Radar anti-jamming method and system based on sliding window multi-beam blind source separation

Abstract

The application provides a radar anti-interference method and system based on a sliding window multi-beam blind source separation, and the method comprises the following steps: interference sensing is performed by using array elements or subarray data of a phased array surface to obtain the angle of interference, the full array is divided into two windows, beam synthesis is performed according to the angle of interference sensing and the main working angle of the radar, then the beam domain data of the two windows are subjected to joint blind source separation to separate the interference and target signals, finally, the separated target signals are detected and target angle measurement is performed, so that interference suppression and high-precision target angle measurement are realized. The application solves the problems of angle measurement error caused by the anti-interference process, anti-interference when the subarray spacing cannot meet the sliding invariance, and the technical problem of the decline of sidelobe interference suppression performance caused by subarray dimension reduction.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing technology, and specifically to a radar anti-jamming method and system based on sliding window multi-beam blind source separation. Background Technology

[0002] With the increasing complexity of the electromagnetic environment in radar applications, the demand for radar anti-jamming capabilities is becoming more and more urgent. The ability to adapt to complex interference environments is crucial to the success or failure of a radar system. To improve radar performance in complex electromagnetic interference environments, anti-jamming measures such as frequency agility, ultra-low sidelobes, sidelobe concealment, adaptive sidelobe cancellation, adaptive beamforming, and blind source separation have been widely adopted. Adaptive sidelobe cancellation and adaptive beamforming suppress interference by exploiting the difference in spatial response between interference and target echoes, and by creating nulls in the radiation pattern in the interference direction through the interference covariance statistical matrix. However, when the interference is located within the main lobe beamwidth, the target direction vector is partially correlated with the interference characteristic vector, causing the beam in the target direction to fail to form effectively, and introducing a shift in the main lobe radiation pattern. For interference entering through the main lobe, researchers have proposed anti-jamming methods based on blind source separation theory. These methods utilize the statistical independence of interference and target echoes, and achieve target echo separation and extraction by estimating the system response matrix. Among these, the eigenma approximation joint diagonalization (JADE) algorithm based on fourth-order cumulants has relatively stable performance and is suitable for weak target, strong interference environments.

[0003] For example, the existing invention application CN109270499A, entitled "A Multi-Target Main Lobe Anti-Interference Method Based on Joint Diagonalization of Feature Matrix," first models multiple targets under main lobe interference; then, it uses JADE to separate the interference component from the target echo component; finally, according to the matched filtering principle, the target echo component obtained by blind source separation is passed through a matched filter to ultimately suppress the interference and achieve target detection. From the specific implementation of this prior art, it can be seen that this existing solution involves a multi-target blind source separation method to combat main lobe interference. However, radar target detection in complex electromagnetic interference environments not only involves interference suppression but also requires detection and parameter estimation of the signal after interference suppression. Especially when the interference is located within the main lobe, interference suppression often leads to distortion of the main lobe manifold, making it impossible to achieve angle measurement through amplitude envelope fitting after interference suppression. Furthermore, before the target angle is measured, it is impossible to know whether the interference is main lobe interference or side lobe interference, thus hindering the selection of a suitable method architecture. Therefore, it is necessary to study a method that can both resist interference and perform target angle measurement.

[0004] For example, the existing invention patent application document CN114114163A, entitled "A Method for Anti-Deception Jamming of Array Radar Based on Blind Source Separation," combines blind source separation technology with the angle of arrival (ADR) information of the received signal. Specifically, by recovering the multi-source signal waveform of the excitation array antenna through blind source separation, it can effectively separate the interference and the target. Then, interference identification technology is used to effectively extract the target signal. From the specific implementation of this prior art, it can be seen that this prior art utilizes prior data such as the array structure to obtain the ADR parameters of the target and the interference, thereby separating the target signal and the interference. Meanwhile, the existing invention application document CN111044979A, entitled "A Method for Main Lobe Interference Cancellation and Target Angle Estimation Based on Blind Source Separation," includes the following steps: 1. Receiving signals and interference data; 2. Obtaining the output data of the receiving antenna array; 3. Weighting the array output to obtain sum and difference beam data; 4. Repeating steps 1 and 2 for another antenna and performing blind source separation on the received data; 5. Extracting the separated interference data; 6. Using the extracted interference data to cancel interference in the obtained sum and difference beam data; 7. Using the sum and difference beam data after interference cancellation to perform sum and difference beam angle measurement to obtain the azimuth and elevation angles of the target. However, this existing technology is primarily applied to distributed radar. The blind source separation radar anti-jamming method based on sliding window subarrays has the ability to achieve interference suppression and angle measurement, but the subarray architecture requires the subarray spacing to have sliding invariance. Modern radar systems often use non-equidistant antenna arrays to reduce the cost of microwave systems, making it difficult to guarantee the consistency of the sliding window subarray spacing combination. In addition, when synthesizing subarrays, the fewer subarrays synthesized, the lower the computational load, but this corresponds to a narrower beam of the subarray, making it impossible to achieve high-performance suppression of far-field sidelobe interference.

