WO2021068494A1 - 基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法 - Google Patents

基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法 Download PDF

Info

Publication number
WO2021068494A1
WO2021068494A1 PCT/CN2020/088567 CN2020088567W WO2021068494A1 WO 2021068494 A1 WO2021068494 A1 WO 2021068494A1 CN 2020088567 W CN2020088567 W CN 2020088567W WO 2021068494 A1 WO2021068494 A1 WO 2021068494A1
Authority
WO
WIPO (PCT)
Prior art keywords
array
tensor
virtual domain
signal
axis
Prior art date
Application number
PCT/CN2020/088567
Other languages
English (en)
French (fr)
Inventor
周成伟
郑航
陈积明
史治国
Original Assignee
浙江大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 浙江大学 filed Critical 浙江大学
Priority to PCT/CN2020/088567 priority Critical patent/WO2021068494A1/zh
Priority to JP2021541668A priority patent/JP7044289B2/ja
Publication of WO2021068494A1 publication Critical patent/WO2021068494A1/zh
Priority to US17/395,478 priority patent/US11300648B2/en

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S3/00Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
    • G01S3/02Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using radio waves
    • G01S3/14Systems for determining direction or deviation from predetermined direction
    • G01S3/143Systems for determining direction or deviation from predetermined direction by vectorial combination of signals derived from differently oriented antennae
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S3/00Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
    • G01S3/02Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using radio waves
    • G01S3/74Multi-channel systems specially adapted for direction-finding, i.e. having a single antenna system capable of giving simultaneous indications of the directions of different signals
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S3/00Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
    • G01S3/02Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using radio waves
    • G01S3/023Monitoring or calibrating
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S3/00Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
    • G01S3/02Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using radio waves
    • G01S3/04Details
    • G01S3/043Receivers
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S3/00Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
    • G01S3/02Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using radio waves
    • G01S3/04Details
    • G01S3/046Displays or indicators
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S3/00Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
    • G01S3/02Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using radio waves
    • G01S3/14Systems for determining direction or deviation from predetermined direction
    • G01S3/46Systems for determining direction or deviation from predetermined direction using antennas spaced apart and measuring phase or time difference between signals therefrom, i.e. path-difference systems
    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01QANTENNAS, i.e. RADIO AERIALS
    • H01Q21/00Antenna arrays or systems
    • H01Q21/06Arrays of individually energised antenna units similarly polarised and spaced apart
    • H01Q21/061Two dimensional planar arrays

Definitions

  • the present invention belongs to the field of array signal processing technology, and in particular relates to a statistical signal processing technology based on planar coprime array tensor signals, in particular to a high-resolution accurate two-dimensional direction of arrival based on a planar coprime array virtual domain tensor space spectrum search
  • the estimation method can be used for passive detection and spatial positioning.
  • the planar coprime array As a two-dimensional sparse array with a systematic architecture, the planar coprime array has the characteristics of large aperture and high resolution. Compared with the traditional uniform array, it can achieve the comprehensive performance of the two-dimensional direction of arrival estimation in terms of estimation accuracy and resolution.
  • the planar coprime array by constructing a two-dimensional virtual domain, it is possible to perform signal processing that meets the Nyquist matching condition on a uniform surface array of the virtual domain, thereby solving the signal mismatch problem of the planar coprime array.
  • the spatial spectrum of the planar coprime array is constructed based on the virtual domain signal, and then through the two-dimensional peak search, an accurate two-dimensional direction of arrival estimation can be obtained.
  • the traditional method usually expresses the incident signal with two-dimensional spatial structure information as a vector, and calculates the second-order statistics of the multi-sampled signal in a time-averaged manner, and then derives the second-order virtual domain through vectorization, etc. Price signal.
  • the planar coprime array received signal and its virtual domain equivalent signal represented by a vector not only lose the multi-dimensional spatial structure information of the original signal, but also easily cause dimensional disasters as the amount of data increases, so this is the basic structure
  • the spatial spectrum and the two-dimensional direction of arrival estimation still have defects in accuracy and resolution.
