CN-122001410-A - Mixed source positioning method considering mutual coupling effect under analog-digital mixed architecture

CN122001410ACN 122001410 ACN122001410 ACN 122001410ACN-122001410-A

Abstract

The invention relates to a hybrid source positioning method considering the mutual coupling effect under an analog-digital hybrid architecture, which comprises the steps of firstly dynamically adjusting a switching state and a phase regulator of antenna connection, reconstructing a covariance matrix by utilizing correlation among received signals, secondly estimating DOA of a far-field source by using a MUSIC algorithm through the reconstructed covariance matrix, simultaneously compensating the mutual coupling effect, then reconfiguring the antenna connection by utilizing estimated far-field DOA information, calculating and compensating the mutual coupling coefficient, and finally separating near-field signals by using an oblique projection method by utilizing the compensated covariance matrix to perform two-dimensional DOA and distance estimation.

Inventors

TIAN YE
CAO YUETING
ZHAO HONGYUN
ZHU YILI
Lin Juntuo

Assignees

宁波大学

Dates

Publication Date: 20260508
Application Date: 20251125

Claims (8)

1. A mixed source positioning method considering mutual coupling effect under an analog-digital mixed architecture is characterized by comprising the following steps: dynamically configuring a switching network and a phase shifter in an HADS architecture, acquiring the correlation of output signals of different radio frequency links, and reconstructing a space covariance matrix in a block mode; Symmetrical Toeplitz structure based on mutual coupling matrix, and constructing a selection matrix Substituting the reconstructed space covariance matrix into the selection matrix To inhibit the cross coupling effect and obtain a corrected space covariance matrix; performing eigenvalue decomposition on the corrected space covariance matrix to obtain a corresponding signal subspace And noise subspace ; Constructing classical MUSIC spatial spectrum function by using the noise subspace, and fixing distance parameters in two-dimensional MUSIC pseudo spectrum Obtaining a one-dimensional spectrum function, searching a peak value of the one-dimensional spectrum function, and obtaining an arrival direction estimated value of a far-field signal source; Based on the arrival direction estimated value of the far-field signal source, establishing an overdetermined linear equation set by reconstructing a connection mode of an antenna and a radio frequency link, and estimating an unknown mutual coupling coefficient of an antenna array by solving the overdetermined linear equation set; and compensating the reconstructed space covariance matrix by using the estimated mutual coupling coefficient, separating a near-field information source signal by adopting an oblique projection operator, and further estimating the arrival direction and the arrival distance of the near-field information source.
2. The method for positioning a hybrid source under an analog-digital hybrid architecture by considering the mutual coupling effect according to claim 1, wherein the specific process of reconstructing the spatial covariance matrix in a block manner is as follows: Defining a spatial covariance matrix as , wherein, In order to make a statistical expectation, For the purpose of conjugate transposition, , , , , Representing the far-field source signal vector, Representing a near-field source signal vector, Representing a noise vector by combining the spatial covariance matrix Divided into With dimensions of Is a sub-matrix of (2) , wherein, Indicating the number of radio frequency links, Indicating the number of antennas connected for each radio frequency link, ; For each sub-matrix Obtaining all diagonal elements of the object; For each sub-matrix Acquiring all off-diagonal elements of the object; According to all diagonal elements and off-diagonal elements of all obtained submatrices, all submatrices are corresponding to the space covariance matrix according to the submatrices And (3) splicing the row and column positions to reconstruct the space covariance matrix.
3. The method for mixed source positioning considering mutual coupling effect under an analog-digital mixed architecture according to claim 2, wherein for each submatrix The specific process for obtaining all diagonal elements is as follows: For submatrices Performs the following diagonal element estimation operations for each diagonal element of: For submatrices The mth RF link and the nth RF link of (a) are set to have their switching vectors set to be , wherein, Representing a Q-dimensional unit vector, wherein only the Q-th bit of the Q-dimensional unit vector is 1, and the rest is 0; , the mth RF link and the nth RF link receive only the signal of the qth antenna in their corresponding sub-matrix; Collecting output signals of an mth RF link and an nth RF link And , wherein, , Calculating the sample cross-correlation value of two output signals for the snapshot number, and taking the sample cross-correlation value as a submatrix Is the first of (2) First, the Estimation of diagonal elements of a column ; Indexing diagonal elements From 1 traversal to To perform the diagonal element estimation operation to obtain a submatrix A kind of electronic device Diagonal elements.
4. A hybrid source positioning method considering mutual coupling effect under an analog-digital hybrid architecture as claimed in claim 2 or 3, wherein said method is characterized by comprising the following steps for each sub-matrix The specific process for obtaining all the off-diagonal elements is as follows: For submatrices Performs the following off-diagonal element estimation operations for each off-diagonal element in (a): Configuring a switch vector So as to activate the q antenna and the u antenna in the sub-matrix corresponding to the m RF link at the same time, and the q antenna and the u antenna in the sub-matrix corresponding to the n RF link, ; Configuring the phase shifter vector as Generating a weight matrix ; Receiving and calculating the output signal correlation of the mth RF link and the nth RF link to obtain a first correlation measurement value of the received signals of the two RF links ; Maintaining the configuration of the switch vector, adjusting the phase shifter vector to shift the signal phase of the u-th antenna Signal phase offset of the q-th antenna The rest antennas have no offset, and weight matrix is generated Wherein, the method comprises the steps of, , A switch connection vector representing the q-th row and column elements of the covariance of the estimation sub-matrix, Is a unit vector with 0 except the u element being 1, Is a unit vector with 0 except the q-th element being 1; , a phase shift vector representing the q columns of the nth row of estimated sub-covariance, Is a full 1-phase vector of the vector, And All of which represent the phase disturbance factor, Is a unit vector with 0 except the u element being 1, Is a unit vector with 0 except the q-th element being 1; Receiving and calculating the output signal correlation of the mth RF link and the nth RF link to obtain a second correlation measurement value of the received signals of the two RF links ; Based on the first correlation measurement And a second correlation measurement By using Is a binary linear system of equations is constructed: ; solving the binary linear equation system to obtain the submatrix Middle position Off-diagonal elements at Is a function of the estimated value of (2); Traversing a submatrix All of them satisfy And performing the off-diagonal element estimation operation to obtain all off-diagonal element estimates.