[0005] In summary, existing technologies suffer from problems such as angle measurement errors during the anti-interference process, anti-interference issues when the subarray spacing cannot meet the sliding invariance requirement, and technical problems such as decreased sidelobe interference suppression performance due to subarray dimensionality reduction. Summary of the Invention

[0006] The technical problem to be solved by this invention is how to solve the problems of angle measurement error caused by the anti-interference process, the anti-interference problem when the subarray spacing cannot meet the sliding invariance, and the technical problem of the decline in sidelobe interference suppression performance caused by subarray dimensionality reduction in the prior art.

[0007] This invention solves the above-mentioned technical problems by employing the following technical solution: a radar anti-jamming method based on sliding window multi-beam blind source separation includes:

[0008] S1. Obtain full array data and sliding window subarray data;

[0009] S2. Based on the full array data, perform covariance estimation to obtain signal eigenvalues ​​and signal eigenvectors. Then, use the MUSIC algorithm to perform spatial spectrum estimation to obtain the angular distribution of all signal sources and thus detect signal interference.

[0010] S3. Construct an interference beamforming steering vector for each signal interference direction, and use the interference beamforming steering vector to perform interference beamforming on the sliding window subarray data. Construct a target beamforming steering vector for the target direction, and use the target beamforming steering vector to perform target beamforming on the sliding window subarray data to construct sliding window beam domain data.

[0011] S4. Perform joint blind source separation on the sliding window beam domain data, estimate the signal for each sliding window, perform joint preprocessing to obtain the joint covariance matrix, perform eigenvalue decomposition on the joint covariance matrix, construct the whitening matrix, and perform whitening preprocessing to obtain the whitened signal.

[0012] S5. Construct a joint cumulant matrix based on all whitened signals, obtain and eigendecompose the joint cumulant matrix to obtain joint cumulant eigenvalues ​​and cumulant eigenvectors, process them to obtain a joint feature matrix, and perform joint approximate diagonalization on the joint feature matrix to obtain a sliding window joint blind separation matrix.

[0013] S6. Using the sliding window joint blind separation matrix, the sliding window beam signal vector is separated to obtain the separated signal data, and the target signal and interference signal are obtained accordingly.

[0014] S7. Perform constant false alarm rate (CFAR) detection on the separated signal data path, perform angle measurement on the distance cell where the target is located, and calculate the phased array path difference to obtain the direction angle corresponding to the current phase angle.

[0015] This invention utilizes the array data of the phased array surface to sense the angle of interference. The entire array is divided into two windows, and beamforming is performed separately according to the interference angle and the radar's main operating angle. The beam domain data of the two windows are then used for joint blind source separation to separate the interference from the target signal. This allows for the detection of the separated target signal and target angle measurement, thus solving the angle measurement error caused by the anti-interference process. While achieving interference suppression, high-precision target angle measurement is obtained, effectively solving the problem that traditional anti-main lobe interference methods cannot measure angles.

[0016] In a more specific technical solution, step S1 includes:

[0017] S11. Based on the source signal and system model, represent the signal of an array with N receiving elements using the following logic:

[0018] X = [x1, x2, x3, ..., x N ]T

[0019]

[0020] Where M is the number of all signal sources received by the antenna, and satisfies:

[0021] M <N,

[0022] In the formula, a(θ)=[1,e j2πdsinθ / λ ,...,e j2π(N-1)dsinθ / λ ] T Let 'a' be the direction guide vector, 'a' be the direction vector, and 'e' be the natural constant. d is the spacing between the receiving channels, θ i Let i = 1, ..., M be the angle of the signal relative to the antenna normal, λ be the wavelength, and j be the m-th component. m ∈C 1×P Let 1 ≤ m ≤ M be the signal source. The signal source can be represented using the following logic:

[0023] j m =[1,exp(j2πf m / f s ),...,exp(j2πf m (P-1) / f s )]

[0024] Among them, f s f is the radar sampling frequency. m Let X be the frequency of the signal source, and let X be a complex matrix of N×P, where N is the spatial dimension, P is the fast time sampling dimension, and Γ is the noise.

[0025] S12. Divide the entire signal matrix into two groups of spatially diverse signal matrices by sliding windows with W array elements in each window:

[0026]

[0027]

[0028] This invention, by employing a sliding window beamforming system architecture, reduces the requirement for the sliding invariance of subarray spacing in anti-interference methods. Compared with existing technologies, this invention has a wider range of applications. This invention solves the anti-interference problem in existing technologies where the subarray spacing cannot satisfy sliding invariance.

