideband array main lobe interference suppression method, which effectively avoids the defects in the above-mentioned technical methods, reduces the amount of operation, ensures the correctness of the main lobe directional diagram, and reduces the sidelobe level to a certain extent.

[0005] To achieve the above purpose, the present application provides a wideband array main lobe interference suppression method, the steps are as follows:

[0006] S1: using Taylor estimation method to obtain a higher precision covariance matrix for the received signal model;

[0007] S2: eigenvalue decomposition is performed on the covariance matrix obtained in S1 to construct an eigenvalue projection matrix;

[0008] S3: eigenvalue projection matrix preprocessing is performed on the received signal to obtain a frequency domain output;

[0009] S4: using focusing transformation to convert the wideband signal into a narrowband signal, and then reconstructing the narrowband signal covariance matrix;

[0010] S5: under the premise that the weight vector projection in the main lobe range is unchanged, the optimal weight vector of the narrowband signal is obtained;

[0011] S6: obtaining the optimal weight vector corresponding to each frequency of the wideband by focusing inverse transform, and weighting the wideband signal after removing the main lobe interference;

[0012] S7: calculating the new frequency domain output, and performing inverse Fourier transform on the new frequency domain output to complete the wideband main lobe interference suppression adaptive beam forming.

[0013] Preferably, the expression of the covariance matrix in S1 is as follows:

[0014]

[0015] wherein, is the covariance matrix of the received signal; A matrix is a matrix composed of steering vectors a(θ i ), A = [a(θ1), …, a(θ m )], m is the sum of the main lobe and side lobe interference numbers; q i is the power of the i-th interference, is the noise power.

[0016] Preferably, the expression of the eigenvalue in S2 is as follows:

[0017]

[0018] λ i is the eigenvalue of , i = 1, 2, …, m, u i is the eigenvector corresponding to the i-th eigenvalue, Λ g is the interference subspace U g is the eigenvalue diagonal matrix, Λ n is the noise subspace U n is the eigenvalue diagonal matrix, p is the number of array elements, is the covariance matrix of the received signal, u g is the eigenvector corresponding to the g-th eigenvalue, u n is the eigenvector corresponding to the n-th eigenvalue;

[0019] Assuming that u m is the eigenvector corresponding to the main lobe interference, the expression of the eigenvalue projection matrix is:

[0020]

[0021] wherein, I is the unit matrix, |u m H a(θ0)| 2 ≥ ρ |a(θ0)| 2, where a(0) is the steering vector, p is a scalar parameter that limits the correlation between the eigenvector and the steering vector, and u m is the eigenvector corresponding to the main lobe interference, otherwise it is the eigenvector corresponding to the side lobe interference.

[0022] Preferably, S3 is specifically as follows:

[0023] Suppose the received signal is x(n), and the discrete Fourier transform (DFT) of x(n) is performed to obtain the frequency domain signal at f i , whose expression is as follows:

[0024] X(f i )=A(f i )S(f i )+N(f i ),i=1,2…,J

[0025] where A(f i ) is the array manifold vector matrix A(f i , θ) = [a(f i , θ1), a(f i , θ2), … a(f i , θ K )], S(f i ) is the signal spectrum component, N(f i ) is the noise spectrum component, and J is the number of Fourier transform points;

[0026] X(f i ) is multiplied by the eigenvector projection matrix obtained in S2 to obtain the frequency domain output X o (f i ), whose expression is as follows:

[0027] X o (f i )=B(f i )A(f i )S(f i )+B(f i )N(f i ),

[0028] where B(f i ) is the eigenvector projection matrix.

[0029] Preferably, the specific steps of S4 are as follows:

[0030] Step 1: Calculate the cross-correlation matrix R(f i , f0) and the noise-free covariance matrix P(f0), whose expressions are as follows:

[0031] R(f i ,f0)=A(f i, θ)S(f i )S H (f0)A H (f0) / L

[0032] P(f0) = A(f0, θ)S(f0)S H (f0)A H (f0) / L

[0033] where L is the number of received signal snapshots, A(f i , θ) = [a(f i , θ1), a(f i , θ2),... a(f i , θ K )], S(f i ) is the signal spectral component, the available approximations of the cross-correlation matrix R(f i , f0) and the noise covariance matrix P(f0) are expressed as follows:

[0034]

[0035]

[0036] where is the noise variance at the reference frequency, and the expression obtained after rearrangement is:

[0037] T(f i )R(f i , f0) = P(f0)