  • the two-dimensional direction of arrival estimation method of planar coprime array based on tensor space spectrum search has begun to attract attention.
  • tensors can store the original multi-dimensional information of signals.
  • multi-dimensional algebraic theories such as high-order singular value decomposition and tensor decomposition also provide rich analysis tools for multi-dimensional feature extraction of tensor signals. Therefore, the tensor signal model can make full use of the multi-dimensional spatial structure information of the incident signal of the planar coprime array.
  • the existing method is still based on the actual received tensor signal for processing, and does not use the two-dimensional virtual domain of the planar coprime array to construct the tensor space spectrum, and does not solve the problem of signal mismatch of the planar coprime array, resulting in accuracy. Damaged; and the resolution of the generated spectral peaks is low, which is easy to cause mutual aliasing. Therefore, the existing methods still have much room for improvement in accuracy and resolution performance.
  • the purpose of the present invention is to solve the problems of loss of signal multi-dimensional spatial structure information and limited spatial spectrum resolution and accuracy performance in the above method, and propose a high-resolution accurate two-dimensional space spectrum search based on the virtual domain tensor spatial spectrum of the planar coprime array
  • the direction of arrival estimation method provides a feasible way to establish the connection between the planar coprime array tensor signal statistics and the virtual domain spatial spectrum, build a virtual domain tensor spatial spectrum search architecture, and achieve high-resolution, high-precision two-dimensional direction of arrival estimation.
  • a high-resolution accurate two-dimensional direction of arrival estimation method based on the virtual domain tensor space spectrum search of a planar coprime array including the following steps:
  • the receiving end uses 4M x M y + N x N y -1 physical antenna array elements, which are structured according to the structure of a planar coprime array; among them, M x , N x and My , N y are a pair respectively Coprime integers, and M x ⁇ N x , My y ⁇ N y ; the planar coprime array can be decomposed into two sparse uniform sub-arrays with
  • s k [s k,1 ,s k,2 ,...,s k,L ] T is the multi-shot sampling signal waveform corresponding to the k-th incident signal source
  • [ ⁇ ] T represents the transposition operation
  • represents the outer product of the vector
  • Is a noise tensor independent of each signal source
  • the steering vector in the x-axis and y-axis directions corresponds to the direction of the incoming wave as
  • the signal source is expressed as:
  • the received signal can be another three-dimensional tensor Expressed as:
  • x 1 (l) and x 2 (l) respectively represent the l-th slice of x 1 and x 2 in the third dimension (ie snapshot dimension) direction, and ( ⁇ ) * represents the conjugation operation;
  • the steering vector in the x-axis and y-axis directions corresponds to the direction of the incoming wave as Signal source; Represents the power of the k-th incident signal source; here, Represents the Kronecker product; the tensor subscript represents the modulus expansion operation of the PARAFAC decomposition of the tensor;
  • the steering vector in the x-axis and y-axis directions corresponds to the direction of the incoming wave as Signal source;
  • the orthogonal complement of the factor matrices C x and Cy is calculated; the orthogonal complement of C x is denoted as The orthogonal complement of C y is Among them, min( ⁇ ) represents the operation of taking the minimum value; then take As the noise subspace, use a tensor Represents the noise subspace, Means The h-th slice along the third dimension is expressed as:
  • ⁇ ⁇ Q ⁇ > represents the modulo ⁇ Q ⁇ shrinking operation of two tensors along the Q-th dimension, requiring the same size of the Q-th dimension of the two tensors;
  • ⁇ F represents the Frobenius norm; with A vector is obtained by condensing modulo ⁇ 1,2 ⁇ along the first and second dimensions Get the spatial spectrum function After that, the spatial spectrum corresponding to the two-dimensional search direction of arrival can be constructed, and then the two-dimensional direction of arrival corresponding to the position of the search spectrum peak is estimated to be the two-dimensional direction of arrival of the incident source.
