A method for array radar anti-deceptive jamming based on blind source separation

Abstract

The present invention discloses an array radar anti-deceptive interference method based on blind source separation. Based on the blind source separation technology, the present invention combines the wave arrival angle information of the received signal, that is, recovers the multi-source signal waveform of the excitation array antenna through blind source separation, which can effectively realize the separation of interference and target, and then uses the interference identification technology to effectively extract the target signal. This method can perform interference identification on each signal channel after blind source separation, and adaptively suppress deceptive interference. From the simulation results, the present invention has a significant anti-interference effect on new deceptive interference such as dense false target interference and forwarding slice interference.

Description

Technical Field

[0001] The present invention relates to the technical field of radar anti-interference signal processing, and in particular to an array radar anti-deceptive interference method based on blind source separation. Background Art

[0002] Array signal processing is a very important branch of signal processing. It has been widely used in radar, sonar, communication, navigation, voice signal processing, electronic countermeasures and other fields. For example, radar radiates energy in all directions in space by emitting electromagnetic waves, and then receives reflected signals from target objects, while suppressing various clutter and interference in the radiation environment. With the rapid development of array radar technology, various interference technologies have emerged, such as passive interference, active interference, and new types of false target deception interference. These interferences seriously reduce the target detection and tracking performance of array radar. Therefore, how to effectively reduce the impact of interference on radar is a problem that must be solved at present.

[0003] Blind source separation refers to the process of separating and recovering relatively independent source signals from observed multi-source mixed signals. The so-called "blind" means that little is known about the source signals and the mixing process, and the only available information is the observed mixed signal. These source signals contain useful information, but they cannot be observed directly. Only by separating them from the mixed signal through blind signal processing can the extraction of useful signals be achieved. Blind source signal separation plays an increasingly important role in the field of digital signal processing, and has been widely used in communications, speech processing, seismic exploration, biomedicine, image processing, radar, and economic data analysis.

[0004] With the improvement of jammer performance, the modern battlefield environment is becoming more and more complex. Since the existing blind source separation method does not utilize the known prior knowledge of array antennas, it is difficult to guarantee the algorithm accuracy and convergence in the actual interference environment, and it is difficult to deal with coherent deceptive interference. Summary of the invention

[0005] In view of this, the present invention provides an array radar anti-deceptive interference method based on blind source separation. The present invention makes full use of the known prior knowledge of the array antenna structure, directly converts the mixing matrix and unmixing matrix estimation problem into the time-space parameter estimation problem of multi-source signals, and assists in constructing the unmixing matrix through the time-space parameter estimation of the source signal to achieve effective decomposition of the target signal and the interference, and finally effectively identifies the deceptive interference through the interference signal identification technology to achieve the purpose of anti-interference. The present invention can not only suppress the influence of the deceptive interference signal on the target signal and achieve anti-deceptive interference, but also separate the interference signal from the mixed signal, and further analyze the identified deceptive interference signal, providing information support for the selection of radar anti-interference strategy and optimization of working mode.

[0006] An array radar anti-deceptive jamming method based on blind source separation estimates the source parameters of the radar receiving signal covariance matrix to obtain the number of source signals; the radar receiving signal is pre-whitened using the number of source signals, and the pre-whitened signal is unmixed to obtain multiple separated signals; the center distance characteristic parameter of each separated signal is obtained, and the signal corresponding to the minimum center distance characteristic parameter is taken as the target signal.

[0007] Preferably, the MUSIC algorithm or the Capon algorithm is used to estimate the source parameters.

[0008] Preferably, the specific algorithm of the MUSIC algorithm to obtain the number of source signals is:

[0009] use Perform spectrum peak search to obtain the source signal quantity Q, where the search range is: The angle corresponding to the spectrum peak is the signal incident direction. The orthogonal projection matrix of the signal subspace formed by the eigenvectors corresponding to the Q principal eigenvalues ​​of the covariance matrix of the received signal; is the element-level array manifold vector, and H is the complex conjugate transpose operation.

[0010] Preferably, the specific method of the pre-whitening treatment is:

[0011] According to the number of source signals, the covariance matrix is ​​eigen-decomposed to obtain the eigenvalues, eigenvectors and noise powers corresponding to the source signals, and then a pre-whitening matrix is ​​constructed; the pre-whitening matrix is ​​multiplied with the radar receiving signal to realize pre-whitening processing.

[0012] Preferably, using Implement eigendecomposition of the covariance matrix, where R sub is the covariance matrix of the received signal, Q is the number of source signals, λ i is the eigenvalue corresponding to the source signal, u i isi The corresponding eigenvector, H is the complex conjugate transpose operation, is the noise power, and L is the number of sub-arrays.

[0013] Preferably, using Construct a pre-whitening matrix, where is the noise power, λ Q is the eigenvalue corresponding to the Qth source signal, u Q is Q The corresponding feature vector.

