A spatial polarization domain adaptive cancellation method based on blind source separation

Abstract

This invention belongs to the field of radar signal processing and discloses a spatial polarization domain adaptive cancellation method based on blind source separation. This method combines blind source separation with a spatial polarization adaptive cancellation algorithm and applies it to the suppression of main lobe-variable polarization interference. It solves the problem that traditional orthogonal polarization channel blind source separation is difficult to effectively separate the target signal when variable polarization interference enters from the main lobe. It uses a polarization array beamforming algorithm to solve the problem that blind source separation is difficult to separate the target by utilizing polarization information differences when the interference polarization state changes. It solves the problem that traditional adaptive cancellation algorithms weight and cancel the target signal together when interference exists in the main lobe through polarization blind source separation. Finally, it uses the spatial polarization adaptive cancellation method to achieve effective suppression of interference.

Description

Technical Field

[0001] This invention belongs to the field of radar signal processing, specifically relating to a spatial polarization domain adaptive cancellation method based on blind source separation. Background Technology

[0002] In recent years, the electronic warfare environment has become increasingly complex. When interference enters the radar from the antenna main lobe, sidelobe interference suppression methods become almost ineffective, and countering main lobe interference remains a hot topic and a challenge in the radar field. Blind source separation, the process of recovering unknown source signals from observed mixed signals, is widely used to suppress main lobe interference. Meanwhile, polarization, as the fourth dimension of electromagnetic waves beyond the spatial, temporal, and frequency domains, is also finding increasingly widespread applications. Blind source separation can also utilize the difference in polarization information between the target and the interference to achieve effective separation. However, when variable polarization interference enters from the main lobe, traditional blind source separation algorithms using orthogonal polarization channels struggle to suppress the interference. Summary of the Invention

[0003] To address the shortcomings of existing technologies, the present invention aims to propose a spatial polarization domain adaptive cancellation method based on blind source separation. This method combines blind source separation with a spatial polarization adaptive cancellation algorithm, applying it to suppress main lobe polarization-changing interference. It utilizes a polarization array beamforming algorithm to address the problem that blind source separation struggles to separate the target signal based on polarization differences when the interference polarization state changes. Furthermore, it solves the problem that traditional adaptive cancellation algorithms weight and cancel the target signal along with the main lobe when interference exists, through polarization blind source separation. Finally, the spatial polarization adaptive cancellation method effectively suppresses interference. To achieve the above objectives, the present invention employs the following technical methods.

[0004] A spatial polarization domain adaptive cancellation method based on blind source separation, the method comprising the following steps:

[0005] Constructing a polarization-shifting interference signal;

[0006] Acquire the received signals of the orthogonal polarization channels of each array element;

[0007] The polarization angle of the variable polarization interference in each pulse repetition period is determined using the interference signal amplitude of an orthogonal polarization receiving channel.

[0008] After beamforming the orthogonal polarization array signal, the horizontal and vertical polarizations and beams of the target and the interference are obtained respectively.

[0009] The separated interference signal is obtained by performing blind source separation on the sum beam of the horizontal polarization array and the sum beam of the vertical polarization array respectively;

[0010] Using the interference signal as an auxiliary channel and the target vertical polarization and beam as the main channels, adaptive cancellation of the empty poles is performed in each pulse repetition period.

[0011] The canceled signal is pulse compressed.

[0012] Furthermore, the variable polarization interference is slow-variable polarization interference, meaning that the interference polarization does not fluctuate within the duration of a single pulse repetition interval (PRI). Therefore, the interference signal has different polarization vectors h within different PRIs. j for:

[0013]

[0014] In the formula, τ is the ellipse tilt angle of the disturbance.