  • the ideal (no noise scene) can be modeled as:
  • the uniform area array of the virtual domain described in step (5) Equivalent signal
  • the spatial structure information of the virtual domain array is saved in the, however, due to It can be regarded as a single snapshot of the virtual domain signal, and its statistics often have a rank deficit problem. Therefore, based on the idea of two-dimensional spatial smoothing, the virtual domain signal After processing, construct multiple equivalent snapshot virtual domain sub-array signals, and after summing and averaging these virtual domain sub-array signals, the fourth-order autocorrelation tensor is calculated. Subarray The position of the middle element is expressed as:
  • step (6) extraction of the multi-dimensional features of the fourth-order autocorrelation tensor of the virtual domain described in step (6) to realize the signal and noise subspace classification can be achieved by high-order singular value decomposition in addition to CANDECOMP/PARACFAC decomposition, which specifically represents for:
  • ⁇ Q represents the modulo Q inner product of the tensor and the matrix along the Q-th dimension; Represents a kernel tensor containing high-order singular values, with Represents the singular matrix corresponding to the four dimensions of v.
  • the spatial spectrum function is obtained in step (7)
  • the specific steps for searching for two-dimensional peaks are as follows: gradually increase with a° as the step size.
  • the value of the two-dimensional direction of arrival The search start point is (-90°,0°), and the end point is (90°,180°); each You can calculate a corresponding The spatial spectrum value of, thus can construct a corresponding to The space spectrum.
  • step (7) the construction of the tensor space spectrum described in step (7) can also be implemented using a noise subspace obtained based on high-order singular value decomposition, which is expressed as
  • ( ⁇ ) H represents the conjugate transpose operation.
  • ( ⁇ ) H represents the conjugate transpose operation.
  • the present invention has the following advantages:
  • the present invention uses a tensor to represent the actual received signal of the plane coprime, which is different from the traditional method of vectorizing the two-dimensional spatial information and averaging the snapshot information to obtain the second-order statistics.
  • the present invention takes the snapshots of each sample The signal is superimposed in the third dimension, and the second-order cross-correlation tensor containing the four-dimensional spatial information is used to estimate the spatial spectrum, and the multi-dimensional spatial structure information of the actual incident signal of the planar coprime array is retained;
  • the present invention constructs the subspace classification idea of the virtual domain signal through the tensor statistics analysis of the virtual domain equivalent signal, which provides a theoretical basis for the construction of the tensor space spectrum, thereby solving the problem of the signal mismatch of the planar coprime array. Problem, the construction of the virtual domain tensor space spectrum conforming to the Nyquist matching condition is realized;
  • the present invention uses tensor CANDECOMP/PARACFAC decomposition and high-order singular value decomposition to extract the multi-dimensional features of the fourth-order autocorrelation tensor of the virtual domain signal, thereby establishing the relationship between the virtual domain model and the signal and noise subspaces.
  • the connection of provides a basis for realizing high-precision, high-resolution tensor space spectrum.
  • Figure 1 is a block diagram of the overall flow of the present invention.
  • Fig. 2 is a schematic diagram of the structure of the planar coprime array in the present invention.
  • FIG. 3 is a schematic diagram of the structure of the augmented virtual domain area array derived by the present invention.
  • Figure 4 is a schematic diagram of the tensor space spectrum constructed by the present invention.
  • the present invention provides a high-resolution accurate two-dimensional wave based on a planar coprime array virtual domain tensor spatial spectrum search.
  • Direction of arrival estimation method Through the statistical analysis of the tensor signal received by the planar coprime array, the virtual domain equivalent signal with the spatial structure information of the virtual domain array is constructed; based on the multi-dimensional feature analysis method of the virtual domain signal tensor statistics, the virtual domain model and Zhang are established.
