Search

CN-121978615-A - Array signal processing method, device, equipment and storage medium

CN121978615ACN 121978615 ACN121978615 ACN 121978615ACN-121978615-A

Abstract

The application relates to an array signal processing method, device, equipment and storage medium in the technical field of array signal processing, wherein the method comprises the steps of obtaining a sampling covariance matrix of an array signal, constructing a CMSP-TNN optimization model by applying conjugate symmetry constraint conditions to the sampling covariance matrix, solving the CMSP-TNN optimization model to obtain a low rank matrix corresponding to the sampling covariance matrix, and estimating a direction of arrival of the array signal based on the low rank matrix. According to the method, the inherent conjugate symmetry characteristics of the signal covariance matrix are fused, the conjugate symmetry constraint condition is synchronously applied in the low-rank matrix recovery process, and the recovered covariance matrix not only maintains the low-rank essence, but also strictly accords with the structural characteristics of the signal covariance matrix by constructing the optimized model special for the signal processing scene, so that the matrix recovery precision is remarkably improved, and the DOA estimation robustness in the complex electromagnetic environment is improved.

Inventors

  • GAO YUNFENG
  • LI WUTAO
  • ZHANG CHI
  • WANG ZHENMING
  • Liu Shuocen
  • FAN BING
  • Wang sihang

Assignees

  • 北京遥感设备研究所

Dates

Publication Date
20260505
Application Date
20251229

Claims (10)

  1. 1. An array signal processing method, comprising: acquiring a sampling covariance matrix of an array signal; constructing a CMSP-TNN optimization model by applying conjugate symmetry constraint conditions to the sampling covariance matrix; solving the CMSP-TNN optimization model to obtain a low-rank matrix corresponding to the sampling covariance matrix; and estimating the direction of arrival of the array signal based on the low rank matrix.
  2. 2. The method of claim 1, wherein the applying conjugate symmetry constraints to the sampling covariance matrix comprises: The error between the ideal matrix of the low rank matrix corresponding to the sampling covariance matrix and the conjugate symmetry function of the sampling covariance matrix is minimized.
  3. 3. The method of claim 2, wherein constructing a CMSP-TNN optimization model by applying conjugate symmetry constraints to the sampling covariance matrix comprises: Applying conjugate symmetry constraint conditions to the sampling covariance matrix to construct an error matrix; the magnitude of the error matrix is calculated using the l 1 norm and added to the objective function of the TNN optimization model, when the CMSP-TNN optimization model.
  4. 4. The method of claim 1, wherein the CMSP-TNN optimization model is: In the middle of All are complex matrices, Z is a low rank matrix, Z * is the kernel norm of matrix Z, E is a sparse matrix, E 1 is the l 1 norm of matrix E, X is a sampling covariance matrix, lambda and gamma are regularization parameters, tr (AZB H ) are the traces of matrix, A and B are truncated left and right singular matrices of matrix X, Z-J N X * J N is an error matrix, and Z-J N X * J N || 1 is the l 1 norm of the error matrix.
  5. 5. The method of claim 4, wherein solving the CMSP-TNN optimization model comprises solving the CMSP-TNN optimization model by an iterative method, further comprising: The first stage, fixing the matrixes Z and E, and calculating matrixes A and B; in the second stage, matrices Z and E are updated with matrices A and B fixed.
  6. 6. The method of claim 5, wherein at a first iteration, the first stage comprises: For fixed Z l and E l , calculate matrix X l ＝Z l +E l , perform singular value decomposition on matrix X l , X l ＝U l Σ l V l H , where U l ＝(u 1 ,u 2 ,...,u M ),V l ＝(v 1 ,v 2 ,...,v M ); The left singular vectors corresponding to the first r largest singular values are selected from U l to form A l , and the right singular vectors corresponding to the first r largest singular values are selected from V l to form B l .
  7. 7. The method of claim 1, wherein the performing direction of arrival estimation on the array signal based on the low rank matrix comprises: and estimating the direction of arrival of the array signal by adopting a multiple signal classification algorithm MUSIC or a rotation invariant subspace algorithm ESPRIT based on the low rank matrix.
  8. 8. An array signal processing apparatus, comprising: An acquisition unit for acquiring a sampling covariance matrix of the array signal; The optimization model construction unit is used for constructing a CMSP-TNN optimization model by applying conjugate symmetry constraint conditions to the sampling covariance matrix; the low-rank matrix acquisition unit is used for solving the CMSP-TNN optimization model to obtain a low-rank matrix corresponding to the sampling covariance matrix; And the direction-of-arrival estimation unit is used for estimating the direction of arrival of the array signal based on the low-rank matrix.
  9. 9. An electronic device comprising a memory, a processor and a computer program stored in the memory and running on the processor, characterized in that the processor implements the steps of the method according to any of claims 1 to 7 when the computer program is executed.
  10. 10. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the method according to any one of claims 1 to 7.