[0015] Furthermore, the received signal S of the orthogonal polarization array H (t) and S V (t) are respectively:

[0016]

[0017]

[0018] In the formula, h H h is the polarization vector corresponding to the radar horizontal receiving antenna. V S is the polarization vector corresponding to the radar's vertically polarized antenna. p Let h be the polarization scattering matrix of the target. t h is the polarization vector corresponding to the radar transmitting antenna. j For the interference polarization vector, g H For the gain of the radar horizontal receiving polarized antenna, g V For the gain of the radar vertical receiving polarized antenna, n H (t) and n V (t) represents the background noise within the polarized receiving channel, which follows a zero-mean Gaussian distribution. s and a j Spatial steering vectors for signals and interference, respectively:

[0019]

[0020]

[0021] In the formula, M is the number of array elements, λ is the wavelength of the incident signal, and d represents the spacing between equally spaced linear array elements. For the desired signal direction, The direction of interference.

[0022] Furthermore, the interference polarization does not fluctuate within a single PRI duration, and the pulse repetition period of the radar transmitted signal is generally much larger than the pulse width. Therefore, within the period from the maximum echo delay time to the next detection pulse transmission time, only the interference signal will exist, and the amplitude V of the interference signal in the orthogonal polarization receiving channel will be [not specified]. H (t) and V V (t) is:

[0023]

[0024]

[0025] The method for determining the polarization angle includes:

[0026] The polarizability ρ of the disturbance is calculated using the horizontal and vertical polarization components. J for:

[0027]

[0028] The polarization angle τ of the interference within each PRI is obtained using the interference polarization:

[0029]

[0030] Furthermore, the horizontal and vertical polarization and beam Y of the target signal sH (t) and Y sV (t) is:

[0031] Y sH (t)=W s H S H (t),

[0032] Y sV (t)=W s H S V (t),

[0033] The horizontal and vertical polarization and beam Y of the interference signal JH and Y JV for:

[0034] Y JH (t)=W J H S H (t),

[0035] Y JV (t)=W J H S H (t),

[0036] In the formula, [·]H The conjugate transpose operation is represented by the weighted vectors for the target and interference directions, respectively. in,[·] T This indicates a transpose operation, where φ1 and φ2 are the phase differences between adjacent array elements, and... λ represents the wavelength of the signal, and d represents the spacing between the array elements.

[0037] Furthermore, the polarization blind source separation method includes:

[0038] The signal to be processed is demeaned and whitened to obtain a preprocessed signal;

[0039] The fourth-order cumulant matrix of the preprocessed signal is obtained, and its eigenvalues ​​are decomposed to obtain the estimated matrix of the unitary matrix, thus finally obtaining the separated interference signal.

[0040] Furthermore, the signal to be processed is polarization and beam X, wherein the horizontal polarization and beam of the target and the interference are defined as X. H Vertical polarization and beam are defined as X V Blind source separation was performed separately.

[0041] The method for removing the mean is as follows:

[0042]

[0043] In the formula, The polarization and beam are preprocessed with zero mean.

[0044] The whitening matrix C in the whitening process is:

[0045]

[0046] In the formula, Λ k For the correlation matrix A diagonal matrix consisting of the first k largest eigenvalues, σ 2 E represents the noise variance. k V is a unit vector of order k. k It is a matrix consisting of the eigenvectors corresponding to k eigenvalues.

[0047] The preprocessed signal It can be represented as:

[0048]

[0049] The fourth-order cumulant matrix D z (P) is:

[0050]

[0051] In the formula, λ jLet [·] be the fourth-order cumulant of the signal source. H For the conjugate transpose, u j Let Λ be the j-th column vector of matrix U. P It is a diagonal matrix, and

[0052] The feature decomposition is as follows:

[0053] D z (P)=VΛV H ,

[0054] In the formula, Λ is a diagonal matrix, and matrix V is an estimate of matrix U.

[0055] The estimation matrix V is:

[0056] V=UD E T,

[0057] In the formula, D E It is a diagonal matrix with diagonal elements of ±1, and T is a permutation matrix.