[0038] Step 2: Calculate the focusing matrix T(f i ), the expression of which is:

[0039] T(f i ) = P(f0)(P H (f0)P(f0)) -1 / 2 ·(R H (f i , f0)R(f i , f0)) -1 / 2 R H (f i , f0);

[0040] Step 3: Calculate the focused output signal Y(f i ) and the narrowband signal covariance matrix R Y , the expressions of which are:

[0041] Y(f i ) = T(f i )X0(f i ),

[0042]

[0043] Where Y(f) i T(f) is the focused output signal. i X is the focusing matrix. o (f i () represents the frequency domain output after removing main lobe interference, J is the number of points in the Fourier transform, and R... Y The covariance matrix of the narrowband signal;

[0044] Step 4: For the narrowband signal covariance matrix R Y The eigenvalue decomposition is expressed as follows:

[0045]

[0046] Where λ i R is the covariance matrix of the narrowband signal. Y The eigenvalues ​​are i = 1, 2, ..., m, and λ1 ≥ λ2 ≥ ... λ i-1 u i Let p be the eigenvector corresponding to the i-th eigenvalue, and p be the number of matrix elements.

[0047] Step 5: At the eigenvalue λ i Select the eigenvalue λ corresponding to the noise. i-d , λ i-d+1 、…λ i-1 d is the number of eigenvalues ​​corresponding to the noise, expressed as:

[0048]

[0049] Calculate the noise eigenvalue λ based on the number d of noise eigenvalues. i-d , λ i-d+1 、…λ i-1 average

[0050]

[0051] Step 6: Reconstruct the covariance matrix, its expression is:

[0052]

[0053] in Y = [y1, y2, ..., y] is a diagonal matrix composed of the corrected eigenvalues. P-1 ] is the eigenvector matrix, and R is the reconstructed covariance matrix.

[0054] Preferably, the optimal weight vector described in S5 is obtained through the following steps:

[0055] Step 1: Determine the main lobe subspace according to the formula:

[0056] F = ∫ φ a(θ)a(θ) H dθ

[0057]

[0058] where the main lobe region subspace is spanned by the column vectors of U1, ω i is the i-th eigenvalue, v i is the corresponding eigenvector, Λ1 is a diagonal matrix composed of the first K2 eigenvalues, U1 is a matrix composed of the eigenvectors corresponding to the eigenvalues in Λ1, Λ2 is composed of P-K2 eigenvalues, U2 is a matrix composed of the eigenvectors corresponding to the eigenvalues in Λ2, and a(θ) is the steering vector of the desired direction θ;

[0059] Step 2: Determine the projection matrix E of the main lobe subspace through the main lobe subspace, and the expression is:

[0060] E = U1(U1 H U1) -1 U1 H

[0061] Step 3: Finally, calculate the optimal weight vector W(f0) under the minimum variance criterion with the constraint that the projection of the weight vector in the main lobe subspace is invariant, and the expression is:

[0062]

[0063] The expression obtained by arranging the above formula is:

[0064]

[0065] where E is the projection matrix of the main lobe subspace, R is the modified signal covariance, W(f0) is the optimal weight vector, U1 is a matrix composed of eigenvectors corresponding to eigenvalues in Λ1, and Λ1 is a diagonal matrix composed of modified eigenvalues.

[0066] Preferably, the specific steps of S7 are as follows:

[0067] Step 1: Perform focus inverse transformation on the optimal weight vector W(f0) obtained in S5 to obtain the optimal weight vector at frequency f i , and the expression is:

[0068] W'(f i ) = T -1 (f i )W(f0);

[0069] where T -1 (f i) is the focused matrix inverse transform, W(f0) is the optimal weight vector at frequency f0, W'(f i ) is the optimal weight vector at frequency f i ;

[0070] Step 2: calculate the new frequency domain output, the expression of which is:

[0071] Z(f) = [W'(f1)X o (f1), W'(f2)X o (f2), …, W'(f i )X o (f i ), …, W'(f J )X o (f J )]

[0072] wherein X o (f i ) is the frequency domain output after removing the main lobe interference, W'(f i ) is the optimal weight vector at frequency f i , and Z(f) is the frequency domain output;

[0073] Step 3: perform Fourier inverse transform on the new frequency domain output obtained in Step 2 to obtain the final signal, the expression of which is:

[0074] z(n) = IDFT[Z(f)].

