1. A mainlobe interference suppression method based on feature oblique projection covariance matrix reconstruction is characterized by comprising the following steps:

(1a.) the noise covariance matrix of the echo data X is subjected to characteristic decomposition, and a main lobe interference signal subspace U is output_mSide lobe interference signal subspace U_pAnd the eigenvalue lambda of the main lobe interference signal_m；

(1b.) Using the main lobe interference signal subspace U_mAnd said sidelobe interfering signal subspace U_pConstructing an oblique projection matrix O, preprocessing the echo data X, and outputting a preprocessed target signal X_o；

(1c.) use of the eigenvalue λ of the main lobe interference signal_mReconstructing a target interference noise covariance matrix, eliminating main lobe interference and target signal components, and outputting a corrected covariance matrix

(1d.) use of the modified covariance matrixSolving the optimal weight vector W_optAnd outputting the preprocessed target signal X_oPassing through an optimal weight vector W_optAnd performing weighting processing to obtain an output signal Y.

2. The method of claim 1, wherein the noise covariance matrix of the echo data X is characterized by an output mainlobe interference signal subspace U_mSide lobe interference signal subspace U_pAnd the eigenvalue lambda of the main lobe interference signal_mThe specific method comprises the following steps:

(2a.) when the number of uniform linear array elements is M, the interference noise covariance matrix R_XThe expression of (a) is:

wherein λ is_iIs the characteristic value of the ith interference signal, i is 1, 2. u. of_iThe feature vector corresponding to the ith feature value; lambda_sFor interference signal subspace U_sCharacteristic value diagonal matrix of_nAs noise signal subspace U_nEigenvalue diagonal matrix of (1), interference signal subspace U_sDivision into main-lobe interference signal subspaces U_mSum sidelobe interferer subspace U_pThe expression is:

U_p＝U_s-U_m

wherein the noise signal subspace U_nSide lobe jamming signal subspace U_pSubspace U synthesized with target signal_paComprises the following steps:

U_pa＝[U_p,a(θ₀)]

wherein, a (theta)₀) As a guide vector of the target signal, theta₀Is the angle of incidence of the target signal.

3. The method of claim 2, said utilizing the main lobe interference signal subspace U_mAnd said sidelobe interfering signal subspace U_pConstructing an oblique projection matrix O, preprocessing the echo data X, and outputting a preprocessed target signal X_oThe specific method comprises the following steps:

(3a.) is along with U_paIn parallel direction in the main lobe interference signal subspace U_mObtaining a characteristic oblique projection matrix O;

wherein the content of the first and second substances,is U_mThe space of the orthogonal complements of (a),is U_paThe hermite matrix of (a) is,is U_mHermitian matrix of;

(3b.) the expression for the echo data X is:

wherein P is the number of interference signals, s₀Represented as a complex envelope form, s, of the target signal in the echo data_jComplex envelope shape expressed as jth interference signal in echo data XN is a noise signal in the echo data; theta_jIs the incident angle of the jth interference signal, j is 1, 2.

Performing characteristic oblique projection on the received data:

X_o＝OX

obtaining a preprocessed target signal X_oAnd output, X_oThe specific expression of (A) is as follows:

wherein, a (theta)₁) Is the steering vector of the 1 st target signal, theta₁Is the incident angle, s, of the 1 st interference signal₁In the form of a complex envelope of the 1 st interfering signal in the echo data,is a pre-processed noise signal.