[0058] The interference after separation is the separation signal Y. H Independent component elements in (k), where:

[0059]

[0060] In the formula, [·] H This indicates the conjugate transpose.

[0061] Furthermore, the empty pole adaptive cancellation method includes:

[0062] The signal within each PRI is segmented, and the weights of each segment are calculated and then canceled out. The two interfering signals within each PRI are defined as J. l (k), whose covariance matrix R HV for:

[0063]

[0064] In the formula, l represents the number of pulse repetition cycles, E{·} represents the expected value, [·] H This indicates the conjugate transpose operation.

[0065] Calculate target vertical polarization and beam y l (k) and the cross-correlation vector P of the interference signal yJ for:

[0066] P yJ =E{J l (k)y l * (k)},

[0067] Obtain the empty pole adaptive weighted W l for:

[0068]

[0069] In the formula [·] -1 This represents the matrix inversion operation. The target output s after weighted cancellation. l (k) is:

[0070] s l (k)=y l (k)-W l H (k)J l (k),

[0071] Furthermore, the output signal s after pulse compression o (t) is represented as:

[0072] s o (t) = IFFT{H(f)·S(f)},

[0073] In addition, the present invention also provides a computer-readable storage medium on which a computer program is stored, which, when executed by a processor, implements the steps of the method described above.

[0074] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0075] This invention combines blind source separation with an adaptive polarization cancellation algorithm and applies it to the suppression of main lobe polarization-changing interference. It solves the problem that traditional orthogonal polarization channel blind source separation is unable to effectively separate the target signal when polarization-changing interference enters from the main lobe. It utilizes a polarization array beamforming algorithm to solve the problem that blind source separation is unable to separate the target signal by utilizing polarization information differences when the interference polarization state changes. Through polarization blind source separation, it solves the problem that traditional adaptive cancellation algorithms weight and cancel the target signal together when interference exists in the main lobe. Attached Figure Description

[0076] Figure 1 System processing block diagram;

[0077] Figure 2 This is a block diagram illustrating the principle of polarization array beamforming.

[0078] Figure 3 Diagram of the principle of adaptive cancellation of empty poles;

[0079] Figure 4 For polarization domain parameter estimation;

[0080] Figure 5 Let be the horizontal polarization of the target and the interference, and the real part of the beam signal;

[0081] Figure 6 For the vertical polarization of the target and the interference, and the real part of the beam signal;

[0082] Figure 7 This represents the real part of each signal after horizontal polarization and beam splitting;

[0083] Figure 8 This represents the real part of each signal after vertical polarization and beam splitting;

[0084] Figure 9 The amplitude of the signal pulse compression before and after the empty pole adaptive cancellation. Detailed Implementation

[0085] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0086] The overall flowchart of the present invention is as follows: Figure 1 As shown.

[0087] 1. Polarization interference

[0088] According to polarization theory, the electric field characteristics of electromagnetic waves are called polarization. The polarization information of electromagnetic waves mainly depends on the amplitude ratio and phase difference of signals in two orthogonal directions on an equiphase surface, that is, the Jones vector E of the electromagnetic wave is:

[0089]

[0090] Where φ=φ y -φ x The relative phase of the two field components.

[0091] The corresponding polarization ratio ρ xy Defined as:

[0092]

[0093] In the formula, tan γ is the ratio of the amplitudes of the two orthogonal polarization components of the electric field, and φ is their phase difference. (γ, φ) is called the polarization phase descriptor of the electromagnetic wave, and its values ​​range from γ∈[0, π / 2] to φ∈[0, 2π]. Therefore, it can be seen that at a polarization ratio ρ... xy There is a one-to-one correspondence between them and the phase descriptors (γ, φ).

[0094] Substituting equation (2) into the expression for the Jones vector (1), and considering only the polarization state of the electric field while ignoring its amplitude information, we can assume that the electric field has a unit power density, i.e. Therefore, the Jones vector of the electromagnetic wave on the polarization basis (x, y) becomes:

[0095]

[0096] This is the polarization ratio representation of the electric field vector. The polarization ratio is a very important descriptor for electromagnetic wave polarization. Its value depends on the choice of the polarization basis. Once the polarization basis is given, the polarization ratio can completely and uniquely characterize the polarization of an electromagnetic wave.