[0075] Therefore, by adopting the above-mentioned wideband array main lobe interference suppression method, the main lobe interference can be effectively suppressed, the sidelobe interference suppression is strengthened, the main lobe width of the shaped directional diagram remains unchanged, the main lobe direction is correct, and the operation amount is small. BRIEF DESCRIPTION OF DRAWINGS

[0076] Figure 1 is the implementation flowchart of the wideband array main lobe interference suppression method of the present application;

[0077] Figure 2 is the comparison diagram of the directional diagram at the center frequency of the method of the present application and the conventional wideband constant beam main lobe suppression method in Example 1;

[0078] Figure 3 is the comparison diagram of the directional diagram at the center frequency of the method of the present application and the blocking matrix algorithm main lobe interference suppression method in Example 1. DETAILED DESCRIPTION

[0079] The technical solutions of the present application are further described below by means of the drawings and examples.

[0080] As Figure 1As shown, the present invention provides a method for suppressing main lobe interference in a broadband array, comprising the following steps:

[0081] Step 1: Obtain a more accurate covariance matrix using the Taylor estimation method.

[0082]

[0083] in, Let A be the covariance matrix of the received signal; let A be the steering vector a(θ). i A matrix consisting of [a(θ1), ..., a(θ)] is given by [a(θ1), ..., a(θ)]. m )], m is the sum of the interference numbers of the main lobe and side lobes; q i Let be the power of the i-th interference. This represents noise power.

[0084] Step 2: Perform eigenvalue decomposition on the covariance matrix and construct the eigenprojection matrix.

[0085] This step involves the covariance matrix obtained in step 1. Decompose and construct the feature projection matrix.

[0086]

[0087] λ i for The eigenvalues, i = 1, 2, ..., m; u i Let Λ be the eigenvector corresponding to the i-th eigenvalue. g For the interference subspace U g Eigenvalue diagonal matrix, Λ n For the noise subspace U n Eigenvalue diagonal matrix, where p is the number of elements. Let u be the covariance matrix of the received signal. g Let u be the eigenvector corresponding to the g-th eigenvalue. n This is the eigenvector corresponding to the nth eigenvalue.

[0088] Assume u m The feature vector corresponding to the main lobe interference has the following feature projection matrix:

[0089]

[0090] I is the identity matrix, |u m H a(θ0)| 2 ≥ρ|a(θ0)| 2 ρ is a scalar parameter that restricts the correlation between the eigenvector and the steering vector, and α(θ0) is the steering vector; u satisfies the above equation mThat is the main lobe interference corresponding to the feature vector, otherwise it is the side lobe interference feature vector.

[0091] Step three, pre-processing the received signal and getting the frequency domain output.

[0092] Assume that the received signal is x(n), and the discrete Fourier transform DFT is performed on x(n) to obtain the frequency domain signal f i at:

[0093] X(f i )=A(f i ,θ)S(f i )+N(f i )i=1,2…,J

[0094] Wherein, A(f i ) is the array manifold vector matrix, A(f i ,θ)=[a(f i ,θ1),a(f i ,θ2),…a(f i ,θ K )], S(f i ) is the signal spectrum component, N(f i ) is the noise spectrum component, and J is the number of Fourier transform points.

[0095] X(f i ) is multiplied by the characteristic projection matrix B(f i ) to obtain the frequency domain output X o (f i ), the formula is as follows:

[0096] X o (f i )=B(f i )A(f i ,θ)S(f i )+B(f i )N(f i )

[0097] Step four, this step is mainly to convert the wideband signal to narrowband signal by using frequency domain focusing transform.

[0098] According to the array manifold vector matrix A(f i ), the cross-correlation matrix R(f i ,f0), and the noiseless covariance matrix P(f0), the focusing matrix T(f i )

[0099] R(f i ,f0)=A(f i ,θ)S(f i )S H(f0)A H (f0) / L

[0100] P(f0) = A(f0, θ)S(f0)S H (f0)A H (f0) / L

[0101] where L is the number of received signal snapshots, A(f i , θ) = [a(f i , θ1), a(f i , θ2),... a(f i , θ K )], S(f i ) is the signal spectral component.

[0102] An approximate value of the cross-correlation matrix R(f i , f0) and the noise covariance matrix P(f0) can be obtained as follows:

[0103]

[0104]

[0105] where σ2(f0) is the noise variance at the reference frequency.