4. The method of claim 3, wherein the using the eigenvalues λ of the mainlobe interference signal_mReconstructing a target interference noise covariance matrix, eliminating main lobe interference and target signal components, and outputting a corrected covariance matrixThe specific method comprises the following steps:

(4a.) the interference noise covariance matrix R_XThe expression of (a) is:

wherein U is an interference signal subspace U_sSum noise subspace U_nAnd, U ═ U_s,U_n]；U^HHermite matrix of U, a is eigenvalue diagonal matrix, and Λ ═ diag (λ)₁,λ₂,…,λ_m,…,λ_P,λ_P+1,…,λ_M) (ii) a The eigenvalue of the mainlobe interference signal is found by the following formula:

wherein the content of the first and second substances,correcting the noise characteristic value by using a correction method for the mean value of the noise characteristic value, and replacing the mean value of the noise characteristic value with the characteristic value corresponding to the noise signal, namely:

wherein the content of the first and second substances,for the corrected eigenvalues, the covariance matrix after correction at that timeComprises the following steps:

wherein the content of the first and second substances,for the modified eigenvalue diagonal matrix,

5. the method of claim 4, wherein the utilizing the modified covariance matrixSolving the optimal weight vector W_optAnd outputting the preprocessed target signal X_oPassing through an optimal weight vector W_optWeighting processing is carried out to obtain an output signal Y, and the specific method comprises the following steps:

(5a.) based on the corrected covariance matrixCalculating an optimal weight vector W_opt：

Wherein mu is a constant coefficient; lambda^-1An inverse matrix of the target interference noise covariance matrix;

(5b.) use of the optimal weight vector W_optAnd a preprocessed target signal X₀And obtaining and outputting an output signal Y, wherein the expression of the output signal Y is as follows:

wherein the content of the first and second substances,is W_optHermitian matrix of.

Background

With the continuous improvement of the informatization level, the working environment of modern information systems is increasingly complex, and a large amount of interference seriously hinders the functional realization of the information systems. In a complex electromagnetic environment, the radar adopts anti-interference means such as low side lobe, side lobe cancellation, Conventional Beam Forming (CBF) algorithm, Adaptive Beam Forming (ABF) and the like. Conventional adaptive beamforming algorithms are all directed to a scene in which an interference signal is incident from a side lobe, and interference can be effectively suppressed only when only side lobe interference exists in a space. However, for the main lobe interference, since the interference signal and the target signal have strong correlation in space, the main beam formed by the spatial domain interference resistance may be distorted, which results in the loss of the directional gain of the target signal and the reduction of the output signal-to-interference-and-noise ratio gain, thereby seriously affecting the performance of the main lobe interference resistance.

In order to effectively inhibit the main lobe interference, domestic and foreign scholars propose a series of solutions, but each method has application conditions. The adaptive polarization filtering algorithm cannot inhibit broadband interference signals or forwarding interference signals with different polarizations; the auxiliary array algorithm obtains narrow main lobe width at the cost of increasing the size of the antenna, and cannot be applied to a limited scene of the size of the radar array antenna; the Blocking Matrix (BMP) algorithm reduces the system freedom of the array antenna; the feature projection (EMP) algorithm is high in calculation complexity and has peak deviation, when main lobe interference is suppressed, a part of target signals still have loss, and although the EMP improvement algorithm reduces the calculation amount, the problem of target signal loss still cannot be solved.

In order to solve the problems of large calculation amount and peak shift in the EMP algorithm, a feature projection covariance matrix reconstruction (EMP-CMR) algorithm is proposed by scholars. This algorithm allows the signal to be nulled no longer in the main lobe and the target gain is preserved. However, after the target signal is subjected to the feature projection processing, a certain loss exists, and the smaller the interval between the main lobe interference and the target angle is, the larger the target signal loss is.

Therefore, a method applying a feature projection covariance matrix reconstruction (EMP-CMR) algorithm is needed to suppress the main lobe interference and reduce the loss of the target signal after the algorithm processing.

Disclosure of Invention

In view of the above, the invention provides a main lobe interference suppression method based on feature oblique projection covariance matrix reconstruction, which can effectively suppress main lobe interference and reduce loss of a target signal, aiming at the problem that the target signal has a certain loss due to the existing EMP algorithm and the improved algorithm EMP-CMR thereof.