  • the implementation steps of the present invention are as follows:
  • Step 1 Construct a planar coprime array.
  • 4M x M y +N x N y -1 physical antenna elements are used to construct a planar coprime array, as shown in Figure 2:
  • Contains N x ⁇ N u antenna array elements the distance between the array elements in the x-axis direction and the y-axis direction is M x d and My y d, and its
  • Step 2 Modeling the received signal tensor of the planar coprime array.
  • K from Direction of the far-field narrow-band incoherent signal source, the sparse sub-array of the planar coprime array
  • the snapshot signals of each sample are superimposed in the third dimension to obtain a three-dimensional tensor signal (L is the number of sampled snapshots), expressed as:
  • s k [s k,1 ,s k,2 ,...,s k,L ] T is the multi-shot sampling signal waveform corresponding to the k-th incident signal source
  • [ ⁇ ] T represents the transposition operation
  • represents the outer product of the vector
  • Is a noise tensor independent of each signal source
  • the steering vector in the x-axis and y-axis directions corresponds to the direction of the incoming wave as
  • the signal source is expressed as:
  • the received signal can be another three-dimensional tensor Expressed as:
  • Compute subarray with Receive the cross-correlation statistics of the tensor signal x 1 and x 2 to obtain a second-order cross-correlation tensor with four-dimensional spatial information Expressed as:
  • x 1 (l) and x 2 (l) respectively represent the l-th slice of x 1 and x 2 in the third dimension (ie snapshot dimension) direction, and ( ⁇ ) * represents the conjugation operation;
  • Step 3 Derive the equivalent signal of the virtual domain based on the second-order cross-correlation tensor of the planar coprime array.
  • the second-order cross-correlation tensor of the tensor signal received by the two sub-arrays of the planar coprime array The ideal modeling (no noise scene) is:
  • the equivalent of the received signal will be the cross-correlation tensor
  • the first and third dimensions representing the spatial information in the x-axis direction are merged into one dimension, and the second and fourth dimensions representing the spatial information in the y-axis direction are merged into another dimension.
  • the dimensional merging of tensors can be realized through the modulus expansion operation of its PARAFAC decomposition. Specifically, the dimension set is defined with Cross-correlation tensor Model for PARAFAC decomposition Expand to get an augmented virtual domain area array Equivalent received signal Expressed as:
  • Step 4 Construct the equivalent received signal of the uniform area array in the virtual domain.
  • the structure of is shown in the dashed box in Figure 3, expressed as:
  • the steering vector in the x-axis and y-axis directions corresponds to the direction of the incoming wave as Signal source;
  • Step 5 Derive the fourth-order autocorrelation tensor of the smooth signal in the virtual domain.
  • a uniform area array of the virtual domain is obtained by the above steps Equivalent signal Virtual domain signal
  • the spatial structure information of the virtual domain array is saved in the, however, due to It can be regarded as a single snapshot of the virtual domain signal, and its statistics often have a rank deficit problem. Therefore, based on the idea of two-dimensional spatial smoothing, the virtual domain signal After processing, construct multiple equivalent snapshot virtual domain sub-array signals, and after summing and averaging these virtual domain sub-array signals, the fourth-order autocorrelation tensor is calculated.
  • the specific method is to uniform area array in the virtual domain In the x-axis and y-axis directions, take a sub-array with a size of Y 1 ⁇ Y 2 for every other element in the x-axis and y-axis directions. Divide into L 1 ⁇ L 2 uniform sub-arrays partially overlapping each other, and L 1 , L 2 , Y 1 , and Y 2 satisfy the following relationship:
  • Y 2 +L 2 -1 M y N y +M y +N y -1.
  • Step 6 Multi-dimensional feature extraction based on the fourth-order autocorrelation tensor of the virtual domain realizes signal and noise subspace classification.