Description

Array signal processing method, device, equipment and storage medium Technical Field The application belongs to the technical field of array signal processing, and particularly relates to an array signal processing method, device, equipment and storage medium. Background In the field of array signal processing, the estimation technique of direction of arrival (Direction of Arrival, DOA) based on subspace decomposition is widely used due to its super-resolution characteristic. The general algorithm is based on the theory of characteristic decomposition of the covariance matrix of the received signal. The method constructs a space spectrum function to perform peak search by constructing orthogonal characteristics of a noise subspace and a signal subspace of an ideal covariance matrix, and finally realizes DOA parameter estimation. However, there are two typical constraints in practical engineering applications (1) when the source is too far from the array spacing resulting in significant attenuation of the beamforming gain and (2) when the device sampling capability is limited in a dynamic monitoring environment resulting in an insufficient number of active snapshots. Under the two scenes, the actual sampling covariance matrix of the received signal can deviate from the statistical characteristics of an ideal mathematical model seriously, and the method is characterized in that the dispersion effect of matrix eigenvalues is aggravated under the condition of limited samples, the orthogonality of a noise subspace and a signal subspace is destroyed, and then the pseudo-peak phenomenon and main lobe distortion of a spatial spectrum function are caused. Therefore, the accuracy of DOA estimation of array signals is currently difficult to guarantee. Disclosure of Invention The application aims to provide an array signal processing method, device, equipment and storage medium, so as to improve DOA estimation robustness in a complex electromagnetic environment. In a first aspect of an embodiment of the present application, there is provided an array signal processing method, including: acquiring a sampling covariance matrix of an array signal; Constructing a CMSP-TNN optimization model by applying conjugate symmetry constraint conditions to the sampling covariance matrix; Solving a CMSP-TNN optimization model to obtain a low-rank matrix corresponding to the sampling covariance matrix; direction of arrival estimation is performed on the array signals based on the low rank matrix. In one embodiment, applying a conjugate symmetry constraint to the sampling covariance matrix includes: The error between the ideal matrix of the low rank matrix corresponding to the sampling covariance matrix and the conjugate symmetry function of the sampling covariance matrix is minimized. In one embodiment, constructing the CMSP-TNN optimization model by applying conjugate symmetry constraints to the sampling covariance matrix includes: applying conjugate symmetry constraint conditions to the sampling covariance matrix to construct an error matrix; the magnitude of the error matrix is calculated using the l 1 norm and added to the objective function of the TNN optimization model, when the CMSP-TNN optimization model. In one embodiment, the CMSP-TNN optimization model is: s.t.Z+E=X In the middle of All are complex matrices, Z is a low rank matrix, Z * is the kernel norm of matrix Z, E is a sparse matrix, E 1 is the l 1 norm of matrix E, X is a sampling covariance matrix, lambda and gamma are regularization parameters, tr (AZB H) are the traces of matrix, A and B are truncated left and right singular matrices of matrix X, Z-J NX*JN is an error matrix, and Z-J NX*JN||1 is the l 1 norm of the error matrix. In one embodiment, solving the CMSP-TNN optimization model includes solving the CMSP-TNN optimization model by an iterative method, further comprising: The first stage, fixing the matrixes Z and E, and calculating matrixes A and B; in the second stage, matrices Z and E are updated with matrices A and B fixed. In one embodiment, at the first iteration, the first phase comprises: For fixed Z l and E l, calculate matrix X l＝Zl+El, singular value decompose matrix X lWherein the method comprises the steps of Ul＝(u1,u2,...,uM),Vl＝(v1,v2,...,vM); The left singular vectors corresponding to the first r largest singular values are selected from U l to form A l, and the right singular vectors corresponding to the first r largest singular values are selected from V l to form B l. In one embodiment, performing direction of arrival estimation on an array signal based on a low rank matrix comprises: and estimating the direction of arrival of the array signals by adopting a multiple signal classification algorithm MUSIC or a rotation invariant subspace algorithm ESPRIT based on the low rank matrix. In a second aspect of an embodiment of the present application, there is provided an array signal processing apparatus including: An acquisition unit for acquiri