A Parameter Estimation Method for Near-Field Polarimetric MIMO Radar Based on Parallel Factorization

Abstract

The present invention relates to a near-field polarization MIMO radar parameter estimation method based on parallel factor decomposition. A system model of a bistatic MIMO radar based on a uniform linear array of crossed dipole-magnetic loop antennas is established. The transmitting end comprises a uniform linear transmitting array composed of 2M+1 crossed dipole-magnetic loop antennas, and the receiving end comprises a uniform linear receiving array composed of 2N+1 crossed dipole-magnetic loop antennas. The spacing between transmitting array elements and the spacing between receiving array elements are represented by dt and dr, respectively. K uncorrelated near-field narrowband fully polarized targets are located. The two-dimensional transmitting angle and the two-dimensional receiving angle of the kth target are represented by θ tk and θ rk Represented by k = 1, 2, ... K, this method can accurately calculate the two-dimensional transmission angle, two-dimensional receiving angle, two-dimensional transmission polarization angle, two-dimensional receiving polarization angle and target position coordinates under the premise of a uniform linear array of cross-dipole-magnetic ring antenna. The method is simple and flexible.

Description

Technical Field

[0001] The present invention relates to the technical field of radar signal processing, and in particular to a near-field polarization MIMO radar parameter estimation method based on parallel factor decomposition. Background Art

[0002] MIMO radar parameter estimation is a current research hotspot and a challenge. Compared to phased array radars, MIMO radars can achieve higher angular resolution, greater array freedom, and flexible transmit waveform design by leveraging mutually orthogonal waveform characteristics. In recent years, numerous excellent algorithms have been proposed for joint parameter estimation of transmit and receive angles in bistatic MIMO radars. Although these algorithms can accurately locate target parameters in various backgrounds, the transmit and receive arrays used are either scalar uniform or non-uniform, and the target sources are mostly far-field. Compared to scalar arrays, polarimetric arrays provide not only target angle information but also polarization information. Therefore, polarimetric arrays offer higher target resolution and parameter estimation capabilities. Furthermore, when targets are located in the near-field of the MIMO radar's radiation, the nonlinearity of the signal wavefront representation means that directly applying far-field source direction-finding algorithms will result in significant estimation errors due to model mismatch. Currently, no multidimensional parameter estimation algorithms for MIMO radars that utilize polarimetric information have been proposed for near-field scenarios. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a near-field polarization MIMO radar parameter estimation method based on parallel factor decomposition, which can accurately calculate the two-dimensional transmission angle, two-dimensional reception angle, two-dimensional transmission polarization angle, two-dimensional reception polarization angle and target position coordinates under the premise of a uniform linear array of crossed dipole-magnetic loop antennas.

[0004] The technical solution adopted by the present invention is a near-field polarization MIMO radar parameter estimation method based on parallel factor decomposition, which includes the following steps:

[0005] S1. Establish a system model for a bistatic MIMO radar based on a uniform linear array of crossed dipole-magnetic loop antennas. In the system model, the transmitter consists of a uniform linear transmitting array composed of 2M+1 crossed dipole-magnetic loop antennas, and the receiver consists of a uniform linear receiving array composed of 2N+1 crossed dipole-magnetic loop antennas. The spacing between the transmitting and receiving elements is represented by dt and dr, respectively, and satisfies dt≤λ / 2 and dr≤λ / 2, where λ is the wavelength of the electromagnetic wave.

[0006] S2. Position K uncorrelated near-field narrowband fully polarized targets. The two-dimensional transmission angle and two-dimensional receiving angle of the k-th target are respectively expressed as θtk and θ rk Denotes, where k = 1, 2, ... K;

[0007] S3. During the first snapshot period, the data received by the receiving array is obtained, which is expressed as:

[0008] X l =A r (θ rk ,ρ rk ,r rk ,η rk )diag{b (l)}A t T (θ tk ,ρ tk ,r tk ,η tk )S+Z (l) , where A r (θ rk ,ρ rk ,r rk ,η rk ) is K