  • the fourth-order autocorrelation tensor v is subjected to CANDECOMP/PARACFAC decomposition to extract multi-dimensional features. The obtained results are expressed as follows:
  • ⁇ Q represents the modulo Q inner product of the tensor and the matrix along the Q-th dimension; Represents a kernel tensor containing high-order singular values, with Represents the singular matrix corresponding to the four dimensions of v.
  • Step 7 High-resolution accurate two-dimensional direction of arrival estimation based on virtual domain tensor space spectrum search. Define the two-dimensional direction of arrival for spectral peak search Construct a uniform area array corresponding to the virtual domain Guide information Expressed as:
  • ⁇ ⁇ Q ⁇ > represents the modulo ⁇ Q ⁇ shrinking operation of two tensors along the Q-th dimension, requiring the same size of the Q-th dimension of the two tensors;
  • ⁇ F represents the Frobenius norm; with A vector is obtained by condensing modulo ⁇ 1,2 ⁇ along the first and second dimensions
  • the two-dimensional direction of arrival estimation result is obtained through the two-dimensional spectral peak search, the specific steps are: gradually increase the search step size a° The value of the two-dimensional direction of arrival
  • the search start point is (-90°,0°), and the end point is (90°,180°); each You can calculate a corresponding The spatial spectrum value of, thus can construct a corresponding to The space spectrum.
  • ( ⁇ ) H represents the conjugate transpose operation.
  • ( ⁇ ) H represents the conjugate transpose operation.
  • the spatial spectrum of the high-resolution accurate two-dimensional direction of arrival estimation method based on the virtual domain tensor spatial spectrum search of the planar coprime array proposed by the present invention is shown in FIG. 4. It can be seen that the method proposed in the present invention can effectively construct a two-dimensional spatial spectrum, in which there is a sharp peak corresponding to the two-dimensional direction of arrival of the incident signal source, and the x-axis and y-axis corresponding to the peak The value of is the elevation angle and azimuth angle of the incident source.
  • the present invention fully considers the multi-dimensional structure information of the planar coprime array signal, uses tensor signal modeling, constructs virtual domain equivalent signals with virtual domain area array spatial structure information, and analyzes its tensor statistics Features, constructs a subspace classification idea based on multi-dimensional feature extraction of virtual domain autocorrelation tensor, establishes the connection between the virtual domain model of the planar coprime array and the tensor space spectrum, and solves the signal mismatch problem of the planar coprime array