In order to achieve the purpose, the technical scheme of the invention is as follows:

a mainlobe interference suppression method based on feature oblique projection covariance matrix reconstruction comprises the following steps:

(1a.) the noise covariance matrix of the echo data X is subjected to characteristic decomposition, and a main lobe interference signal subspace U is output_mSide lobe interference signal subspace U_pAnd the eigenvalue lambda of the main lobe interference signal_m。

(1b.) use of the Main lobe interference Signal subspace U_mSum sidelobe interferer subspace U_pConstructing an oblique projection matrix O, preprocessing the echo data X, and outputting a preprocessed target signal X_o。

(1c.) use of the eigenvalue λ of the main lobe interference signal_mReconstructing a target interference noise covariance matrix, eliminating main lobe interference and target signal components, and outputting a corrected covariance matrix

(1d.) use of the modified covariance matrixSolving the optimal weight vector W_optAnd outputting the preprocessed target signal X_oPassing through an optimal weight vector W_optAnd performing weighting processing to obtain an output signal Y.

Further, the noise covariance matrix of the echo data X is subjected to characteristic decomposition, and a main lobe interference signal subspace U is output_mSide lobe interference signal subspace U_pAnd the eigenvalue lambda of the main lobe interference signal_mThe specific method comprises the following steps:

(2a.) when the number of uniform linear array elements is M, the interference noise covariance matrix R_XThe expression of (a) is:

wherein λ is_iIs the characteristic value of the ith interference signal, i is 1, 2. u. of_iThe feature vector corresponding to the ith feature value; lambda_sFor interference signal subspace U_sCharacteristic value diagonal matrix of_nAs noise signal subspace U_nEigenvalue diagonal matrix of (1), interference signal subspace U_sDivision into main-lobe interference signal subspaces U_mSum sidelobe interferer subspace U_pThe expression is:

U_p＝U_s-U_m

wherein the noise signal subspace U_nSide lobe jamming signal subspace U_pSubspace U synthesized with target signal_paComprises the following steps:

U_pa＝[U_p,a(θ₀)]

wherein, a (theta)₀) As a guide vector of the target signal, theta₀Is the angle of incidence of the target signal.

Further, a main lobe interference signal subspace U is utilized_mSum sidelobe interferer subspace U_pConstructing an oblique projection matrix O, preprocessing the echo data X, and outputting a preprocessed target signal X_oThe specific method comprises the following steps:

(3a.) is along with U_paParallel directionIn the main lobe interference signal subspace U_mObtaining a characteristic oblique projection matrix O;

wherein the content of the first and second substances,is U_mThe space of the orthogonal complements of (a),is U_paThe hermite matrix of (a) is,is U_mHermitian matrix of.

(3b.) the expression for the echo data X is:

wherein P is the number of interference signals, s₀Represented as a complex envelope form, s, of the target signal in the echo data_jRepresenting the complex envelope form of the jth interference signal in the echo data X, and n is a noise signal in the echo data; theta_jJ is the incident angle of the jth interference signal, 1, 2.

Performing characteristic oblique projection on the received data:

X_o＝OX

obtaining a preprocessed target signal X_oAnd output, X_oThe specific expression of (A) is as follows:

wherein, a (theta)₁) Is the steering vector of the 1 st target signal, theta₁Is the incident angle, s, of the 1 st interference signal₁For the 1 st in the echo dataThe complex envelope form of the individual interfering signals,is a pre-processed noise signal.

Further, the characteristic value lambda of the main lobe interference signal is utilized_mReconstructing a target interference noise covariance matrix, eliminating main lobe interference and target signal components, and outputting a corrected covariance matrixThe specific method comprises the following steps:

(4a.) interference noise covariance matrix R_XThe expression of (a) is:

wherein U is an interference signal subspace U_sSum noise subspace U_nAnd, U ═ U_s,U_n]；U^HHermite matrix of U, a is eigenvalue diagonal matrix, and Λ ═ diag (λ)₁,λ₂,…,λ_m,…,λ_P,λ_P+1,…,λ_M) (ii) a The eigenvalue of the mainlobe interferer is found by:

wherein the content of the first and second substances,correcting the noise characteristic value by using a correction method for the mean value of the noise characteristic value, and replacing the mean value of the noise characteristic value with the characteristic value corresponding to the noise signal, namely:

wherein the content of the first and second substances,for the corrected eigenvalues, the covariance matrix after correction at that timeComprises the following steps:

wherein the content of the first and second substances,for the modified eigenvalue diagonal matrix,

further, the corrected covariance matrix is utilizedSolving the optimal weight vector W_optAnd outputting the preprocessed target signal X_oPassing through an optimal weight vector W_optWeighting processing is carried out to obtain an output signal Y, and the specific method comprises the following steps:

(5a.) based on the corrected covariance matrixCalculating an optimal weight vector W_opt：

Wherein mu is a constant coefficient; lambda^-1Is the inverse of the target interference noise covariance matrix.

(5b.) use of the optimal weight vector W_optAnd a preprocessed target signal X₀And obtaining and outputting an output signal Y, wherein the expression of the output signal Y is as follows:

wherein the content of the first and second substances,is W_optHermitian matrix of.

Has the advantages that: the method is applied to radar anti-interference, and the characteristic oblique projection covariance matrix reconstruction (EOMP-CMR) algorithm used by the method can effectively solve the problem of projection loss of a target signal, so that a better main lobe interference resisting effect is obtained, and the method belongs to a steady anti-interference method. Aiming at the problem that the target signal has certain loss due to the EMP algorithm and the improved EMP-CMR algorithm, the method utilizes the trained snapshot number to solve the covariance matrix, and carries out characteristic decomposition on the covariance matrix to determine the interference signal subspace; and then constructing the basis of an oblique projection matrix through a target signal guide vector and an interference signal subspace, and performing self-adaptive beam forming to realize effective suppression of main lobe interference. The algorithm used by the invention can simultaneously reduce the loss of the target signal when the main lobe interference is suppressed, and the algorithm does not lose the degree of freedom of the system.

Drawings

Fig. 1 is a signal processing flow diagram of the method of the present invention.

Fig. 2 (a), (b), (c) and (d) correspond to the array patterns of the method of the invention compared to the patterns of the Conventional Beam Forming (CBF), Adaptive Beam Forming (ABF), eigenprojection preprocessing (EMP) and eigenprojection covariance matrix reconstruction (EMP-CMR) algorithms, respectively.

FIG. 3 is a comparison graph of the variation of the output signal-to-interference-and-noise ratio (SINR) with the number of snapshots between the method of the present invention and other spatial domain anti-interference methods.

FIG. 4 is a graph comparing the output signal-to-interference-and-noise ratio (SINR) with the signal-to-noise ratio (SNR) of a target signal in the method of the present invention with other spatial domain interference rejection methods.

Detailed Description

The invention is described in detail below by way of example with reference to the accompanying drawings.

The invention provides a main lobe interference suppression method based on characteristic oblique projection covariance matrix reconstruction, wherein the processing flow is shown as figure 1, and the method comprises the following specific steps:

step one, carrying out characteristic decomposition on a noise covariance matrix of echo data X, and outputting a main lobe interference signal subspace U_mSide lobe interference signal subspace U_pAnd the eigenvalue lambda of the main lobe interference signal_mThe specific method comprises the following steps:

when the number of the uniform linear array elements is M, the interference noise covariance matrix R_XThe expression of (a) is:

wherein λ is_iIs the characteristic value of the ith interference signal, i is 1, 2. u. of_iThe feature vector corresponding to the ith feature value; lambda_sFor interference signal subspace U_sCharacteristic value diagonal matrix of_nAs noise signal subspace U_nEigenvalue diagonal matrix of (1), interference signal subspace U_sDivision into main-lobe interference signal subspaces U_mSum sidelobe interferer subspace U_pThe expression is:

U_p＝U_s-U_m

wherein the noise signal subspace U_nSide lobe jamming signal subspace U_pSubspace U synthesized with target signal_paComprises the following steps:

U_pa＝[U_p,a(θ₀)]

wherein, a (theta)₀) As a guide vector of the target signal, theta₀Is the angle of incidence of the target signal

Step two, utilizing the main lobe interference signal subspace U_mAnd said sidelobe interfering signal subspace U_pConstructing an oblique projection matrix O, preprocessing the echo data X, and outputting a preprocessed target signal X_oThe specific method comprises the following steps:

according to the known definition of oblique projection, the interference signal and the target signal are both directedAre not the same, so that the subspace U_paAnd subspace U_mThere is no intersection.