[0014] Preferably, the specific method of the demixing process is:

[0015] The delay covariance matrix of the pre-whitening processed signal is obtained, and eigendecomposition is performed to obtain an unmixing matrix; the unmixing matrix is ​​multiplied by the pre-whitening processed signal to implement unmixing processing.

[0016] Preferably, the specific method for determining the target signal is:

[0017] For each separated signal, short-time Fourier transform is used to perform time-frequency analysis on the separated signal to form a two-dimensional grayscale image; the scale center of gravity of the two-dimensional grayscale image is obtained; the scale center of gravity is used to obtain the central moment characteristic parameter under the set order; the central moment characteristic parameters corresponding to all separated signals are arranged from large to small, and the signal corresponding to the smallest center distance characteristic parameter is the target echo signal.

[0018] Preferably, the set order value is 4.

[0019] Preferably, using Get the fourth-order central moment characteristic parameters of the image, where x c and c is the scale centroid, A(i,j) is the two-dimensional grayscale image after time-frequency distribution preprocessing in the i-th row and j-th column, and I and J are the row and column dimensions of the two-dimensional grayscale image after time-frequency distribution preprocessing, respectively.

[0020] Beneficial effects:

[0021] (1) Compared with the existing technology, the array radar anti-deceptive jamming method based on blind source separation proposed by the present invention can make full use of prior knowledge such as array structure to obtain the target and interference angle parameters, thereby effectively separating the target signal and interference, and performing interference identification on each signal channel after blind source separation, and adaptively suppressing deceptive jamming. From the simulation results, the present invention has a significant anti-jamming effect on new deceptive jamming such as dense false target jamming and forwarding slice jamming.

[0022] (2) The present invention adopts a pre-whitening method to pre-process the radar receiving signal, which can effectively reduce the correlation between receiving channels and reduce the impact of system noise.

[0023] (3) The present invention constructs an unmixing matrix based on the pre-whitened signal, which can achieve accurate separation of the target and the interference and achieve a significant anti-interference effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 The array structure model studied in the present invention;

[0025] Figure 2 It is the single channel time domain waveform of radar receiving signal;

[0026] Figure 3 is the DOA parameter estimation result of multi-source signals;

[0027] Figure 4 It is the pulse compression result of single channel of radar receiving signal;

[0028] Figure 5 It is the time domain waveform of each separation channel of blind source separation;

[0029] Figure 6 The time-frequency analysis results of each separation channel based on interference identification;

[0030] Figure 7 The pulse pressure result of the target echo signal channel that is correctly identified;

[0031] Figure 8 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0032] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0033] The present invention provides an array radar anti-deceptive jamming method based on blind source separation. Since the new deceptive jamming pattern is complex, it has a serious performance impact on the tracking and detection of the radar, and the traditional anti-jamming method cannot effectively suppress such interference. Based on the blind source separation technology, the present invention combines the wave arrival angle information of the received signal (that is, restores the multi-source signal waveform of the excitation array antenna through blind source separation) to effectively separate the interference and the target, so that the interference signal and the target signal can be basically distinguished; and then the interference identification technology (that is, using time-frequency analysis) is used to effectively extract the target signal to achieve the separation of interference and the target.

[0034] The present invention first estimates the time and space parameters of the radar received signal, uses the estimated information to assist the blind source separation algorithm to construct an unmixing matrix, and separates the received sub-array output signals to obtain independent source signal channels. The separated channels include target echo signals and deceptive interference signals. Then, STFT time-frequency analysis is performed on each separated channel, and the 4×4-order central moment characteristic parameters are calculated. The interference signal and target echo are identified according to the size of the characteristic parameters. Finally, the pulse compression result of the target echo is output as the anti-interference result. The processing flow of the present invention is as follows: Figure 8 As shown, the specific steps include:

[0035] Step 1: Establish a radar array receiving signal model to obtain multi-channel radar data;

[0036] The planar array composed of N omnidirectional array elements is located on the YOZ plane, such as Figure 1 The first array element is located at the origin of the coordinate system and is the reference array element. The coordinates of the nth array element are (y n ,z n ), the array element spacing is d. The array is divided into L sub-arrays, which can be overlapping or non-overlapping sub-arrays. There are Q far-field narrow-band radiation sources, where the azimuth and elevation angles of the kth (k = 1, ..., Q) source signal are The received signal, that is, the subarray output can be expressed as

[0037] X sub (t) = A sub S(t)+N(t) (1)

[0038] Among them, X sub (t) = [x 1 (t),x 2 (t),…,x L (t)] T , is a known value, x L (t) is the output of the Lth sub-matrix;

[0039] is the subarray-level array manifold vector; is the element-level array manifold vector.