[0097] The conversion relationships between the elliptic tilt angle τ, the ellipticity angle ε, and the electromagnetic wave parameters γ and φ are as follows:

[0098]

[0099] When the polarization state is linear polarization, the ellipticity angle ε = 0, and the range of the elliptic inclination angle τ is defined as τ∈[-π / 2, π / 2]. Substituting into (4), we get γ = |τ|, that is, when τ∈[-π / 2, 0], φ = π, and when τ∈[0, π / 2], φ = 0. Therefore, the linear polarization ratio ρ xy for:

[0100]

[0101] Substituting into formula (3), we can obtain the polarization vector h of the linear polarization interference. j for:

[0102]

[0103] Slowly polarized jamming refers to jamming where the polarization transition period changes slowly relative to the radar detection period. It is generally assumed that the jamming polarization does not fluctuate within a single radar detection period (PRI). Therefore, the jamming signal has different polarization vectors h within different PRIs. j This achieves a polarized interference signal.

[0104] 2. Orthogonal polarization channel model of the array

[0105] For far-field narrowband signals, there will be a delay when the same signal arrives at different array elements. This delay causes a phase difference between the various receiving array elements. At the same time, the polarization angles of the target and the interference are also different. Therefore, the steering vectors of the target signal and the interference are described by two-dimensional features in the spatial domain and polarization domain.

[0106] The polarization state of the interference signal is generally different from that of the target echo. Here, a horizontally polarized signal is transmitted while a variable-polarization interference signal is received. Let the target echo be s(t) and the interference signal be J(t), then the signals S received by the horizontal and vertical polarization arrays... H (t) and S V (t) are respectively represented as:

[0107]

[0108]

[0109] In the formula, h H h is the polarization vector corresponding to the radar horizontal receiving antenna. V S is the polarization vector corresponding to the radar's vertically polarized antenna. p Let h be the polarization scattering matrix of the target. t h is the polarization vector corresponding to the radar transmitting antenna. j For the interference polarization vector, g H For the gain of the radar horizontal receiving polarized antenna, g V For the gain of the radar vertical receiving polarized antenna, n H (t) and n V (t) represents the background noise within the polarized receiving channel, which follows a zero-mean Gaussian distribution. s and a j Spatial steering vectors for signals and interference, respectively:

[0110]

[0111]

[0112] In the formula, M is the number of array elements, λ is the wavelength of the incident signal, and d represents the spacing between equally spaced linear array elements. For the desired signal direction, The direction of interference.

[0113] 3. Measure the interference polarization angle

[0114] Since the interference polarization does not fluctuate within a single PRI duration, and the pulse repetition period of the radar transmitted signal is generally much larger than the pulse width, according to the radar detection period, only the interference signal will exist during the period from the maximum echo delay time to the next detection pulse transmission time. The horizontal polarization component V of the interference signal... H and vertical polarization component V V for:

[0115]

[0116]

[0117] In the formula, h H h is the polarization vector corresponding to the radar horizontal receiving antenna. V h is the polarization vector corresponding to the radar vertical polarization antenna. j For the interference polarization vector, g H For the gain of the radar horizontal receiving polarized antenna, g V For the gain of the radar vertical receiving polarized antenna, n H(t) and n V (t) represents the background noise within the polarized receiving channel.

[0118] The polarizability ρ of the interference can be calculated using the horizontal and vertical polarization components of the interference. J for:

[0119]

[0120] Furthermore, the polarization angle τ of the interference within each pulse repetition interval can be calculated as follows:

[0121]

[0122] 4. Polarized array beamforming

[0123] By weighted summing of the horizontal and vertical polarization outputs of each array element, the horizontal and vertical polarization array beams of the antenna are "guided" to one direction within a certain time, thus obtaining the horizontal and vertical polarization beams of the target and the horizontal and vertical polarization beams of the interference. Figure 2 This is a block diagram illustrating the principle of polarization array beamforming.