[0106] After rearrangement, T(f i )R(f i , f0) = P(f0), and the focusing matrix is:

[0107] T(f i ) = P(f0)(P H (f0)P(f0)) -1 / 2 ·(R H (f i ,f0)R(f i ,f0)) -1 / 2 R H (f i ,f0)

[0108] The focusing matrix T(f i ) is multiplied by the frequency-domain output X o (f i ) after the main-lobe interference is removed, to obtain the focused output signal Y(f i ) and the covariance matrix R Y , which are expressed as follows:

[0109] Y(f i ) = T(f i )X0(f i )

[0110] ​

[0111] Where R Y Let J be the covariance matrix of the narrowband signal, and J be the number of points in the Fourier transform.

[0112] Step 5: Reconstruct the covariance matrix of the received signal.

[0113] For narrowband signal covariance matrix R Y Eigenvalue decomposition yields:

[0114]

[0115] Where λ i For R Y The eigenvalues ​​are i = 1, 2, ..., m, and λ1 ≥ λ2 ≥ ... λ i-1 u i Let be the eigenvector corresponding to the i-th eigenvalue.

[0116] At the eigenvalue λ i Select the eigenvalue λ corresponding to the noise. i-d , λ i-d+1 、…λ i-1 d is the number of eigenvalues ​​corresponding to the noise, expressed as:

[0117]

[0118] Calculate the noise eigenvalue λ based on the number d of the noise eigenvalues. i-d , λ i-d+1 、…λ i-1 average The expression is:

[0119]

[0120] Using noise mean The noise characteristic value is corrected, and the noise characteristic value λ is set to... i-d , λ i-d+1 、…λ i-1 The average value is equal to the noise mean. The corrected signal covariance matrix is ​​obtained. The expression is:

[0121]

[0122] in, The diagonal matrix formed by the corrected eigenvalues. Y = [y1, y2, ... y P-1 ] is the eigenvector matrix.

[0123] Step six, calculate the optimal weight vector W, this step mainly in the main lobe subspace projection is not changed as a constraint condition to calculate the minimum variance criterion under the optimal weight vector W(f0), according to the following formula:

[0124] F = ∫ φ a(θ)a(θ) H dθ

[0125]

[0126] Determine the main lobe region subspace, the main lobe region subspace is the column vector of U1. i The i-th eigenvalue, v i The corresponding eigenvector, Λ1 is a diagonal matrix composed of the first K2 eigenvalues, U1 is a matrix composed of the eigenvectors corresponding to the eigenvalues in Λ1, Λ2 is composed of P-K2 eigenvalues, U2 is a matrix composed of the eigenvectors corresponding to the eigenvalues in Λ2, a(θ) is the steering vector of the expected direction θ.

[0127] According to the above main lobe subspace U1, construct the main lobe subspace projection matrix E, the formula is as follows:

[0128] E = U1(U1 H U1) -1 U1 H

[0129] With the projection of the weight vector in the main lobe subspace U1 as a constraint condition, according to the modified signal covariance matrix R obtained in step five, calculate the optimal weight vector under the minimum variance condition, the expression is:

[0130]

[0131] Where a(θ) is the steering vector of the expected direction θ, W is the optimal weight vector, E is the projection matrix of the main lobe subspace.

[0132] Finally, the optimal weight vector is obtained, the expression is:

[0133]

[0134] Where E is the projection matrix of the main lobe subspace, R is the modified signal covariance, W(f0) is the optimal weight vector, U1 is a matrix composed of the eigenvectors corresponding to the eigenvalues in Λ1, A diagonal matrix composed of the modified eigenvalues.

[0135] Step seven, this step mainly obtains the final frequency domain output according to the optimal weight vector at each frequency.

[0136] Perform focus inverse transform on the optimal weight vector obtained in step six to obtain the frequency f iThe expression of the optimal weight vector at the frequency f0 is:

[0137] W'(f i )=T -1 (f i )W(f0)

[0138] Where T -1 (f i ) is the focusing matrix inverse transform, W(f0) is the optimal weight vector, and W'(f i ) is the optimal weight vector at the frequency f i .

[0139] The frequency domain output Z(f) is obtained according to the frequency domain output X o (f i ) after the main lobe interference is removed and the optimal weight vector W'(f i ) at the frequency f i , and the expression is:

[0140] Z(f)=[W'(f1)X o (f1),W'(f2)X o (f2),…,W'(f i )X o (f i ),…,W'(f J )X o (f J )]

[0141] Where X o (f i ) is the frequency domain output after the main lobe interference is removed, W'(f i ) is the optimal weight vector at the frequency f i , and Z(f) is the frequency domain output.