[0040] λ is the wavelength;

[0041] N(t)=[n 1 (t),…,n L ] T is additive white Gaussian noise, which is independent of the received signal. S(t) is the source signal.

[0042] The covariance matrix of the received signal is:

[0043] R sub =E[X sub (t)X sub (t) H ] (2)

[0044] Step 2: Estimate the parameters of each signal source based on the prior information of the array structure;

[0045] Using the received signal covariance matrix obtained in step 1, the Capon algorithm or the MUSIC algorithm is used to complete the source parameter estimation. The MUSIC algorithm is a subspace decomposition algorithm, which uses the orthogonality between the noise subspace and the steering vector to perform super-resolution estimation of the signal arrival angle, which can obtain more accurate data. Therefore, this embodiment uses the subarray-level MUSIC algorithm to complete the source parameter estimation.

[0046] The spectrum estimation expression of subarray-level MUSIC is:

[0047]

[0048] Perform peak search according to the above formula, search range: The angle corresponding to the spectrum peak is the signal incident direction. Represents R sub The Q main eigenvalues ​​correspond to the eigenvectors that form the orthogonal projection matrix of the signal subspace. According to the number of spectrum peaks estimated by MUSIC, the number of source signals Q and the angle of each source signal can be obtained. These spatiotemporal parameters are used to assist in constructing the unmixing matrix to achieve effective decomposition of the target signal and interference.

[0049] Step 3: recover the multi-source signal waveform that excites the array antenna through blind source separation;

[0050] According to the source signal parameters obtained by spectral estimation in step 2, the covariance matrix R sub Perform eigendecomposition:

[0051]

[0052] In the formula, λ i is the eigenvalue corresponding to the source signal, u i is i The corresponding eigenvector, is the noise power. After obtaining the source signal subspace, the whitening matrix B can be constructed as follows:

[0053]

[0054] Using B to receive signal X sub (t) is pre-whitened, and the processed signal is

[0055] Z(t)=BXsub (t) = B(A sub S(t)+N(t))=BA sub S(t)=CS(t) (6)

[0056] In the above formula, since the signal subspace generated by B is orthogonal to the noise subspace, the noise part is eliminated after pre-whitening. C is the unitary matrix required to obtain the pre-whitened signal Z(t), and its delay covariance matrix can be expressed as

[0057] R Z (τ)=E{Z(t)Z H (t-τ)} (7)

[0058] Where E{·} refers to the mathematical expectation operation, H is the complex conjugate transpose operation, τ is the time delay component and τ>0. Choose a suitable time delay τ so that the matrix have different eigenvalues. Then Perform eigendecomposition:

[0059]

[0060] In the above formula, Λ is the eigenvalue matrix, and the eigenvector matrix C obtained after eigendecomposition is 0 That is, the estimated matrix of the unitary matrix C is the unmixing matrix, and the estimated matrix C 0 The unitary matrix C satisfies the following relationship:

[0061] C 0 =CPD (9)

[0062] In the above formula, P is in the form of a general permutation matrix, and D is in the form of a general diagonal matrix with non-zero diagonals. That is, the estimated matrix C 0 It can be regarded as a linear permutation matrix of the unitary matrix C. Get the unmixing matrix C 0 Afterwards, the whitened signal can be separated using the unmixing matrix to obtain the output signals of the blind source separation system, a total of Q.

[0063]

[0064] Step 4: Use the separated signal waveform to identify the interference signal and the target echo;

[0065] The present invention uses short-time Fourier transform to perform time-frequency analysis on the Q output signals separated in step 3. For each output signal: before feature extraction, STFT (t, f) is first preprocessed to convert its value range to the value range [0, 255] of the grayscale image, as shown in the following formula:

[0066]

[0067] In the above formula, STFT(t,f) is the short-time Fourier transform of the output signal Y(t), A is the two-dimensional grayscale image after time-frequency distribution preprocessing, and then the feature parameters are extracted from it.

[0068] The scale centroid of the grayscale image A is defined as:

[0069] C=[x c ,y c ] (12)

[0070] in, I and J are the row and column dimensions of A, that is, the number of points in the time domain and frequency domain after the short-time Fourier transform of Y(t). A(i,j) is the i-th row and j-th column element of the two-dimensional grayscale image after time-frequency distribution preprocessing.

[0071] The 4th order central moment of image A is defined as:

[0072]

[0073] Among them, x c and c is the scale center of gravity. The central moment of the image reflects the degree to which the pixel in the image deviates from the center. Here, the 4×4 order central moment feature parameter μ of A is extracted. 44 .