[0009] The steering vector of the receiving array, A t (θ tk ,ρ tk ,r tk ,η tk ) is the steering vector of the K transmitting arrays, S is the transmitting waveform matrix; b (l) is the scattering coefficient, Z (l) is the noise matrix, ρ rk is the distance from the receiving array to the target, r rk is the two-dimensional receiving polarization angle of the kth target, η rk is the two-dimensional receiving phase difference of the kth target, θ rk is the two-dimensional acceptance angle of the k-th target, θ tk is the two-dimensional emission angle of the kth target, ρ tk is the distance from the transmitting array to the target, r tk is the two-dimensional emission polarization angle of the kth target, η tk is the two-dimensional emission phase difference of the kth target, where 0≤r tk ,r rk ≤π / 2, 0≤η tk ,η rk ≤2π;

[0010] S4. Consider obtaining the output signal matrix of the matched filter at the receiving end under L snapshots, which is expressed as:

[0011] Y=(Q t ⊙Vt ⊙Q r ⊙V r )B+N, where Q t is the transmission signal steering vector; Q r is the received signal steering vector; V t is the spatial response of the transmitting array; V r is the spatial response of the receiving array; B = [b (1) ,…,b (L) ]; N = [n (1) ,…,n (L) ],in is the filtered noise vector;

[0012] S5. Arrange the output signal matrix of the receiving end matched filter in step (4) into a fifth-order PARAFAC tensor decomposition model, expressed as: Then convert the fifth-order PARAFAC tensor decomposition model into a third-order PARAFAC tensor decomposition model, namely:

[0013] S6. Based on the third-order PARAFAC tensor decomposition model, the complex parallel factor decomposition algorithm is used to decompose A t ,A r and B, and we get A t ,A r and the estimated values of B, respectively, are denoted as and

[0014] S7, according to the expression Come to Perform eigendecomposition to obtain the eigenvector P t =[o t1 ,o t2 ,…,o tK ] and eigenvalue R t =diag{λ t1 ,λ t2 ,…λ tK}, where Ψ t is a K×K dimensional matrix, Δ1 is the scale fuzzy matrix, J t1 =[Ι 2M ,0 2M ], J t2 =[0 2M ,I 2M ], I 2M is the identity matrix of 2M×2M, 0 2M is a 2M×2M zero matrix, Ι2 is a 2×2 dimensional identity matrix, A t is the steering vector of the transmitting array;

[0015] S8. The estimated value of Π is obtained based on the obtained eigenvector and eigenvalue. Among them, round{·} is a rounding function, Re{P t} means taking P t The real part of V t The estimated value of is: in, Indicated by The submatrix consisting of the 2m-1th row to the 2mth row; according to V t The estimated value of is used to derive the two-dimensional emission phase difference of the kth target: The two-dimensional emission polarization angle of the kth target is: in, express The element in the second row and k column of express The element in the first row and k column of ;

[0016] S9. According to the expression Get w tk The estimated value of is: According to w tk The estimated value of is the two-dimensional launch angle of the kth target:

[0017] S10, according to the expression Get V r The estimated value of is:

[0018] Among them, J r1 =[Ι 2N ,0 2N ], J r2 =[0 2N ,I 2N ], I 2N is the identity matrix of 2N×2N, 0 2N is a 2N×2N zero matrix, Δ2 is the scale fuzzy matrix, Ψ r It is a K×K dimensional matrix, according to V r The estimated value of gives the two-dimensional receiving phase difference of the k-th target as: The two-dimensional receiving polarization angle of the kth target is: in, express The element in the second row and k column of express The element in the first row and k column of ;

[0019] S11. According to the expression Get w rkThe estimated value of is: According to w rk The estimated value of the two-dimensional acceptance angle of the k-th target is: The position of the kth target is:

[0020] The beneficial effects of the present invention are as follows: through the above-mentioned near-field polarization MIMO radar parameter estimation method based on parallel factor decomposition, it is possible to accurately calculate parameters such as two-dimensional transmission angle, two-dimensional reception angle, two-dimensional transmission polarization angle, two-dimensional reception polarization angle and target position coordinates under the premise of a uniform linear array of cross-dipole-magnetic ring antenna, and the method is simple and flexible.

[0021] Preferably, in step S2, when positioning K uncorrelated near-field narrowband fully polarized targets, 2M+1 transmitting array elements simultaneously transmit 2M+1 orthogonal signals at the transmitting end, which are reflected upon encountering the target to form echoes received by the receiving array elements; and it is set that within the same signal period, the cross-sectional scintillation of the target remains constant, the cross-sectional scintillation fluctuations of different targets are unrelated to each other, and the fluctuation statistics within different pulse times are also independent of each other.