[0029] In a more specific technical solution, step S2 includes:

[0030] S21. Estimate the covariance matrix of the full array data using the following logic: R N×N =E{X*XH}

[0031] Where H represents the conjugate transpose;

[0032] S22. Use the following logic to represent signal eigenvalues ​​and signal eigenvectors:

[0033]

[0034] Wherein, the eigenvalue λ i V is a positive real number, corresponding to the energy of the signal. i System characteristics corresponding to i signal sources;

[0035] S23. Arrange the signal characteristic values ​​in descending order according to the following logical order:

[0036]

[0037] The signal feature vectors corresponding to each signal feature value are as follows:

[0038]

[0039] S24. Construct the signal subspace S based on the signal feature vector. j and noise subspace S n :

[0040]

[0041]

[0042] The spatial spectrum estimate obtained from the MUSIC algorithm is as follows:

[0043]

[0044] S25. Obtain the angular distribution of all signal sources by processing the spatial spectrum estimate:

[0045]

[0046] In the formula, This represents the i-th estimated angle value.

[0047] The beam domain method based on interference sensing in this invention reduces the uncertainty that may be introduced in non-interference directions by introducing interference sensing information, thereby concentrating the system's spatial resources on anti-interference and achieving better anti-interference performance.

[0048] In a more specific technical solution, step S3 includes:

[0049] S31, the direction of interference for each signal: Construct the interference beam synthesis steering vector using the following logic:

[0050]

[0051] S32. Based on the interference beamforming steering vector, perform interference beamforming on the sliding window subarray data using the following logic to obtain the interference beamforming result:

[0052]

[0053]

[0054] S33. Using the following logic, construct the target beam synthesis steering vector for the target direction θ0:

[0055]

[0056] S34. Based on the following logic, use the target beamforming steering vector to perform target beamforming on the sliding window subarray data to obtain the target beamforming result:

[0057]

[0058]

[0059] S35. Based on the interference beamforming results and the target beamforming results, construct the sliding window beam domain data according to the following logic:

[0060] and

[0061] Where T represents transpose.

[0062] The sliding window beam domain method used in this invention forms a beam based on the results of interference sensing to achieve dimensionality reduction. This avoids the problem of decreased sidelobe interference suppression performance caused by subarray dimensionality reduction, making dimensionality reduction more accurate. At the same time, the beam can be pointed to various interferences, thereby avoiding the contradiction between far-field sidelobe interference and subarray size and number in the subarray method.

[0063] In a more specific technical solution, step S4 includes:

[0064] S41, Sliding window combined with whitening pretreatment

[0065] Based on the signal from each sliding window, the covariance matrix is ​​estimated as follows:

[0066] and

[0067] in, The vector formed by the elements of the i-th column of the first window. The vector formed by the elements of the i-th column of the second window;

[0068] S42. Based on the covariance matrix, construct the joint covariance matrix using the following logic:

[0069]

[0070] S43. For the joint covariance matrix R zz Perform eigenvalue decomposition and represent the joint covariance matrix using the following logic:

[0071] R zz =UΛU H

[0072] Where Λ is a diagonal matrix and the diagonal elements are eigenvalues;

[0073] S44. Sort the eigenvalues ​​in descending order: λ1≥λ2≥...,≥λ Mj+1 The eigenvectors corresponding to the eigenvalues ​​form the eigenma matrix U, where the eigenma matrix is ​​represented as U = [U1, U2, ..., U...]. Mj+1 ];

[0074] S45. Construct a whitening matrix and perform whitening preprocessing. Construct a joint whitening processing matrix using the following logic:

[0075]

[0076] The following logic is used to whiten the data from the two sliding window beam domains respectively, thereby obtaining the first whitened signal. and the second whitening signal

[0077] In a more specific technical solution, step S5 includes:

[0078] S51. The first sliding window signal vector is represented using the following logic. The fourth-order cumulant:

[0079]

[0080] And construct the first fourth-order cumulant matrix based on the fourth-order cumulants:

[0081]

[0082] The second sliding window signal vector is processed using the aforementioned logic in step S51. Construct the second and fourth order cumulant matrix

[0083] S52. Based on the first and second fourth-order cumulant matrices, construct a joint fourth-order cumulant matrix using the following logic:

[0084]

[0085] S53. Perform eigenvalue decomposition on the joint fourth-order cumulant matrix, and use the following logic to represent the joint cumulant eigenvalues ​​and cumulant eigenvectors:

[0086]

[0087] Where, λ i For the i-th eigenvalue, V i The corresponding feature vector;

[0088] S54, Take the first M j +1 joint cumulant eigenvalues ​​and cumulant eigenvectors are used to construct an eigenvector matrix using the following logic:

[0089]

[0090] S55, The (M) in the eigenvector matrix j +1) 2 ×1 dimensional column vector V i Arranged by column as (M) j +1)×(M j +1) characteristic matrix G i Based on this, the joint characteristic matrix is ​​obtained:

[0091]

[0092] S56. Perform JADE processing based on joint approximate diagonalization on the joint feature matrix to obtain the sliding window joint blind separation matrix:

[0093]

[0094] in, For matrix D i =q H G i The element in row k and column l of q.