5. The method for positioning mixed sources considering mutual coupling effect under an analog-digital mixed architecture as claimed in claim 4, wherein said symmetric Toeplitz structure based on mutual coupling matrix constructs a selection matrix The expression of (2) is: , representing dimensions as Is used for the zero-matrix of (c), Representing an (M-2P) order identity matrix, wherein the expression of the corrected spatial covariance matrix is as follows: , wherein, Representing the fully reconstructed spatial covariance matrix, C representing the cross-coupling matrix, a representing the ideal steering matrix, Representing the source covariance matrix, Representing the variance of the noise of the sub-matrix, Represents an (M-2P) order identity matrix, Representing the sub-matrix steering matrix, Representing the conjugate transpose matrix.
6. The method for mixed source positioning considering mutual coupling effect under an analog-digital mixed architecture as claimed in claim 5, wherein said signal subspace For the feature vector corresponding to the first L maximum feature values of the corrected space covariance matrix, the noise subspace Is left to And feature vectors corresponding to the smaller feature values.
7. The method for positioning a hybrid source under an analog-digital hybrid architecture by considering mutual coupling effect according to claim 6, wherein the specific process of estimating the unknown mutual coupling coefficient of the antenna array is as follows: for each mutual coupling coefficient to be estimated Performing an antenna connection reconfiguration of switching a first RF chain to a first of the corresponding sub-matrices A root antenna, wherein, While the remaining RF chains remain connected to the antenna corresponding to the first element in the respective submatrix; In this configuration, a corresponding selection matrix is obtained The selection matrix The connection state between the RF link and the antenna in the p-th reconfiguration is described as follows: the method comprises the steps of collecting a received signal of a radio frequency link in the p-th reconstruction configuration at the moment, wherein the received signal is specifically expressed as: wherein, the method comprises the steps of, Representing the mutual coupling matrix of the p-th reconfiguration , , Representing the array popularity matrix after antenna selection, Representing the noise vector after the antenna selection, Representing a source signal vector; calculating the space covariance matrix of the received signal, and decomposing the eigenvalue to obtain a corresponding noise subspace matrix ; Symmetric Toeplitz structure based on mutual coupling matrix, array popularity matrix after antenna selection Is a vector for each array of array steering Modeling is as follows: wherein, the method comprises the steps of, , Representing a structured matrix associated with a direction, Representing the mutual coupling matrix at the p-th reconfiguration; based on the direction of arrival value of each far-field signal source And according to the array steering vector after being affected by the cross coupling effect And noise subspace matrix The orthogonality principle of (2) is obtained: Establishing an overdetermined linear equation set: solving the overdetermined linear equation to obtain a mutual coupling coefficient in the p-th reconfiguration configuration Is a function of the estimated value of (2); Updating the antenna connection reconfiguration to obtain the estimated values of all unknown mutual coupling coefficients, thereby obtaining a mutual coupling matrix 。
8. The method for positioning a hybrid source under an analog-digital hybrid architecture by considering the mutual coupling effect according to claim 7, wherein the method is characterized in that the reconstructed spatial covariance matrix is compensated by using the estimated mutual coupling coefficient, and a near-field information source signal is separated by using an oblique projection operator, so as to estimate the arrival direction and the arrival distance of the near-field information source, and comprises the following specific steps: Compensating the reconstructed complete space covariance matrix R by using the estimated mutual coupling matrix C to obtain a compensated space covariance matrix; constructing a signal subspace of the far-field signal source based on the estimated arrival direction values of all the far-field signal sources; constructing an oblique projection operator by utilizing the signal subspace of the far-field information source, wherein the oblique projection operator is specifically expressed as follows: Wherein Representing the steering matrix of the far-field source, Representing the pseudo-inverse of the compensated spatial covariance matrix, Represents the conjugate transpose matrix of the matrix, Representing an inverse matrix; The oblique projection operator acts on the compensated space covariance matrix to inhibit far-field signal components in the space covariance matrix, and meanwhile near-field signal components are reserved to obtain a separated near-field signal covariance matrix, wherein the separated near-field signal covariance matrix is specifically expressed as: wherein, the method comprises the steps of, , Representing the spatial covariance matrix after compensation, Represents an identity matrix of order M, The oblique projection operator is represented as such, Represents the conjugate transpose of the object, Representing the estimated cross-coupling matrix, Representing an inverse matrix; performing eigenvalue decomposition on the near-field signal covariance matrix, and extracting a noise subspace thereof ; Using the noise subspace Construction of two-dimensional MUSIC spatial spectrum function for near-field parameter estimation The method is specifically expressed as follows: wherein, the method comprises the steps of, Representing the steering vector of the near field source, The noise subspace is represented as such, Represents the conjugate transpose of the object, Representing an inverse matrix; In parameter space Upper search space spectrum function To obtain the estimated value of the arrival direction and the estimated value of the distance of the near-field source.