[0022] As a preference, in step S9, w is obtained tk The specific process of estimating the value is as follows: First, according to get: where unvec(·) is a matrix operator, which is the inverse operation of vec(·); in is a permutation matrix with all elements on the diagonal being 1 and all other elements being 0, and we can further obtain: Finally, we get w tk estimated value.

[0023] As a preference, in step S10, V r The specific process of estimating the value is as follows: First, according to and With the same permutation matrix we get: pass Further we get: Finally, we get V r estimated value.

[0024] As a preference, in step S11, w is obtained rk The specific process of estimating the value of is as follows: get: Among them, unvec(·) is a matrix operator, which is the inverse operation of vec(·); through in is a permutation matrix with all elements on the diagonal being 1 and all other elements being 0, and we can further obtain:

[0025] Finally, we get w rk estimated value. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A schematic diagram of a bistatic COLD uniform linear array MIMO radar on which the method of the present invention is based;

[0027] Figure 2 This is an estimated diagram of a two-dimensional transmission angle and a two-dimensional reception angle of a near-field polarization MIMO radar obtained by using the method of the present invention in an example of an embodiment of the present invention;

[0028] Figure 3 This is a target position estimation image of a near-field polarization MIMO radar obtained by using the method of the present invention in an example of an embodiment of the present invention;

[0029] Figure 4 This is a polarization parameter estimation diagram of a transmit array of a near-field polarization MIMO radar obtained by using the method of the present invention in an example of an embodiment of the present invention;

[0030] Figure 5 This is a diagram of receiving array polarization parameter estimation of a near-field polarization MIMO radar obtained by using the method of the present invention in an example of an embodiment of the present invention. DETAILED DESCRIPTION

[0031] The invention will be further described below with reference to the accompanying drawings and in combination with specific implementations, so that those skilled in the art can implement the invention with reference to the description. The protection scope of the invention is not limited to the specific implementations.

[0032] Those skilled in the art should understand that, in the disclosure of the present invention, the orientation or positional relationship indicated by terms such as "longitudinal", "transverse", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", and "outside" are based on the orientation or positional relationship shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, the above terms should not be understood as limiting the present invention.

[0033] The present invention relates to a near-field polarization MIMO radar parameter estimation method based on parallel factor decomposition, which specifically comprises:

[0034] S1. First, establish a system model of a bistatic MIMO radar based on a uniform linear array of crossed dipole-magnetic loop antennas. The transmitting end is a uniform linear transmitting array composed of 2M+1 crossed dipole-magnetic loop antennas, and the receiving end is a uniform linear receiving array composed of 2N+1 crossed dipole-magnetic loop antennas. The spacing between the transmitting array elements and the receiving array elements is represented by dt and dr respectively, and they satisfy dt,dr≤λ / 2, where λ is the wavelength of the electromagnetic wave. The center of the transmitting array and the receiving array is taken as the reference point O, as shown in the following example: Figure 1 The figure shows a schematic diagram of a bistatic cross-dipole-magnetic loop antenna uniform linear array MIMO radar;

[0035] S2, assuming that K uncorrelated near-field narrowband fully polarized targets in the yoz plane are located, the two-dimensional transmission angle and two-dimensional receiving angle of the k-th target are represented by θ tk and θ rk It means that 2M+1 transmitting elements simultaneously transmit 2M+1 orthogonal signals at the transmitting end. When they encounter the target, they are reflected to form echoes received by the receiving elements. Assuming that within the same signal period, the target's Radar Cross Set (RCS) remains constant, the RCS fluctuations of different targets are uncorrelated, and the fluctuation statistics within different pulse times are also independent of each other;

[0036] S3. During the first snapshot, the received data is expressed as

[0037]

[0038] Where,

[0039] A r (θ rk ,ρ rk ,r rk ,η rk )=[a r (θ r1 ,ρ r1 ,r r1 ,η r1 ),…a r (θ rk ,ρ rk ,r rk ,η rk )] (2)