[0095] In a more specific technical solution, step S53 includes:

[0096] S531. Arrange the joint cumulant eigenvalues ​​in descending order of their modulus:

[0097]

[0098] S532, The cumulative eigenvectors corresponding to the joint cumulative eigenvalues ​​are:

[0099] In a more specific technical solution, in step S6, the sliding window joint blind separation matrix Q is used to process the sliding window signal vector undergoing joint whitening. and Perform signal separation: This is used to separate the target signal from the interference signal.

[0100] In a more specific technical solution, step S7 includes:

[0101] S71. Results of signal separation: and Perform constant false alarm rate (CFAR) detection on each row;

[0102] S72. Using the following logic, perform phase-based angle measurement on the distance cell where the target is located:

[0103]

[0104] Where conj represents taking the conjugate, and ∠ represents taking the complex phase angle;

[0105] S73. Calculate the phased array path difference to obtain the direction angle corresponding to the phase angle:

[0106]

[0107] Where λ is the wavelength, B d The distance between the phase centers of the two sliding windows is denoted as .

[0108] In a more specific technical solution, the system includes:

[0109] The receiving and sliding window data module is used to acquire full array data and sliding window subarray data;

[0110] The interference sensing module is used to perform covariance estimation based on the full array data, obtain signal feature values ​​and signal feature vectors, and then use the MUSIC algorithm to perform spatial spectrum estimation to obtain the angular distribution of all signal sources, thereby sensing signal interference. The interference sensing module is connected to the receiving and sliding window data module.

[0111] The sliding window beam domain acquisition module is used to construct an interference beam synthesis steering vector for each signal interference direction, use the interference beam synthesis steering vector to perform interference beam synthesis on the sliding window subarray data, construct a target beam synthesis steering vector for the target direction, and use the target beam synthesis steering vector to perform target beam synthesis on the sliding window subarray data to construct sliding window beam domain data. The sliding window beam domain acquisition module is connected to the interference sensing module.

[0112] The sliding window beam domain data joint blind source separation module is used to estimate the signal for each sliding window separately, perform joint preprocessing to obtain the joint covariance matrix, perform eigenvalue decomposition on the joint covariance matrix to construct the whitening matrix, and perform whitening preprocessing to obtain the whitened signal. The sliding window beam domain data joint blind source separation module is connected to the receiving and sliding window data module.

[0113] The sliding window joint blind separation matrix estimation module is used to construct a joint cumulant matrix based on all whitening signals, obtain and eigendecompose the joint cumulant matrix to obtain joint cumulant eigenvalues ​​and cumulant eigenvectors, process them to obtain a joint eigenma matrix, and perform joint approximate diagonalization on the joint eigenma matrix to obtain the sliding window joint blind separation matrix. The sliding window joint blind separation matrix estimation module is connected to the sliding window joint whitening preprocessing module.

[0114] The sliding window joint signal separation module is used to separate the sliding window beam signal vector using the sliding window joint blind source separation matrix, thereby obtaining the separated signal data, and thus obtaining the target signal and interference signal. The sliding window joint signal separation module is connected to the joint blind source separation module.

[0115] The separation result detection and angle measurement module is used to perform CFAR detection on each line of the separated signal path data, perform angle measurement on the distance cell where the target is located, and calculate the phased array path difference to obtain the direction angle corresponding to the current phase angle. The separation result detection and angle measurement module is connected to the sliding window joint blind separation matrix estimation module.

[0116] Compared with the prior art, the present invention has the following advantages: The present invention uses the array data of the phased array surface to sense the angle of interference, divides the entire array into two windows and performs beam synthesis according to the interference angle and the radar main working angle, and performs joint blind source separation on the beam domain data of the two windows to separate the interference from the target signal, so as to complete the detection of the separated target signal and the target angle measurement, which solves the angle measurement error caused by the anti-interference process, and obtains high-precision target angle measurement while achieving interference suppression, effectively solving the problem that traditional anti-main lobe interference methods cannot measure angles.

[0117] This invention, by employing a sliding window beamforming system architecture, reduces the requirement for the sliding invariance of subarray spacing in anti-interference methods. Compared with existing technologies, this invention has a wider range of applications. This invention solves the anti-interference problem in existing technologies where the subarray spacing cannot satisfy sliding invariance.

[0118] The beam domain method based on interference sensing in this invention reduces the uncertainty that may be introduced in non-interference directions by introducing interference sensing information, thereby concentrating the system's spatial resources on anti-interference and achieving better anti-interference performance.