[0124] In order to achieve the target direction θ s and interference direction θ J The time delay between the polarization channels of each array element is compensated to form a main beam. The weighting vector W of a conventional beamformer in this direction is... s (θ) and W J (θ) are respectively:

[0125]

[0126]

[0127] In the formula, [·] T This indicates a transpose operation, where φ1 and φ2 are the phase differences between adjacent array elements, and... Where λ represents the wavelength of the signal and d represents the spacing between the array elements.

[0128] At this time, the target signal outputs signal Y in both the horizontal polarization array and the vertical polarization array DBF. sH and Y sV They are represented as follows:

[0129] Y sH (t)=W s H S H (t), (17)

[0130] Y sV (t)=W s HS V (t), (18)

[0131] Interference signal output signal Y in horizontal polarization array and vertical polarization array DBF JH and Y JV They are represented as follows:

[0132] Y JH (t)=W J H S H (t), (19)

[0133] Y JV (t)=W J H S H (t), (20)

[0134] in,[·] H This indicates the conjugate transpose operation.

[0135] 5. Polarization blind source separation method

[0136] Blind source separation can utilize the Joint Approximate Diagonalization of Eigenmatrices (JADE) algorithm to separate interference and the target, thereby suppressing main lobe interference. The sum of the polarization array beams is defined as X, where the horizontal polarization sum of the target and interference beams is defined as X0. H Vertical polarization and beam are defined as X V The JADE algorithm is applied separately. Since the target signal is horizontally polarized and the interference is linearly polarized, the spatial difference and polarization difference between the interference and the target signals received by the horizontal and vertical polarization arrays are used to separate the interference signal and the target signal.

[0137] First, zero-mean preprocessing is used to remove the strongly correlated DC component from the signal X to be processed. This can be achieved by directly subtracting its expected value from the DC component, making its expected value zero. This can be expressed as:

[0138]

[0139] In the formula, The polarization and beam are preprocessed with zero mean.

[0140] Then, whitening treatment is used to weaken or eliminate the correlation between polarization and beam components, and to make it have unit variance. First, the following is obtained: Correlation matrix R XX for:

[0141]

[0142] Therefore, the whitening array C can be obtained as follows:

[0143]

[0144] In the formula, Λ k For R XX A diagonal matrix consisting of the first k largest eigenvalues, σ 2 E represents the noise variance. k V is a unit vector of order k. k It is a matrix consisting of the eigenvectors corresponding to k eigenvalues.

[0145] The polarization and beam after whitening pretreatment It can be represented as:

[0146]

[0147] Next, calculate the polarization and beam after whitening. The fourth-order cumulant matrix D z (P) is:

[0148]

[0149] In the formula, P is an arbitrary matrix, p ij Let be the element in the i-th row and j-th column of matrix P.

[0150] Based on the multilinearity property of higher-order cumulants, the fourth-order cumulant matrix D z (P) can be represented as:

[0151]

[0152] In the formula, λ j Let [·] be the fourth-order cumulant of the signal source. H For the conjugate transpose, u j Let Λ be the j-th column vector of matrix U. P It is a diagonal matrix, and

[0153] For matrix D z (P) Perform eigenvalue decomposition:

[0154] D z (P)=VΛV H (27)

[0155] In the formula, Λ is a diagonal matrix, matrix V is an estimate of matrix U, and we have:

[0156] V=UD E T, (28)

[0157] In the formula, D E It is a diagonal matrix with diagonal elements of ±1, and T is a permutation matrix.