  • the present invention proposes a high-precision, high-resolution tensor spatial spectrum construction mechanism by using two tensor feature extraction methods, namely tensor decomposition and high-order singular value decomposition. Compared with the existing methods, in the spatial spectrum A breakthrough has been made in the resolution and accuracy of the two-dimensional direction of arrival estimation.

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

一种基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,主要解决现有方法中信号多维信息丢失和空间谱分辨度、精确度受限的问题,其实现步骤是:构建平面互质阵列;平面互质阵列接收信号张量建模;推导基于平面互质阵列二阶互相关张量的虚拟域等价信号;构造虚拟域均匀面阵的等价接收信号;推导虚拟域平滑信号的四阶自相关张量;基于虚拟域自相关张量的多维特征提取实现信号与噪声子空间分类;基于虚拟域张量空间谱搜索的高分辨精确二维波达方向估计。该方法基于平面互质阵列虚拟域张量统计量的多维特征提取,实现基于张量空间谱搜索的高分辨精确二维波达方向估计,可用于无源探测和目标定位。

Description

基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法 技术领域
本发明属于阵列信号处理技术领域,尤其涉及基于平面互质阵列张量信号的统计信号处理技术,具体是一种基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,可用于无源探测和空间定位。
背景技术
平面互质阵列作为一种具有系统化架构的二维稀疏阵列,具有大孔径、高分辨的特点,相较于传统均匀阵列,能够实现二维波达方向估计在估计精度、分辨度等综合性能上的突破;与此同时,通过构造二维虚拟域,能够在虚拟域均匀面阵上进行符合奈奎斯特匹配条件的信号处理,从而解决平面互质阵列的信号失配问题。基于虚拟域信号构造平面互质阵列空间谱,进而通过二维谱峰搜索,可以得到精确的二维波达方向估计。以此为思路,传统的方法通常将具有二维空间结构信息的入射信号用矢量进行表示,并以时间平均的方式计算多采样信号的二阶统计量,进而通过矢量化推导虚拟域二阶等价信号。然而,以矢量方式表示的平面互质阵列接收信号及其虚拟域等价信号不仅丢失了原始信号的多维空间结构信息,且随着数据量增大,容易造成维度灾难,因此以此为基础构造空间谱并得到二维波达方向估计在精确度、分辨度等性能上仍存在缺陷。
为了解决上述问题,基于张量空间谱搜索的平面互质阵列二维波达方向估计方法开始受到关注。张量作为一种高维的数据结构,可以保存信号的原始多维信息;同时,高阶奇异值分解、张量分解等多维代数理论也为张量信号的多维特征提取提供了丰富的分析工具。因此,张量信号模型能够充分利用平面互质阵列入射信号的多维空间结构信息。然而,现有方法仍然是基于实际接收张量信号进行处理,并没有利用平面互质阵列的二维虚拟域进行张量空间谱构造, 没有解决平面互质阵列信号失配的问题,导致精确度受损;且生成的谱峰分辨度低,易产生相互混叠。因此,现有方法在精确度和分辨度性能上仍存在较大的提升空间。
发明内容
本发明的目的在于针对上述方法中存在的信号多维空间结构信息丢失和空间谱分辨度、精度性能受限问题,提出一种基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,为建立平面互质阵列张量信号统计量和虚拟域空间谱联系,搭建虚拟域张量空间谱搜索架构,实现高分辨、高精度的二维波达方向估计提供了可行的思路和有效的解决方案。
本发明的目的是通过以下技术方案实现的:一种基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,包含以下步骤:
(1)接收端使用4M xM y+N xN y-1个物理天线阵元,按照平面互质阵列的结构进行架构;其中,M x、N x以及M y、N y分别为一对互质整数,且M x<N x,M y<N y;该平面互质阵列可分解为两个稀疏均匀子阵列
Figure PCTCN2020088567-appb-000001
Figure PCTCN2020088567-appb-000002
(2)假设有K个来自
Figure PCTCN2020088567-appb-000003
方向的远场窄带非相干信号源,将平面互质阵列稀疏子阵列
Figure PCTCN2020088567-appb-000004
的接收信号用一个三维张量信号
Figure PCTCN2020088567-appb-000005
(L为采样快拍数)表示为:
Figure PCTCN2020088567-appb-000006
其中,s k=[s k,1,s k,2,…,s k,L] T为对应第k个入射信源的多快拍采样信号波形,[·] T表示转置操作,ο表示矢量外积,
Figure PCTCN2020088567-appb-000007
为与各信号源相互独立的噪声张量,
Figure PCTCN2020088567-appb-000008
Figure PCTCN2020088567-appb-000009
分别为