[0142] The frequency domain output Z(f) obtained in the above formula is Fourier inverse transformed to obtain the final signal: z(n)=IDFT[Z(f)].

[0143] Embodiment one

[0144] I. Simulation parameters, as shown in Table 1:

[0145] Table 1

[0146] Parameter Parameter value Parameter Parameter value Number of array elements 8 Signal-to-noise ratio 0 dB Inter-element spacing Half-wavelength Sampling frequency 150 MHz Beam pointing 0° Drying ratio 60 dB Center frequency 2 GHz Main lobe interference direction 3° Bandwidth 60 MHz Number of pulse repetition periods 200 us Pulse width 50 us Number of dwell periods 5 Target direction 0° Noise type Gaussian white noise Number of received signal snapshots 128

[0147] II. Simulation content:

[0148] Simulation 1: Under the simulation conditions in Table 1, the antenna patterns of the main lobe suppression method of the present application and the wideband constant beam forming are compared, and the results are shown in Figure 2 .

[0149] Figure 2 The middle black dashed line represents the antenna pattern obtained by the main lobe suppression method of the wideband constant beam forming, and the red line represents the antenna pattern obtained by the application. Figure 2 It can be seen that the antenna pattern of the application has obvious suppression effect on the sidelobe level, and the calculation time is greatly reduced because the application does not need to perform matrix inversion, the operation time of the original technology is 9.193 seconds, and the operation time of the application is 1.65561 seconds.

[0150] Simulation 2: Under the simulation conditions in Table 1, the application is compared with the blocking matrix wideband array main lobe interference suppression method, and the results are shown in Figure 3 .

[0151] Figure 3 The purple line represents the antenna pattern obtained by the blocking matrix main lobe suppression method, and the red line represents the antenna pattern obtained by the method of the application. Figure 2 It can be seen that compared with the blocking matrix main lobe interference suppression method, the main lobe of the application is narrower, the sidelobe suppression is lower, and the pattern is more correct.

[0152] Therefore, by using the above-mentioned wideband array main lobe interference suppression method, compared with the conventional wideband constant beam main lobe suppression method, the application does not need to perform covariance matrix inversion, and the operation time is obviously reduced, and the calculation efficiency is improved; compared with the blocking matrix algorithm for suppressing main lobe interference, the application reconstructs the signal covariance matrix to remove the influence of noise on adaptive beam forming, so that the sidelobe level of the pattern is low, and the notch of the sidelobe interference part is deep. The method of the application effectively suppresses the main lobe interference, strengthens the sidelobe interference suppression, the main lobe direction is correct, and the operation amount is reduced.

[0153] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the application and not to limit them, although the application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the application can still be modified or replaced by equivalents, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the application.

Claims

1. A method of broadband array mainlobe jammer rejection, characterized by: The steps are as follows: S1: Taylor estimation method is used for the received signal model to obtain a higher precision covariance matrix; S2: eigenvalue decomposition is performed on the covariance matrix obtained in S1 and a feature projection matrix is constructed; S3: the received signal is preprocessed by the feature projection matrix and the frequency domain output is obtained; S4: the wideband signal is converted into a narrowband signal by focusing transformation, and then the narrowband signal covariance matrix is reconstructed; S5: the optimal weight vector of the narrowband signal is obtained under the premise that the weight vector projection is unchanged within the main lobe range; S6: the optimal weight vector corresponding to each frequency of the wideband is obtained by inverse focusing transformation, and the wideband signal with the main lobe interference removed is weighted; S7: the new frequency domain output is calculated, and then inverse Fourier transform is performed on the new frequency domain output to complete the wideband main lobe interference suppression adaptive beamforming.

2. The wideband array mainlobe jammer rejection method of claim 1, wherein: The expression of the covariance matrix in S1 is as follows: ； wherein is the covariance matrix of the received signal; A is the steering vector is the matrix composed of , m is the sum of the main lobe and side lobe interference number; , is the power of the i-th interference, is the noise power.

3. The wideband array mainlobe jammer rejection method of claim 1, wherein: The expression of the eigenvalue in S2 is as follows: ； are eigenvalues of , i = 1, 2,..., m, is the eigenvector corresponding to the ith eigenvalue, is the interference subspace is the eigenvalue diagonal matrix, is the noise subspace is the eigenvalue diagonal matrix, p is the number of array elements, is the covariance matrix of the received signal, is the eigenvector corresponding to the gth eigenvalue, is the eigenvector corresponding to the nth eigenvalue; Assume For the main-lobe interference corresponding eigenvector, the expression of the eigen-projection matrix is: ； where I is an identity matrix, , is a steering vector, is a scalar parameter that limits the correlation between the eigen vector and the steering vector, satisfying the above equation is the eigen vector corresponding to the main lobe interference, otherwise it is the eigen vector of the side lobe interference.