[0158] The separated signal Y is obtained through matrix V. H The expression for (k) is:

[0159]

[0160] in,[·] H Y represents the conjugate transpose. H (k) Obtain the independent component parts after separation, namely the separated interference signal and target signal. Use the interference signal obtained after two blind source separations as the auxiliary channel signal, and the target polarization and beam as the main channel signal to perform a space-polarity adaptive cancellation algorithm to suppress interference.

[0161] 6. Hollow Pole Adaptive Cancellation Algorithm

[0162] To address variable polarization interference and achieve better cancellation performance, the signal within each PRI is segmented, and the weights are calculated segment by segment for cancellation. Hollow-polarity adaptive cancellation involves adaptively weighting the two separated interference signals and canceling them against the target's vertical polarization and beam signal. The gain is adjusted in the variable polarization interference direction to create a null point in the interference direction, thus suppressing the interference signal. Figure 3 This is a diagram illustrating the principle of adaptive cancellation of empty poles.

[0163] The two interference signals within each PRI are defined as J. l (k), the target vertical and beam are defined as y l (k), where l is the number of pulse repetition periods, then its covariance matrix R HV It can be represented as:

[0164]

[0165] In the formula, E{·} represents the expected value, [·] H This indicates the conjugate transpose operation.

[0166] Calculate the cross-correlation vector P of the primary and secondary channels yJ for:

[0167] P yJ =E{J l (k)y l * (k)}, (31)

[0168] The empty pole adaptive weighted W can be obtained. l for:

[0169]

[0170] In the formula [·] -1This represents the matrix inversion operation. The target output s after weighted cancellation. l (k) is:

[0171] s l (k)=y l (k)-W l H (k)J l (k), (33)

[0172] 7. Pulse Compression

[0173] Pulse compression, also known as matched filtering, utilizes the cross-correlation function between the delays of the received and transmitted signals to compress a wide transmitted pulse signal into a narrow pulse signal. There are two methods to achieve this: using time-domain correlation processing and implementing it in the frequency domain using FFT.

[0174] Compared to time-domain convolution, pulse compression using the frequency domain significantly reduces computational complexity. Furthermore, windowing functions can be used to suppress sidelobes during pulse compression. This is achieved by pre-multiplying the matched filter coefficients and the window function in the computer simulation using either frequency-domain windowing or time-domain windowing.

[0175] H(f)=FFT{h(n)·ω(n)}, (34)

[0176] In the formula, ω(n) is a window function, and an appropriate window function can be selected as needed.

[0177] Here, frequency domain pulse compression is performed on the signal after pole cancellation. Since the matched filter is a linear time-invariant system, according to the properties of the Fourier transform, we have:

[0178]

[0179] When both signals are correctly sampled, the pulse compression output signal s o (t) is represented as:

[0180] s o (t)=IFFT{H(f)·S(f)}, (36)

[0181] Example

[0182] To verify the performance of the adaptive cancellation method proposed in this invention, computer simulation was used to conduct a suppression experiment on main lobe polarization interference.

[0183] The computer simulation conditions are as follows: the radar transmits 4 pulses, 8 radar receiver elements are set, the radar transmits a linear frequency modulated signal with bandwidth B = 1MHz and pulse width T. p =200μs, f s =2MHz, pulse repetition period Tr =1ms. Signal-to-noise ratio is 10dB, signal-to-interference ratio is -25dB.

[0184] The radar transmitting antenna is horizontally polarized, with a polarization vector h. t =[1,0] T The interference polarization within different PRIs changes from 15° linear polarization to 20°, 25°, and 30° linear polarization; the antenna is set as an orthogonal polarization receiving array, with polarization vectors h, ... HH =[1,0] T and h VV =[0,1] T Main antenna gain g H =30, auxiliary antenna gain g V =10; without loss of generality, let the target's polarization scattering matrix be... The target signal is located at 30°, and the interference signal is noise FM interference located at 27°.