Description

Mixed source positioning method considering mutual coupling effect under analog-digital mixed architecture Technical Field The invention relates to the technical field of wireless communication and signal processing, in particular to a hybrid source positioning method considering the mutual coupling effect under an analog-digital hybrid architecture. Background Estimation of direction of arrival (DOA) based antenna arrays is a core technology for array signal processing, providing critical spatial information for radar, sonar, and wireless communication systems. Traditional high resolution algorithms, such as MUSIC and ESPRIT, achieve fine angular resolution by sub-spatially decomposing the Spatial Covariance Matrix (SCM). However, its performance depends on idealized hardware assumptions, in particular requiring that each antenna element be equipped with a dedicated Radio Frequency (RF) link. As systems expand towards massive MIMO configurations, these assumptions are difficult to implement due to hardware cost and complexity issues. To address this fundamental challenge, hybrid analog-digital architecture (HADS) architectures have evolved. The architecture strategically combines a large-scale antenna array with a limited number of radio frequency links through a configurable phase shifter and switching network, establishing critical connections between the antennas and sparse radio frequency links. This configuration achieves a practical balance between spectral efficiency, achievable spatial resolution, and hardware implementation constraints. While HADS provides a viable path for massive MIMO, it introduces fundamental limitations for direction of arrival (DOA) estimation in that analog signals received by multiple antenna elements need to be combined in the analog domain by phase shifters and switching networks before being digitized by limited radio frequency links. The direct consequence of this is that conventional DOA algorithms that rely on complete phased arrays (SCM) can suffer significant performance degradation and even failure, which presents challenges for applications requiring accurate spatial perceptibility. In addition to the limitations of the hybrid architecture itself, there is another layer of complexity in practical HADS deployments, namely the coexistence of Far Field (FF) and Near Field (NF) sound sources. Conventional localization algorithms generally assume that the sound source is present only in the far field region, which can satisfy planar wavefront approximation conditions. However, the continued expansion of the array aperture and the increase in operating frequency inevitably expands the near field region. In a practical scenario, the signal sources are therefore necessarily distributed at different distances relative to the array, where some of the targets may be in the far field (planar wavefront) region and others in the near field (spherical wavefront) region, requiring estimation of their distance in addition to angle of arrival. For applications such as radar target tracking and acoustic source positioning, accurate joint estimation of the arrival direction of a far-field source and two-dimensional parameters of the near-field source is important. Mutual Coupling (MC) between antenna elements is another key non-ideal factor that plagues the performance of practical antenna arrays, including HADS. MC results from electromagnetic interactions between closely packed cells, resulting in significant distortion of the actual array pointing vector under the influence of MC. Because of the extremely high sensitivity of conventional high resolution direction of arrival (DOA) algorithms to array manifold accuracy, unaccounted for MC effects will lead to severe performance degradation, including angular estimation bias and resolution degradation. Disclosure of Invention The invention aims to solve the technical problem of providing a positioning method capable of realizing high-precision joint positioning of a far-near field mixed information source under the condition that an unknown mutual coupling effect exists in a large-scale MIMO system with a low-cost mixed analog-digital structure (HADS). The invention adopts the technical scheme that the method for positioning the mixed source under the analog-digital mixed architecture considers the mutual coupling effect comprises the following steps: dynamically configuring a switching network and a phase shifter in an HADS architecture, acquiring the correlation of output signals of different radio frequency links, and reconstructing a space covariance matrix in a block mode; Symmetrical Toeplitz structure based on mutual coupling matrix, and constructing a selection matrix Substituting the reconstructed space covariance matrix into the selection matrixTo inhibit the cross coupling effect and obtain a corrected space covariance matrix; performing eigenvalue decomposition on the corrected space covariance matrix to obtain a corresp