[0119] This invention employs a sliding window beam domain method to form a beam based on interference sensing results, achieving dimensionality reduction. This avoids the problem of decreased sidelobe interference suppression performance caused by subarray dimensionality reduction, resulting in more accurate dimensionality reduction. Simultaneously, the beam can be directed at various interference sources, thus avoiding the contradiction between far-field sidelobe interference and subarray size and number in the subarray method. This invention solves the technical problems of angle measurement errors during the anti-interference process and the inability of subarray spacing to meet sliding invariance in existing technologies, as well as the technical problems of decreased sidelobe interference suppression performance caused by subarray dimensionality reduction. Attached Figure Description

[0120] Figure 1 This is a flowchart of the radar anti-jamming method based on sliding window multi-beam blind source separation according to Embodiment 1 of the present invention;

[0121] Figure 2 This is a schematic diagram of data stream processing for the radar anti-jamming method based on sliding window multi-beam blind source separation in Embodiment 1 of the present invention;

[0122] Figure 3 This is a spatial structure diagram of a simulation implementation example of Embodiment 1 of the present invention;

[0123] Figure 4 This is a spectral distribution diagram of interference sensing in Embodiment 1 of the present invention;

[0124] Figure 5 This is a schematic diagram showing the distribution of each sliding window beam after multi-beam synthesis in Embodiment 1 of the present invention;

[0125] Figure 6 This is a schematic diagram of the blind source separation results obtained by two sliding windows in Embodiment 1 of the present invention;

[0126] Figure 7 This is a schematic diagram of the angular distribution of the target obtained from multiple experiments in Embodiment 1 of the present invention. Detailed Implementation

[0127] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0128] Example 1

[0129] like Figure 1 and Figure 2 As shown, the radar anti-jamming method based on sliding window multi-beam blind source separation provided by the present invention includes the following basic steps:

[0130] S1. Obtain full array data and sliding window subarray data;

[0131] In this embodiment, for an array with N receiving elements, the signal is represented as: X = [x1, x2, x3, ..., x N ] T It is determined by both the source signal and the system model, and can be expressed as:

[0132]

[0133] Where M is the number of all signal sources received by the antenna, and satisfies M < N,

[0134] a(θ) = [1, e j2πdsinθ / λ ,...,e j2π(N-1)dsinθ / λ ] T

[0135] like Figure 3 As shown, θ is the direction steering vector, d is the spacing between the receiving channels, and θ is the direction steering vector. i ,i=1,...,M is the angle of the signal relative to the antenna normal, λ is the wavelength, j m ∈C 1×P ,1≤m≤M is the signal source. In this embodiment, the signal source includes: an interference generator or a target, which can be represented as:

[0136] j m =[1,exp(j2πf m / f s ),...,exp(j2πf m (P-1) / f s )]

[0137] Among them, f s f is the radar sampling frequency. m Let be the frequency of the signal source. Then, the received signal X is an N×P complex matrix, where N is the spatial dimension, P is the fast-time sampling dimension, and Γ is the noise. In this embodiment, N is 16, P is 200, M is 4, and d is half the wavelength.

[0138] In this embodiment, the entire signal matrix is ​​divided into two spatially diverse signal matrices by sliding a window according to the spatial dimension:

[0139]

[0140]

[0141] In this embodiment, the value of W can be, for example, 15.

[0142] S2. Interference sensing based on MUSIC spatial spectrum estimation of full array data;

[0143] In this embodiment, since the sliding window reduces the effective array size, interference sensing is performed using the full array signal, specifically including: estimating the covariance matrix of the received full array signal.

[0144] R N×N =E{X*X H}

[0145] Where H represents the conjugate transpose.

[0146] Perform eigenvalue decomposition and express it in the form of eigenvalues ​​and eigenvectors:

[0147]

[0148] In the formula, its eigenvalue λ i V is a positive real number, corresponding to the energy of the signal. i This corresponds to the characteristics of i signal sources.

[0149] Arrange the eigenvalues ​​in descending order. The corresponding feature vectors are as follows: Constructing the signal subspace S j and noise subspace S n :

[0150]

[0151]

[0152] The spatial spectrum estimate obtained by the MUSIC algorithm is as follows:

[0153]

[0154] In this embodiment, the angular distribution of all signal sources can be obtained through spatial spectrum estimation.

[0155] like Figure 4 As shown, in this embodiment, M j =3, and various angle values ​​can be, for example: -10.77, 2.21 and 28.62.

[0156] S3. Obtain sliding window beam domain data based on sliding window subarray data;

[0157] In this embodiment, firstly, based on the interference angle value obtained in step S2, interference beamforming is performed in the sliding window space to obtain each interference direction. (i = 1, 2, ..., M) jIn this embodiment, the beamforming steering vector is constructed in three directions (-10.77, 2.21, and 28.62).

[0158]

[0159] And it is used to perform beamforming on sliding window data:

[0160]

[0161]

[0162] In this embodiment, target beamforming is performed by constructing a beamforming steering vector for the target direction θ0:

[0163]

[0164] It is used to perform beamforming on sliding window data. In this embodiment, the target direction θ0 can be set to, for example, 0.

[0165]

[0166]

[0167] Thus, sliding window beam domain data is constructed:

[0168] and

[0169] Where T represents transpose.

[0170] like Figure 5 As shown, in this embodiment, each sliding window constructs 3 beams pointing to interference and 1 beam pointing to a possible target, for a total of 4 beams, while a double sliding window corresponds to 8 beams.