[0185] Parameter estimation of the polarization domain is as follows Figure 4 As shown, the polarization angles of the interference in each PRI are obtained by using the amplitude of the interference signal in the orthogonal polarization channel as 15°, 20°, 25° and 30° respectively. At the same time, since the interference is linearly polarized interference at this time, the polarization phase difference of the interference is 0°.

[0186] Beamforming is performed on the horizontally and vertically polarized arrays for the received signals from the array, with the horizontal polarization and beamforming of the target and interference as follows: Figure 5 As shown, vertical polarization and beam are Figure 6 As shown, blind source separation is performed separately.

[0187] After horizontal polarization and beam blind source separation, the real parts of each signal are as follows: Figure 7 As shown, the real parts of each signal after vertical polarization and beam splitting are as follows: Figure 8 As shown, the two separated interference signals are used as auxiliary channels, while the target vertical and beam signals are used as the main channel signals for adaptive processing.

[0188] The pulse pressure amplitude before and after the empty pole adaptive cancellation is as follows: Figure 9 As shown, it is evident that the target polarization and beam pulse compression before cancellation were insufficient to detect the target, while the target signal was successfully detected by the pulse compression after cancellation. The signal-to-interference-plus-noise ratio (SIR) before cancellation was measured to be 12.86 dB, and the SIR after cancellation was 32.27 dB. This successfully suppressed the main lobe polarization interference and improved the target's SIR.

[0189] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the protection scope of the present invention.

Claims

1. A spatial polarization domain adaptive cancellation method based on blind source separation, characterized in that, Includes the following steps: Constructing a polarization-shifting interference signal; Acquire the received signals of the orthogonal polarization channels of each array element; The polarization angle of the variable polarization interference in each pulse repetition period is determined using the interference signal amplitude of an orthogonal polarization receiving channel. After beamforming the orthogonal polarization array signal, the horizontal and vertical polarizations and beams of the target and the interference are obtained respectively. The separated interference signal is obtained by performing blind source separation on the sum beam of the horizontal polarization array and the sum beam of the vertical polarization array respectively; Using the interference signal as an auxiliary channel and the target vertical polarization and beam as the main channels, adaptive cancellation of the empty poles is performed in each pulse repetition period. The canceled signal is pulse compressed.

2. The spatial polarization domain adaptive cancellation method based on blind source separation according to claim 1, characterized in that, The polarization-changing interference is slow-polarization interference, meaning that the interference polarization does not fluctuate within the duration of a single pulse repetition interval (PRI). Therefore, the interference signal has different polarization vectors h within different PRIs. j for: In the formula, τ is the ellipse tilt angle of the disturbance.

3. The spatial polarization domain adaptive cancellation method based on blind source separation according to claim 1, characterized in that, The received signal S of the orthogonal polarization array H (t) and S V (t) are respectively: In the formula, h H h is the polarization vector corresponding to the radar horizontal receiving antenna. V S is the polarization vector corresponding to the radar's vertically polarized antenna. p Let h be the polarization scattering matrix of the target. t h is the polarization vector corresponding to the radar transmitting antenna. j For the interference polarization vector, g H For the gain of the radar horizontal receiving polarized antenna, g V For the gain of the radar vertical receiving polarized antenna, n H (t) and n V (t) represents the background noise within the polarized receiving channel, which follows a zero-mean Gaussian distribution. s and a j Spatial steering vectors for signals and interference, respectively: In the formula, M is the number of array elements, λ is the wavelength of the incident signal, d represents the spacing between equally spaced linear array elements, and θ s θ represents the desired signal direction. j The direction of interference.