Along and U_paIn parallel direction, in the main lobe interference signal subspace U_mObtaining a characteristic oblique projection matrix O;

wherein the content of the first and second substances,is U_mThe space of the orthogonal complements of (a),is U_paThe hermite matrix of (a) is,is U_mHermitian matrix of.

Wherein, the expression of the echo data X is:

wherein P is the number of interference signals, s₀Represented as a complex envelope form, s, of the target signal in the echo data_jRepresenting the complex envelope form of the jth interference signal in the echo data X, and n is a noise signal in the echo data; theta_jJ is the incident angle of the jth interference signal, 1, 2.

Performing characteristic oblique projection on the received data:

X_o＝OX

obtaining a preprocessed target signal X_oAnd output, X_oThe specific expression of (A) is as follows:

wherein, a (theta)₁) Is the steering vector of the 1 st target signal, theta₁Is the incident angle, s, of the 1 st interference signal₁In the form of a complex envelope of the 1 st interfering signal in the echo data,is a pre-processed noise signal.

From the above formula, it can be seen that after the oblique projection matrix processing, the target signal is retained while the main lobe interference component is removed and the side lobe interference component is retained, and compared with the EMP and EMP-CMR algorithms, the oblique projection preprocessing has no loss of the target signal.

Step three, utilizing the characteristic value lambda of the main lobe interference signal_mReconstructing a target interference noise covariance matrix, eliminating main lobe interference and target signal components, and outputting a corrected covariance matrixThe specific method comprises the following steps:

since the target signal in the processed signal is not cancelled, in order to reduce the influence on the interference suppression performance in the subsequent adaptive beamforming, when the covariance is reconstructed, the eigenvalue corresponding to the eigenvector of the target signal needs to be modified, and the eigenvalue and the main lobe interference are merged into the noise subspace. Namely:

in the above formula, the first and second carbon atoms are,andand the characteristic values respectively represent the corrected target signal and the characteristic value corresponding to the main lobe interference signal, and similarly, in order to reduce the influence of noise disturbance on the adaptive algorithm, the corrected characteristic values are as follows:

wherein the content of the first and second substances,is the corrected covariance.

From step one, the interference noise covariance matrix R_XThe expression of (a) is:

wherein U is an interference signal subspace U_sSum noise subspace U_nAnd, U ═ U_s,U_n]；U^HHermite matrix of U, a is eigenvalue diagonal matrix, and Λ ═ diag (λ)₁,λ₂,…,λ_m,…,λ_P,λ_P+1,…,λ_M)。

Covariance matrix at this timeIs composed of

U＝[U_s,U_n]

Wherein, U_sRepresenting the interference signal subspace, U_nWhich represents the noise subspace, is,for the new eigenvalue diagonal matrix, U is the sum of the interference signal subspace and the noise subspace.

Step four, utilizing the corrected covariance matrixSolving the optimal weight vector W_optAnd output, will pre-treatProcessed target signal X_oPassing through an optimal weight vector W_optWeighting processing is carried out to obtain an output signal Y, and the specific method comprises the following steps:

according to the corrected covariance matrixCalculating an optimal weight vector W_opt：

Wherein mu is a constant coefficient; lambda^-1Is the inverse of the target interference noise covariance matrix.

As can be seen from the above equation, there is no target signal and main lobe interference, so the directional pattern does not generate main beam distortion. Also, the operation complexity is low because the matrix inversion process is not included.