[0040] A t (θ tk ,ρ tk ,r tk ,η tk )=[a t (θ t1 ,ρ t1 ,r t1 ,ηt1 ),…a t (θ tk ,ρ tk ,r tk ,η tk )] (3)

[0041]

[0042]

[0043] Among them, a t (θ tk ,ρ tk ,r tk ,η tk ) is the launch array steering vector of the kth target, expressed as

[0044]

[0045] Among them, q tk (θ tk ,ρ tk ) is the transmission signal steering vector of the kth target, expressed as:

[0046]

[0047] Where, λ is the signal wavelength;

[0048] V t (r t ,η t ) is the spatial response of the transmitting array, expressed as

[0049] V t (r t ,η t )=[v t1 (r t1 ,η t1 ),…,v tk (r tk ,η tk )] (8)

[0050] Then v tk (r tk ,η tk ) is the spatial response of the transmitted signal, expressed as

[0051]

[0052] a r (θ rk ,ρ rk ,r rk ,η rk) is the receiving array steering vector of the kth target, expressed as

[0053]

[0054] Among them, q rk (θ rk ,ρ rk ) is the received signal steering vector of the kth target, expressed as

[0055]

[0056] Where, λ is the signal wavelength;

[0057] V r (r r ,η r ) is the spatial response of the receiving array, expressed as:

[0058] V r (r r ,η r )=[v r1 (r r1 ,η r1 ),…,v rk (r rk ,η rk )] (12)

[0059] Then, v rk (r rk ,η rk ) is the spatial response of the received signal, which can be expressed as:

[0060]

[0061] is the scattering coefficient of the kth target, expressed as Among them, β k , f dk , t l They represent the target’s echo area, Doppler shift and slow time respectively;

[0062] s m ∈c 1×p represents the symbol sequence of length p transmitted by the mth transmitting array in one snapshot, m=-M,...,-1,0,1,...M; assuming Mutually orthogonal, for any m≠n, we have:

[0063] Z (l) is the noise matrix, which is independent and identically distributed Gaussian white noise with a mean of 0;

[0064] S4, the received signal is transmitted through the transmitting signal s m (m=-M,…,-1,0,1,…,M) filtering, then the output signal matrix can be expressed as

[0065] y (l) =(Q t ⊙V t ⊙Q r ⊙V r )b (l) +n (l) (14)

[0066] Where, is the filtered noise vector;

[0067] S5. Considering L snapshots, the output signal matrix of the matched filter at the receiving end can be expressed as:

[0068] Y=(Q t ⊙V t ⊙Q r ⊙V r )B+N (15)

[0069] Among them, Q t is the transmission signal steering vector, denoted as Q t =[q t1 ,…,q tk ],q tk represents the transmission signal steering vector of the kth target; Q r is the received signal steering vector, denoted as Q r =[q r1 ,…,q rk ],q rk V represents the received signal steering vector of the kth target; t is the spatial response of the transmit array, denoted as V t (r t ,η t )=[v t1 (r t1 ,η t1 ),…,v tk (r tk ,η tk )],v tk (r tk ,η tk ) is the spatial response of the transmitted signal of the kth target, expressed as V r is the spatial response of the receiving array, V r (r r ,η r )=[v r1 (r r1 ,ηr1 ),…,v rk (r rk ,η rk )],v rk (r rk ,η rk ) is the spatial response of the received signal of the kth target, expressed as B=[b (1) ,…,b (L) ],in and The scattering coefficient of the kth target is expressed as N=[n (1) ,…,n (L) ],in is the filtered noise vector;

[0070] From the above formula, we can see that Y can be regarded as a fifth-order PARAFAC tensor decomposition model, so Y can be expressed as follows:

[0071]

[0072] Among them, Γ 5,K×1 It is a symbolic representation without specific meaning; is the tensor form of N; Q t×2 ,V t×3 ,Q r×4 ,V r×5 With the above Q t ,V t ,Q r ,V r Meaning the same; B = [b (1) ,…,b (L) ],in and The scattering coefficient of the kth target can be expressed as

[0073] S6. Considering that there is a relatively mature fast algorithm for the third-order PARAFAC decomposition model, the complex parallel factor decomposition (COMFAC), this algorithm considers converting the above-mentioned fifth-order parallel factor model into a third-order PARAFAC model. According to the definition of PARAFAC model rearrangement, let O1 = {1, 2}, O2 = {3, 4}, O3 = {5}, then Y can be rearranged into a third-order PARAFAC tensor decomposition model:

[0074]

[0075] Among them, Γ 3,K×1 It is just a symbol, without any specific meaning; A t×2 ,Ar×3 and A in formula (1) t , A r Same meaning; is the tensor form of N; B=[b (1) ,…,b (L) ],in and The scattering coefficient of the kth target can be expressed as

[0076] According to the definition of parallel factor modulo n matrix expansion, the modulo n matrix expansion of Y is

[0077]

[0078]

[0079] Y3=(A t ⊙A r )B T (20)

[0080] Use estimated value Alternative theoretical value Y n , then the PARAFAC decomposition of the tensor Y can be completed through joint optimization:

[0081]

[0082]

[0083]

[0084] For this type of optimization problem, the trilinear alternating least squares algorithm is generally used. Its basic idea is to assume that A t ,A r ,B, any two are known, then one matrix can be updated in each step, that is, for the remaining matrices, according to the results of the previous estimation, the least squares method is used to update them, and the other matrices are updated, and the above steps are repeated until the algorithm converges. t ,A r , the least squares estimates of B are

[0085]

[0086]

[0087]

[0088] in, Represents the matrix pseudo-inverse. Since the alternating least squares algorithm is sensitive to the initial value, this algorithm uses the COMFAC algorithm to speed up the convergence;

[0089] S7. When PARAFAC decomposition is completed, A t ,A r and the estimated values of B, respectively, are denoted as and Its column fuzziness and scale fuzziness can be expressed as:

[0090]

[0091]

[0092]

[0093] Where Π is a permutation matrix, Δ1, Δ2 and Δ3 are the corresponding scale fuzzy matrices, all of which are diagonal matrices, and Δ1Δ2Δ3 = Ι, N1, N2 and N3 are error matrices;

[0094] S8, let J t1 =[Ι 2M ,0 2M ], J t2 =[0 2M ,I 2M ], where I 2M is the identity matrix of 2M×2M, 0 2M is a 2M×2M zero matrix, then the transmit array steering vector Q t satisfy

[0095] J t2 Q t =J t1 Q t Ψ t (30)

[0096] Because A t =Q t ⊙V t , so we can get

[0097]

[0098] Where Ι2 is the 2×2 dimensional identity matrix, A t is the steering vector of the transmitting array;

[0099] Ignoring the error term in equation (27), we can obtain from equations (27) and (31):

[0100]

[0101] in, A is obtained by using the COMFAC algorithm tThe estimated value of ; Π is a permutation matrix, Δ1 is the scale fuzzy matrix, which is a diagonal matrix;

[0102] S9, command From formula (32), we can get

[0103]

[0104] right Performing eigendecomposition, we can get the eigenvector and eigenvalue as P respectively. t =[o t1 ,o t2 ,…,o tK ], R t =diag{λ t1 ,λ t2 ,…λ tK}; From formula (33), we can see that the estimated value of Π is

[0105]

[0106] Among them, round{·} is a rounding function. Re{P t} means taking P t The real part of

[0107] S10, so we get V t The estimated value is

[0108]

[0109] in, Indicated by The submatrix consisting of the 2m-1th row to the 2mth row;

[0110] Then the polarization parameters of the transmitting array of the kth target can be calculated by the following two formulas:

[0111]

[0112]

[0113] in, express The element in the second row and k column of express The element in the first row and k column of ;

[0114] S11. Due to So we get

[0115]

[0116] Among them, unvec(·) is a matrix operator, which is the inverse operation of vec(·);

[0117] S12, order in is a permutation matrix with all elements on the diagonal being 1 and all other elements being 0, so

[0118]

[0119] From the above formula, we can get

[0120]

[0121] w tk The estimated value of can be calculated by the following formula

[0122]

[0123] Then the two-dimensional emission angle of the kth target is

[0124]

[0125] S13, From formula (27) and formula (28), we can know that and With the same permutation matrix, we can calculate

[0126]

[0127] It can be further calculated that

[0128]

[0129] Among them, J r1 =[Ι 2N ,0 2N ], J r2 =[0 2N ,I 2N ], I 2N is the 2N×2N identity matrix, 0 2N is a 2N×2N zero matrix; then we can get