4. The spatial polarization domain adaptive cancellation method based on blind source separation according to claim 1, characterized in that, The interference polarization remains constant throughout the duration of a single PRI, and the pulse repetition period of the radar transmitted signal is much longer than the pulse width. Therefore, during the period from the maximum echo delay to the next detection pulse transmission time, only the interference signal will exist. The horizontal polarization component signal V of the interference in the orthogonal polarization receiving channel... H (t) and vertical polarization component signal V V (t) are respectively: The polarizability ρ of the disturbance is calculated using the horizontal and vertical polarization components. J for: The polarization angle τ of the interference within each PRI is obtained using the interference polarization:

5. The spatial polarization domain adaptive cancellation method based on blind source separation according to claim 1, characterized in that, Horizontal and vertical polarization of the target signal and beam output signal Y sH (t) and Y sV (t) are respectively: Y sH (t)=W s H S H (t), Y sV (t)=W s H S V (t), Horizontal and vertical polarization of the interference signal and beam output signal Y JH and Y JV They are respectively: Y JH (t)=W J H S H (t), Y JV (t)=W J H S H (t), In the formula, [·] H The conjugate transpose operation is represented by the weighted vectors for the target and interference directions, respectively. in,[·] T This indicates a transpose operation, where φ1 and φ2 are the phase differences between adjacent array elements, and... λ represents the wavelength of the signal, and d represents the spacing between the array elements.

6. The spatial polarization domain adaptive cancellation method based on blind source separation according to claim 1, characterized in that, Polarization blind source separation methods include: The signal to be processed is zero-mean and whitened to obtain a preprocessed signal; Find the fourth-order cumulant matrix of the preprocessed signal, perform eigenvalue decomposition on it, obtain the estimated matrix of the unitary matrix, and finally obtain the separated interference signal.

7. The spatial polarization domain adaptive cancellation method based on blind source separation according to claim 6, characterized in that, The signal to be processed is the polarization and beam X, where the horizontal polarization and beam of the target and the interference are defined as X. H Vertical polarization and beam are defined as X V Blind source separation is performed separately; The method for calculating the zero mean is as follows: In the formula, For polarization and beamforming that have undergone zero-mean preprocessing; The whitening matrix C in the whitening process is: In the formula, Λ k For the correlation matrix A diagonal matrix consisting of the first k largest eigenvalues, σ 2 E represents the noise variance. k V is a unit vector of order k. k It is a matrix consisting of the eigenvectors corresponding to k eigenvalues; The preprocessed signal It can be represented as: The fourth-order cumulant matrix D z (P) is: In the formula, λ j Let [·] be the fourth-order cumulant of the signal source. H For the conjugate transpose, u j Let Λ be the j-th column vector of matrix U. P It is a diagonal matrix, and The feature decomposition is as follows: D z (P)=VΛV H , In the formula, Λ is a diagonal matrix, and matrix V is an estimate of matrix U. The estimation matrix V is: V=UD E T, In the formula, D E It is a diagonal matrix with diagonal elements of ±1, and T is a permutation matrix; The interference after separation is the separation signal Y. H Independent component elements in (k), where: In the formula, [·] H This indicates the conjugate transpose.

8. The spatial polarization domain adaptive cancellation method based on blind source separation according to claim 1, characterized in that, The vacuum pole adaptive cancellation method uses the interference obtained after two blind source separations as the auxiliary channel signal, and the target and beam as the main channel signal. The signal within each PRI is segmented, the weights of each segment are calculated, and then the segments are canceled out. The two interference signals within each PRI are defined as J. l (k), whose covariance matrix R is: In the formula, l represents the number of pulse repetition cycles, E{·} represents the expected value, [·] H This represents the conjugate transpose operation; Calculate target vertical polarization and beam y l The cross-correlation vector P between (k) and the interference signal is: P=E{J l (k)y l * (k)}, The empty-pole adaptive weighted W is obtained as follows: W=R -1 P, In the formula [·] -1 This represents the matrix inversion operation, and the target output s after weighted cancellation. l (k) is: s l (k)=y l (k)-W H (k)J l (k)。 9. The spatial polarization domain adaptive cancellation method based on blind source separation according to claim 1, characterized in that, The pulse-compressed output signal s o (t) is represented as: s o (t)=IFFT{H(f)·S(f)}。 10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1-9.

Images