[0171] S4. Perform joint blind source separation on sliding window beam domain data;

[0172] In this embodiment, a sliding window combined with whitening pretreatment is performed:

[0173] The covariance matrix is ​​estimated for each sliding window signal:

[0174] and

[0175] in, The vector formed by the elements of the i-th column of the first window. The vector formed by the elements of the i-th column of the second window.

[0176] Construct the joint covariance matrix

[0177] In this embodiment, for R zz Perform eigenvalue decomposition, denoted as R zz =UΛU H The form is λ1≥λ2≥...,≥λ2, where Λ is a diagonal matrix and the diagonal elements are eigenvalues, arranged in descending order. Mj+1 The corresponding eigenvectors form the eigenmatrix U, which is represented as:

[0178] U = [U1, U2, ..., U Mj+1 ].

[0179] In this embodiment, a whitening matrix is ​​constructed and whitening preprocessing is performed. The combined whitening processing matrix is ​​then constructed as follows: The beam domain data from the two sliding windows were then whitened.

[0180]

[0181]

[0182] S5. Sliding window joint blind separation matrix estimation;

[0183] In this embodiment, the whitened signal The fourth-order cumulant is expressed as:

[0184]

[0185] Construct a fourth-order cumulant matrix:

[0186]

[0187] The whitened signal was processed using the aforementioned operations. Construct a fourth-order cumulant matrix

[0188] In this embodiment, a joint fourth-order cumulant matrix is ​​constructed:

[0189]

[0190] Perform eigenvalue decomposition on the joint fourth-order cumulant matrix and express it in the form of eigenvalues ​​and eigenvectors:

[0191]

[0192] Where, λ i For the i-th eigenvalue, V i These are the corresponding eigenvectors. The eigenvalues ​​are arranged in descending order of their magnitudes. The corresponding feature vectors are as follows:

[0193]

[0194] In this embodiment, the first M is taken. j Construct a matrix using +1 eigenvalue and eigenvector:

[0195]

[0196] In this embodiment, (M) j +1) 2 ×1 dimensional column vector V i Arranged by column as (M) j +1)×(M j +1) characteristic matrix G i Then we obtain the joint characteristic matrix:

[0197]

[0198] In this embodiment, the separation matrix can be obtained by performing Joint Approximate Diagonalization (JADE) on the joint characteristic matrix:

[0199]

[0200] in, For matrix D i =q H G i The element in row k and column l of q.

[0201] S6, sliding window combined signal separation;

[0202] like Figure 5 As shown, in this embodiment, the separation matrix Q is used to analyze the sliding window signal vector undergoing joint whitening processing. and Perform signal separation:

[0203]

[0204] This allows for the separation of the target signal and the interference signal.

[0205] like Figure 6 As shown, in the separation results of this embodiment, four signals are separated each time the window slides out, and the fourth signal contains the target signal.

[0206] S7. Separation result detection and angle measurement.

[0207] like Figure 7 As shown, in this embodiment, the angular distribution of 100 experiments is... and For each row, CFAR detection is performed. For the distance cell j in the row h containing the target, phase-based angle measurement is performed. Where conj represents the conjugate, and ∠ represents the complex phase angle. Based on the calculation of the phased array path difference, the corresponding direction angle can be obtained as: Where λ is the wavelength, B d The distance between the phase centers of the two sliding windows is denoted as . In this embodiment, the target signal is . and B d It is half a wavelength.

[0208] In summary, this invention utilizes the array data of the phased array surface to sense the angle of interference, divides the entire array into two windows, and performs beamforming based on the interference angle and the radar's main operating angle. Furthermore, it performs joint blind source separation on the beam domain data of the two windows to separate the interference from the target signal, thereby enabling the detection of the separated target signal and target angle measurement. This solves the angle measurement error caused by the anti-interference process, achieving high-precision target angle measurement while suppressing interference, and efficiently solving the problem that traditional anti-main lobe interference methods cannot measure angles.

[0209] This invention, by employing a sliding window beamforming system architecture, reduces the requirement for the sliding invariance of subarray spacing in anti-interference methods. Compared with existing technologies, this invention has a wider range of applications. This invention solves the anti-interference problem in existing technologies where the subarray spacing cannot satisfy sliding invariance.

[0210] The beam domain method based on interference sensing in this invention reduces the uncertainty that may be introduced in non-interference directions by introducing interference sensing information, thereby concentrating the system's spatial resources on anti-interference and achieving better anti-interference performance.

[0211] This invention employs a sliding window beam domain method to form a beam based on interference sensing results, achieving dimensionality reduction. This avoids the problem of decreased sidelobe interference suppression performance caused by subarray dimensionality reduction, resulting in more accurate dimensionality reduction. Simultaneously, the beam can be directed at various interference sources, thus avoiding the contradiction between far-field sidelobe interference and subarray size and number in the subarray method. This invention solves the technical problems of angle measurement errors during the anti-interference process and the inability of subarray spacing to meet sliding invariance in existing technologies, as well as the technical problems of decreased sidelobe interference suppression performance caused by subarray dimensionality reduction.