[0074] Because the center moment characteristic parameters of different interference types vary greatly and each has a fixed size range, the target echo and interference can be identified by comparing the center distance characteristic parameters of the separated signal. Generally speaking, the largest center moment characteristic parameter in the separated signal corresponds to the suppression interference signal, the smallest center moment characteristic parameter corresponds to the target echo signal, and the rest are deceptive interference signals. Therefore, the center distance characteristic parameters corresponding to all the separated signals obtained are arranged from large to small, and the signal corresponding to the smallest center distance characteristic parameter is the output target echo signal.

[0075] Example

[0076] In order to verify the array radar anti-deceptive jamming method based on blind source separation proposed in the present invention, target echo reception and interference suppression simulation are carried out. The simulation adopts a planar array, and the simulation parameters are shown in Table 1.

[0077] Table 1 Simulation parameter settings

[0078]

[0079]

[0080] The time domain waveform of a single receiving channel of the radar is as follows: Figure 2As shown in the figure, the power of forwarding slice jamming and dense false target jamming is relatively large. Figure 4 It can be seen that there are multiple peaks after the received data pulse compression, which drowns the target echo, making target detection difficult and requiring anti-deception interference processing. According to step 2, the array structure prior information is used to perform sub-array level super-resolution spectrum estimation MUSIC. Figure 3 The azimuth and elevation angles of the target echo and the deceptive interference can be obtained. The two interferences are incident from the main lobe and side lobe of the beam respectively. After blind source separation, the waveforms of each separation channel are as follows: Figure 5 As shown, the output order cannot be determined, and interference identification is required. According to step 4, interference identification is performed on the separated signal. Figure 6 The time-frequency analysis results of each separation channel are obtained, and then the fourth-order central moment characteristic parameters are extracted for interference identification. Figure 7 In order to correctly identify the pulse pressure result of the target echo signal channel, there is only one peak, which achieves the anti-interference effect.

[0081] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for array radar anti-deceptive interference based on blind source separation, characterized in that: Estimate the source parameters of the radar receiving signal covariance matrix to obtain the number of source signals; perform pre-whitening processing on the radar receiving signal using the number of source signals, and perform unmixing processing on the pre-whitening signal to obtain multiple separated signals; obtain the center distance characteristic parameter of each separated signal, and take the signal corresponding to the minimum center distance characteristic parameter as the target signal; wherein, use the MUSIC algorithm or the Capon algorithm to estimate the source parameters; The specific method of the unmixing process is: Obtaining the delay covariance matrix of the pre-whitened signal and performing eigendecomposition to obtain an unmixing matrix; multiplying the unmixing matrix with the pre-whitened signal to achieve unmixing processing; The specific method for determining the target signal is: For each separated signal, a time-frequency analysis is performed on the separated signal using a short-time Fourier transform to form a two-dimensional grayscale image; the scale centroid of the two-dimensional grayscale image is obtained; the central moment characteristic parameter under a set order is obtained using the scale centroid; the central moment characteristic parameters corresponding to all separated signals are arranged from large to small, and the signal corresponding to the smallest center distance characteristic parameter is the target echo signal; The set order value is 4; Get the fourth-order central moment characteristic parameters of the image, where x c and c is the scale centroid, and A(i,j) is the i-th row and j-th column element of the two-dimensional grayscale image after time-frequency distribution preprocessing.

2. The array radar anti-deceptive interference method based on blind source separation as claimed in claim 1, characterized in that: The specific algorithm of the MUSIC algorithm to obtain the number of source signals is: use Perform spectrum peak search to obtain the source signal quantity Q, where the search range is: The angle corresponding to the spectrum peak is the signal incident direction, where The orthogonal projection matrix of the signal subspace formed by the eigenvectors corresponding to the Q principal eigenvalues ​​of the covariance matrix of the received signal; is the array element level array popular vector, H is the complex conjugate transpose operation, are the azimuth and elevation angles of the source signal.

3. The array radar anti-deceptive interference method based on blind source separation as claimed in claim 1, characterized in that: The specific method of the pre-whitening treatment is: According to the number of source signals, the covariance matrix is ​​eigen-decomposed to obtain the eigenvalues, eigenvectors and noise powers corresponding to the source signals, and then a pre-whitening matrix is ​​constructed; the pre-whitening matrix is ​​multiplied with the radar receiving signal to realize pre-whitening processing.

4. The array radar anti-deceptive interference method based on blind source separation as claimed in claim 3, characterized in that: use Perform eigendecomposition on the covariance matrix, where R sub is the covariance matrix of the received signal, Q is the number of source signals, λ i is the eigenvalue corresponding to the source signal, u i is i The corresponding eigenvector, H is the complex conjugate transpose operation, is the noise power, and L is the number of sub-arrays.

5. The array radar anti-deceptive interference method based on blind source separation as claimed in claim 3, characterized in that: use Construct a pre-whitening matrix, where is the noise power, λ Q is the eigenvalue corresponding to the Qth source signal, u Q is Q The corresponding feature vector.

Images