Using the optimal weight vector W_optAnd a preprocessed target signal X₀And obtaining and outputting an output signal Y, wherein the expression of the output signal Y is as follows:

wherein the content of the first and second substances,is W_optHermitian matrix of.

Therefore, the main lobe interference signal and the side lobe interference signal are both suppressed, and the target signal is reserved.

In order to verify the main lobe interference suppression method based on the feature oblique projection covariance matrix reconstruction, two groups of simulation experiments are carried out to deeply analyze the main lobe interference suppression performance of the method (EOMP-CMR). The simulation parameters of the directional diagram contrast analysis are set in the table 1, and the simulation parameters of the three interference signals in the directional diagram contrast analysis are set in the table two, wherein the noise is white gaussian noise.

Table 1 simulation parameter settings

Table 2 interference signal simulation parameter settings

In fig. 2 (a) (b) (c) (d), the proposed method is based on the comparison of the array pattern of the eigen-slant projection covariance matrix reconstruction (EOMP-CMR) with the patterns of the Conventional Beamforming (CBF), Adaptive Beamforming (ABF), eigen-projection preprocessing (EMP), and eigen-projection covariance matrix reconstruction (EMP-CMR) algorithms, respectively. From (a), it can be seen that Adaptive Beamforming (ABF) generates nulls in the main lobe interference direction, which causes the problem that the main beam is severely deformed and the side lobe level is increased; from the graphs (b) and (c), it can be seen that the main beam of the feature projection preprocessing (EMP) and feature projection covariance matrix reconstruction (EMP-CMR) algorithms is not distorted, but the peak thereof is obviously shifted, i.e. the maximum gain is not in the target direction, so that the existing methods have problems. As can be seen from the diagram (d), the array pattern of the method (EOMP-CMR) provided herein is substantially the same as that of conventional beamforming, and the problems of main lobe distortion and side lobe level increase of the array pattern under main lobe interference are effectively solved.

In order to analyze the main lobe interference suppression performance of the proposed algorithm, simulation analysis is performed on the output signal-to-interference-and-noise ratio (SINR) of the proposed algorithm, and the variation relationship between the SINR and the fast beat number is analyzed, assuming that the fast beat number is 10-100, the SNR is 15dB, other simulation conditions are unchanged, and the Monte Carlo frequency is 2000.

Fig. 3 simulates the output signal to interference plus noise ratio SINR of the proposed method EOMP-CMR and compares it with the output SINR of Conventional Beamforming (CBF), Adaptive Beamforming (ABF), feature projection preprocessing (EMP) and feature projection covariance matrix reconstruction (EMP-CMR) algorithms. As can be seen from the results in the figure, the output SINR of EOMP-CMR is the highest, and is superior to all other methods. Because the conventional CBF does not carry out interference suppression, the output SINR is the minimum, the EMP-CMR and the ABF can effectively suppress the main lobe/side lobe interference and can also output higher SINR, but because a target signal is also suppressed to a certain extent, the output SINR of the two methods is not as high as that of the EOMP-CMR.

In order to further analyze the main lobe interference suppression performance of the proposed algorithm EOMP-CMR, simulation analysis is now performed on the output signal-to-interference-and-noise ratio SINR of the proposed algorithm, and the variation relationship between the output signal-to-interference-and-noise ratio SINR of the proposed algorithm and the target signal SNR is analyzed, assuming that the number of snapshots is 100, the signal-to-noise ratio SNR of the target signal is 0dB to 15dB, stepping is 1dB, other simulation conditions are unchanged, and the Monte Carlo frequency is 2000. As can be seen from fig. 4, as the SNR of the target signal increases, the SINR of the output signal of each algorithm increases, but the SINR of the proposed algorithm EOMP-CMR is the largest and higher than that of the algorithms EMP, EMP-CMR, ABF, and the like.

According to the embodiment, the EOMP-CMR algorithm has better main lobe interference suppression performance.

In summary, the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.