[0212] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A radar anti-jamming method based on sliding window multi-beam blind source separation, characterized in that, The method includes: S1. Obtain full array data and sliding window subarray data; S2. Based on the full array data, perform covariance estimation to obtain signal feature values ​​and signal feature vectors, and then use the MUSIC algorithm to perform spatial spectrum estimation to obtain the angular distribution of all signal sources, thereby sensing signal interference. S3. Construct an interference beamforming steering vector for each direction of the signal interference, and use the interference beamforming steering vector to perform interference beamforming on the sliding window subarray data. Construct a target beamforming steering vector for the target direction, and use the target beamforming steering vector to perform target beamforming on the sliding window subarray data to construct sliding window beam domain data. S4. Perform joint blind source separation on the sliding window beam domain data, estimate the signal for each sliding window, perform joint preprocessing to obtain the joint covariance matrix, perform eigenvalue decomposition on the joint covariance matrix to construct the whitening matrix, and perform whitening preprocessing to obtain the whitened signal. S5. Construct a joint cumulant matrix based on all the whitening signals, obtain and eigendecompose the joint cumulant matrix to obtain joint cumulant eigenvalues ​​and cumulant eigenvectors, process them to obtain a joint feature matrix, and perform joint approximate diagonalization on the joint feature matrix to obtain a sliding window joint blind separation matrix. S6. Using the sliding window joint blind separation matrix, perform signal separation on the sliding window beam signal vector to obtain separated signal data, and obtain the target signal and interference signal accordingly; S7. Perform constant false alarm rate (CFAR) detection on the separated signal data path, perform angle measurement on the distance cell where the target is located, and calculate the phased array path difference to obtain the direction angle corresponding to the current phase angle.

2. The radar anti-jamming method based on sliding window multi-beam blind source separation according to claim 1, characterized in that, Step S1 includes: S11. Based on the source signal and system model, represent the signal of an array with N receiving elements using the following logic: Where M is the number of all signal sources received by the antenna, and satisfies: ， In the formula, As the direction guide vector, This represents the direction vector, where e is the natural constant. d is the spacing between the receiving channels. The angle between the signal and the antenna normal. For wavelength, the first Each component As a signal source, the signal source is represented using the following logic: in, For radar sampling frequency, The frequency of the signal source is used to receive the signal from the full array. Let be a complex matrix of N×P, where N is the spatial dimension and P is the fast-time sampling dimension. For noise; S12. Apply a sliding window to the entire signal matrix according to its spatial dimension, with each window containing the following number of elements: The signal matrices are divided into two spatial diversity groups: 。 3. The radar anti-jamming method based on sliding window multi-beam blind source separation according to claim 2, characterized in that, Step S2 includes: S21. Estimate the covariance matrix of the full array data using the following logic: in, Indicates conjugate transpose; S22. Represent the signal characteristic values ​​and signal characteristic vectors using the following logic: Among them, eigenvalues A positive real number, corresponding to the energy of the signal. correspond System characteristics of a signal source; S23. Arrange the signal feature values ​​in descending order as follows: The signal feature vectors corresponding to each of the aforementioned signal feature values ​​are as follows: ； S24. Construct a signal subspace based on the signal feature vector. and noise subspace : The spatial spectrum estimate obtained from the MUSIC algorithm is as follows: ； S25. Based on the spatial spectrum estimate, the angular distribution of all signal sources is obtained: In the formula, Indicates the first An estimated angle value.

4. The radar anti-jamming method based on sliding window multi-beam blind source separation according to claim 3, characterized in that, Step S3 includes: S31, the direction of interference for each of the aforementioned signals: ( The interference beam synthesis steering vector is constructed using the following logic: ； S32. Based on the interference beamforming steering vector, perform interference beamforming on the sliding window subarray data using the following logic to obtain the interference beamforming result: ； S33. Using the following logic, the target direction is... Construct the target beam synthesis steering vector: ； S34. Based on the following logic, using the target beamforming steering vector, perform target beamforming on the sliding window subarray data to obtain the target beamforming result: ； S35. Based on the interference beamforming result and the target beamforming result, construct sliding window beam domain data according to the following logic: and in, This indicates transpose.

5. The radar anti-jamming method based on sliding window multi-beam blind source separation according to claim 4, characterized in that, Step S4 includes: S41, Sliding window combined with whitening pretreatment Based on the signal from each sliding window, the covariance matrix is ​​estimated as follows: and in, For the first window A vector composed of column elements, For the second window A vector composed of column elements; S42. Based on the covariance matrix, construct the joint covariance matrix using the following logic: ； S43, regarding the joint covariance matrix Perform eigenvalue decomposition and represent the joint covariance matrix using the following logic: in, It is a diagonal matrix, and the diagonal elements are eigenvalues; S44. Sort the eigenvalues ​​in descending order: The feature matrix is ​​constructed using the feature vectors corresponding to the feature values. The feature matrix is ​​represented as follows: ; S45. Construct the whitening matrix and perform whitening preprocessing. Construct a joint whitening processing matrix using the following logic: The two sliding window beam domain data are then whitened using the following logic to obtain the first whitened signal. and the second whitening signal .