[0130]

[0131] Among them, Δ2 is the scale fuzzy matrix, which is a diagonal matrix, Ψ r is a K×K dimensional matrix;

[0132] make From the formula we can get

[0133]

[0134] So we get Vr The estimated value is

[0135]

[0136] Then the receiving array polarization parameters of the kth target can be calculated by the following two formulas:

[0137]

[0138]

[0139] because So we get

[0140]

[0141] where unvec(·) is a matrix operator, which is the inverse operation of vec(·);

[0142] make in is a permutation matrix with all elements on the diagonal being 1 and all other elements being 0, so

[0143]

[0144] From the above formula, we can get

[0145]

[0146] w rk The estimated value of can be calculated by the following formula

[0147]

[0148] Then the two-dimensional acceptance angle of the kth target is

[0149]

[0150] according to Figure 1 From the geometric relationship of the schematic diagram of the bistatic cross-dipole-magnetic ring antenna uniform linear array MIMO radar shown in the figure, it can be seen that the position of the kth target is

[0151]

[0152]

[0153] Example: Assume M = N = 7, λ = 1, d t =d r =λ / 4,d=2×λ,L=512. Assume that there are three targets in the near field space of MIMO radar, their (θ t ,θ r , γ t, γ r , η t , η r ,y k , z k ) are (30°, 120°, 10°, 42°, 36°, 17°, 1, 1.7321), (45°, 135°, 22°, 33°, 48°, 27°, 0, 2), (60°, 150°, 49°, 38°, 56°, 29°, -1, 1.7321); Assuming that the signal-to-noise ratio (SNR) of the three targets is 30 dB, 6 simulation experiments are conducted. The parameter estimation results of the near-field targets are shown in the figure. Figure 2 、 Figure 3 、 Figure 4 and Figure 5 As shown in the figure, the experimental results show that the algorithm can correctly identify the three target sources, and the eight-dimensional parameters can achieve automatic pairing with high estimation accuracy.