4. The wideband array mainlobe jammer rejection method of claim 1, wherein: S3 is as follows: Assume that the received signal is The discrete Fourier transform (DFT) is performed on the signal to obtain a frequency domain signal at , whose expression is ， ； wherein is an array manifold vector matrix , is a signal spectral component, is a noise spectral component, J is the number of points of the Fourier transform; The The frequency domain output is obtained by multiplying the characteristic projection matrix obtained in S2 The expression is: ； wherein is the characteristic projection matrix.

5. The wideband array mainlobe jammer rejection method of claim 1, wherein: The specific steps of S4 are as follows: Step 1: Compute the cross-correlation matrix and the noiseless covariance matrix , which is expressed as follows: ； ； where L is the number of received signal snapshots, , is the signal spectral component, the cross-correlation matrix and the available approximation of the noiseless covariance matrix is given by the expression ； ； wherein is the noise variance at the reference frequency, the expression is arranged as follows: ； Step 2: Compute the focus matrix whose expression is: ； Step 3: Calculate focused output signal and narrowband signal covariance matrix , expressed as: ； ； wherein is a focused output signal, is a focusing matrix, is a frequency domain output after removing main lobe interference, J is the number of Fourier transform points, is a narrowband signal covariance matrix; Step 4: Covariance matrix for narrowband signals Eigenvalue decomposition is performed, which is expressed as: ； wherein is an eigenvalue of the narrowband signal covariance matrix , i = 1, 2,... m, and , is an eigenvector corresponding to the ith eigenvalue, and p is the number of array elements. Step 5: the eigenvalue corresponding to the noise is selected in the eigenvalue corresponding to the noise , d is the number of eigenvalues corresponding to the noise, and the expression is: ； According to the number d of the characteristic values corresponding to the noise, the average value of the characteristic values of the noise is calculated ;​ ； Step 6: Reconstruct the covariance matrix, and its expression is: ； wherein is a diagonal matrix of the corrected eigenvalues, is the eigenvector matrix and R is the reconstructed covariance matrix.

6. The wideband array main lobe interference suppression method according to claim 1, characterized in that: The optimal weight vector in S5 has the following specific steps: Step 1: Determine the main lobe subspace according to the formula: ； ； wherein the main lobe region subspace is spanned by the column vectors of is the ith eigenvalue, is the corresponding eigenvector, is a diagonal matrix composed of the first eigenvalues, is a matrix composed of the eigenvectors corresponding to the eigenvalues in is a matrix composed of the eigenvectors corresponding to the eigenvalues in is composed of the first eigenvalues, is a matrix composed of the eigenvectors corresponding to the eigenvalues in is a matrix composed of the eigenvectors corresponding to the eigenvalues in is the steering vector of the desired direction ; Step 2: Determine the projection matrix E of the main lobe subspace through the main lobe subspace, and its expression is: ； Step 3: Finally, the optimal weight vector under the minimum variance criterion is calculated with the constraint that the projection of the weight vector in the main lobe subspace is invariant The expression is: ； The expression obtained by arranging the above formula is: ； where E is the projection matrix of the main lobe subspace, R is the modified signal covariance, is the optimal weight vector, is the modified signal covariance, is the matrix composed of the eigenvectors corresponding to the eigenvalues in the modified eigenvalue set, is the diagonal matrix composed of the modified eigenvalues.

7. The wideband array mainlobe jammer rejection method of claim 1, wherein: The specific steps of S7 are as follows: Step 1: The optimal weight vector at S5 is obtained The focused inverse transform is performed to obtain the frequency The optimal weight vector at S5 is obtained, which is expressed as: ； wherein is a focused matrix inverse transform, is an optimal weight vector, is a frequency optimal weight vector at Step 2: Calculate the new frequency domain output, and its expression is: ； wherein is the frequency domain output after removing the main-lobe interference, frequency at the optimal weight vector, is the frequency domain output; Step 3: Perform inverse Fourier transform on the new frequency domain output obtained in step 2 to obtain the final signal, and its expression is: 。

Images