6. The radar anti-jamming method based on sliding window multi-beam blind source separation according to claim 5, characterized in that, Step S5 includes: S51. The first sliding window signal vector is represented using the following logic. The fourth-order cumulant: And construct the first fourth-order cumulant matrix based on the aforementioned fourth-order cumulants: ； The second sliding window signal vector is processed using the aforementioned logic in step S51. Construct the second and fourth order cumulant matrix ; S52. Based on the first fourth-order cumulant matrix and the second fourth-order cumulant matrix, construct a joint fourth-order cumulant matrix using the following logic: ； S53. Perform eigenvalue decomposition on the joint fourth-order cumulant matrix, and represent the joint cumulant eigenvalues ​​and the cumulant eigenvectors using the following logic: in, For the first 1 eigenvalue, The corresponding feature vector; S54, Take the previous step The eigenvector matrix is ​​constructed using the following logic, based on the given joint cumulant eigenvalues ​​and cumulant eigenvectors: ； S55, the feature vector matrix in dimensional vector Arranged by column Feature matrix The joint feature matrix is ​​obtained accordingly: ； S56. Perform JADE processing based on joint approximate diagonalization on the joint feature matrix to obtain the sliding window joint blind separation matrix: in, For matrix of OK Column elements.

7. The radar anti-jamming method based on sliding window multi-beam blind source separation according to claim 6, characterized in that, Step S53 includes: S531. Arrange the joint cumulative characteristic values ​​in descending order of their modulus: ； S532, the cumulative feature vector corresponding to each of the joint cumulative feature values ​​is: .

8. The radar anti-jamming method based on sliding window multi-beam blind source separation according to claim 7, characterized in that, In step S6, the sliding window joint blind separation matrix is ​​used. The sliding window signal vector of the joint whitening process and Perform signal separation: The target signal and the interference signal are separated accordingly.

9. The radar anti-jamming method based on sliding window multi-beam blind source separation according to claim 8, characterized in that, Step S7 includes: S71. Results of signal separation: and Perform the constant false alarm rate (CFAR) detection on each row; S72. Perform phase-based angle measurement processing on the distance cell where the target is located using the following logic: in, Indicates taking the conjugate. This represents the phase angle when taken as a complex number; S73. Calculate the path difference of the phased array to obtain the direction angle corresponding to the phase angle: in, For wavelength, The distance between the phase centers of the two sliding windows is denoted as .

10. The system of the radar anti-jamming method based on sliding window multi-beam blind source separation according to claim 1, characterized in that, The system includes: The receiving and sliding window data module is used to acquire full array data and sliding window subarray data; The interference sensing module is used to perform covariance estimation based on the full array data, obtain signal feature values ​​and signal feature vectors, and then use the MUSIC algorithm to perform spatial spectrum estimation to obtain the angular distribution of all signal sources, thereby sensing signal interference. The interference sensing module is connected to the receiving and sliding window data module. A sliding window beam domain acquisition module is used to construct an interference beamforming steering vector for each direction of the signal interference, perform interference beamforming on the sliding window subarray data using the interference beamforming steering vector, construct a target beamforming steering vector for the target direction, and perform target beamforming on the sliding window subarray data using the target beamforming steering vector to construct sliding window beam domain data. The sliding window beam domain acquisition module is connected to the interference sensing module. The sliding window beam domain data joint blind source separation module is used to estimate the signal for each sliding window separately, perform joint preprocessing to obtain a joint covariance matrix, perform eigenvalue decomposition on the joint covariance matrix to construct a whitening matrix, and perform whitening preprocessing to obtain a whitened signal. The sliding window beam domain data joint blind source separation module is connected to the receiving and sliding window data module. The sliding window joint blind separation matrix estimation module is used to construct a joint cumulant matrix based on all the whitening signals, obtain and eigendecompose the joint cumulant matrix to obtain joint cumulant eigenvalues ​​and cumulant eigenvectors, process them to obtain a joint feature matrix, and perform joint approximate diagonalization on the joint feature matrix to obtain the sliding window joint blind separation matrix. The sliding window joint blind separation matrix estimation module is connected to the sliding window joint whitening preprocessing module. The sliding window joint signal separation module is used to perform signal separation on the sliding window beam signal vector using the sliding window joint blind source separation matrix, thereby obtaining separated signal data, and thereby obtaining the target signal and interference signal. The sliding window joint signal separation module is connected to the joint blind source separation module. The separation result detection and angle measurement module is used to perform CFAR detection on each line of the separated signal path data, perform angle measurement on the distance cell where the target is located, and calculate the phased array path difference to obtain the direction angle corresponding to the current phase angle. The separation result detection and angle measurement module is connected to the sliding window joint blind separation matrix estimation module.

Images