Claims

1. A near-field polarimetric MIMO radar parameter estimation method based on parallel factorization, characterized by: The method comprises the following steps: S1. Establish a system model for a bistatic MIMO radar based on a uniform linear array of crossed dipole-magnetic loop antennas. In the system model, the transmitter consists of a uniform linear transmitting array composed of 2M+1 crossed dipole-magnetic loop antennas, and the receiver consists of a uniform linear receiving array composed of 2N+1 crossed dipole-magnetic loop antennas. The spacing between the transmitting and receiving elements is represented by dt and dr, respectively, and satisfies dt≤λ / 2 and dr≤λ / 2, where λ is the wavelength of the electromagnetic wave. S2. Position K uncorrelated near-field narrowband fully polarized targets. The two-dimensional transmission angle and two-dimensional receiving angle of the k-th target are respectively expressed as θ tk and θ rk Denotes, where k = 1, 2, ... K; S3. During the first snapshot, the data received by the receiving element is obtained, which is expressed as: X l =A r (θ rk ,ρ rk ,r rk ,η rk )diag{b (l) }A t T (θ tk ,ρ tk ,r tk ,η tk )S+Z (l) , where A r (θ rk ,ρ rk ,r rk ,η rk ) is the steering vector of K receiving arrays, A t (θ tk ,ρ tk ,r tk ,η tk ) is the steering vector of the K transmitting arrays, S is the transmitting waveform matrix; b (l) is the scattering coefficient, Z (l) is the noise matrix, ρ rk is the distance from the receiving array to the target, r rk is the two-dimensional receiving polarization angle of the kth target, η rk is the two-dimensional receiving phase difference of the kth target, θ rk is the two-dimensional acceptance angle of the k-th target, θ tk is the two-dimensional emission angle of the kth target, ρ tk is the distance from the transmitting array to the target, r tk is the two-dimensional emission polarization angle of the kth target, η tk is the two-dimensional emission phase difference of the kth target, where 0≤r tk ,r rk ≤π / 2, 0≤η tk ,η rk ≤2π; S4. Under L snapshots, obtain the output signal matrix of the receiving end matched filter, which is expressed as: Y = (Q t ⊙V t ⊙Q r ⊙V r )B+N, where Q t is the transmission signal steering vector; Q r is the received signal steering vector; V t is the spatial response of the transmitting array; V r is the spatial response of the receiving array; B = [b (1) ,…,b (L) ]; N = [n (1) ,…,n (L) ],in is the filtered noise vector; S5. Arrange the output signal matrix of the receiving end matched filter in step (4) into a fifth-order PARAFAC tensor decomposition model, expressed as: Then convert the fifth-order PARAFAC tensor decomposition model into a third-order PARAFAC tensor decomposition model, namely: S6. Based on the third-order PARAFAC tensor decomposition model, the complex parallel factor decomposition algorithm is used to decompose A t ,A r and B are estimated to obtain A t ,A r The estimated values of and B are respectively denoted as and S7, according to the expression Come to Perform eigendecomposition to obtain the eigenvector P t =[o t1 ,o t2 ,…,o tK ] and eigenvalue R t =diag{λ t1 ,λ t2 ,…λ tK }, where Ψ t is a K×K dimensional matrix, Δ1 is the scale fuzzy matrix, J t1 =[Ι 2M ,0 2M ], J t2 =[0 2M ,I 2M ], I 2M is the identity matrix of 2M×2M, 0 2M is a 2M×2M zero matrix, Ι2 is a 2×2 dimensional identity matrix, A t is the steering vector of the transmitting array; S8. The estimated value of π is obtained based on the obtained eigenvector and eigenvalue. Among them, round{·} is a rounding function, Re{P t } means taking P t The real part of π; V is obtained based on the estimated value of π t The estimated value of is: in, Indicated by The submatrix consisting of the 2m-1th row to the 2mth row; according to V t The estimated value of is used to derive the two-dimensional emission phase difference of the kth target: The two-dimensional emission polarization angle of the kth target is: in, express The element in the second row and k column of express The element in the first row and k column of ; S9. According to the expression Get w tk The estimated value of is: According to w tk The estimated value of is the two-dimensional launch angle of the kth target: S10, according to the expression Get V r The estimated value of is: Among them, J r1 =[Ι 2N ,0 2N ], J r2 =[0 2N ,I 2N ], I 2N is the 2N×2N identity matrix, 0 2N is a 2N×2N zero matrix, Δ2 is the scale fuzzy matrix, Ψ r It is a K×K dimensional matrix, according to V r The estimated value of gives the two-dimensional receiving phase difference of the k-th target as: The two-dimensional receiving polarization angle of the kth target is: in, express The element in the second row and k column of express The element in the first row and k column of ; S11. According to the expression Get w rk The estimated value of is: According to w rk The estimated value of the two-dimensional acceptance angle of the k-th target is: The position of the kth target is:

2. The method for near-field polarization MIMO radar parameter estimation based on parallel factorization according to claim 1, wherein: In step S2, when locating K uncorrelated near-field narrowband fully polarized targets, 2M+1 transmitting array elements simultaneously transmit 2M+1 orthogonal signals at the transmitting end. After encountering the target, the signals are reflected to form echoes received by the receiving array elements. It is also set that within the same signal period, the cross-sectional scintillation of the target remains constant, the cross-sectional scintillation fluctuations of different targets are uncorrelated, and the fluctuation statistics within different pulse times are also independent of each other.

3. The method for near-field polarization MIMO radar parameter estimation based on parallel factorization according to claim 1, wherein: In step S9, w is obtained tk The specific process of estimating the value is as follows: First, according to get: where unvec(·) is a matrix operator, which is the inverse operation of vec(·); in is a permutation matrix with all elements on the diagonal being 1 and all other elements being 0, and we can further obtain: Finally, we get w tk estimated value.

4. The method for near-field polarization MIMO radar parameter estimation based on parallel factorization according to claim 1, wherein: In step S10, V r The specific process of estimating the value is as follows: First, according to and With the same permutation matrix we get: According to the conditions and get: Finally, we get V r estimated value.

5. The method for near-field polarization MIMO radar parameter estimation based on parallel factorization according to claim 1, wherein: In step S11, w is obtained rk The specific process of estimating the value of is as follows: get: Among them, unvec(·) is a matrix operator, which is the inverse operation of vec(·); through in is a permutation matrix with 1 on the secondary diagonal and 0 on the rest of the elements, then we get: Finally, we get w rk estimated value.

Images