Figure PCTCN2020088567-appb-000010
在x轴和y轴方向上的导引矢量,对应于来波方向为
Figure PCTCN2020088567-appb-000011
的信号源,表示为:
Figure PCTCN2020088567-appb-000012
Figure PCTCN2020088567-appb-000013
其中,
Figure PCTCN2020088567-appb-000014
Figure PCTCN2020088567-appb-000015
分别表示稀疏子阵列
Figure PCTCN2020088567-appb-000016
在 x轴和y轴方向上第i 1和i 2个物理天线阵元的实际位置,且
Figure PCTCN2020088567-appb-000017
Figure PCTCN2020088567-appb-000018
稀疏子阵列
Figure PCTCN2020088567-appb-000019
的接收信号可用另一个三维张量
Figure PCTCN2020088567-appb-000020
表示为:
Figure PCTCN2020088567-appb-000021
其中,
Figure PCTCN2020088567-appb-000022
为与各信号源相互独立的噪声张量,
Figure PCTCN2020088567-appb-000023
Figure PCTCN2020088567-appb-000024
分别为稀疏子阵列
Figure PCTCN2020088567-appb-000025
在x轴和y轴方向上的导引矢量,对应于来波方向为
Figure PCTCN2020088567-appb-000026
的信号源,表示为:
Figure PCTCN2020088567-appb-000027
Figure PCTCN2020088567-appb-000028
其中,
Figure PCTCN2020088567-appb-000029
Figure PCTCN2020088567-appb-000030
分别表示稀疏子阵列
Figure PCTCN2020088567-appb-000031
在x轴和y轴方向上第i 3和i 4个物理天线阵元的实际位置,且
Figure PCTCN2020088567-appb-000032
计算子阵列
Figure PCTCN2020088567-appb-000033
Figure PCTCN2020088567-appb-000034
的接收张量信号x 1和x 2的二阶互相关张量
Figure PCTCN2020088567-appb-000035
Figure PCTCN2020088567-appb-000036
表示为:
Figure PCTCN2020088567-appb-000037
这里,x 1(l)和x 2(l)分别表示x 1和x 2在第三维度(即快拍维度)方向上的第l个切片,(·) *表示共轭操作;
(3)由互相关张量
Figure PCTCN2020088567-appb-000038
得到一个增广的非均匀虚拟域面阵
Figure PCTCN2020088567-appb-000039
其中各虚拟阵元的位置表示为:
Figure PCTCN2020088567-appb-000040
其中,单位间隔d取为入射窄带信号波长λ的一半,即d=λ/2。定义维度集合
Figure PCTCN2020088567-appb-000041
Figure PCTCN2020088567-appb-000042
则通过对互相关张量
Figure PCTCN2020088567-appb-000043
的理想值
Figure PCTCN2020088567-appb-000044
(无噪声场景)进行PARAFAC分解的模
Figure PCTCN2020088567-appb-000045
展开,可获得增广虚拟域面阵
Figure PCTCN2020088567-appb-000046
的等价接收信号
Figure PCTCN2020088567-appb-000047
的理想表示为:
Figure PCTCN2020088567-appb-000048
其中,
Figure PCTCN2020088567-appb-000049
Figure PCTCN2020088567-appb-000050
Figure PCTCN2020088567-appb-000051
是增广虚拟域面阵
Figure PCTCN2020088567-appb-000052
在x轴和y轴方向上的导引矢量,对应于来波方向为
Figure PCTCN2020088567-appb-000053
的信号源;
Figure PCTCN2020088567-appb-000054
表示第k个入射信号源的功率;这里,
Figure PCTCN2020088567-appb-000055
表示克罗内克积;张量下标表示张量的PARAFAC分解的模展开操作;
(4)
Figure PCTCN2020088567-appb-000056
中包含一个x轴分布为(-N x+1)d到(M xN x+M x-1)d、y轴分布为(-N y+1)d到(M yN y+M y-1)d的虚拟域均匀面阵
Figure PCTCN2020088567-appb-000057
中共有D x×D y个虚拟阵元,其中D x=M xN x+M x+N x-1,D y=M yN y+M y+N y-1,
Figure PCTCN2020088567-appb-000058
表示为:
Figure PCTCN2020088567-appb-000059
通过选取虚拟域等价接收信号V中与
Figure PCTCN2020088567-appb-000060
各虚拟阵元位置相对应的元素,获取虚拟域均匀面阵
Figure PCTCN2020088567-appb-000061
的等价接收信号
Figure PCTCN2020088567-appb-000062
将其表示为:
Figure PCTCN2020088567-appb-000063
其中,
Figure PCTCN2020088567-appb-000064
Figure PCTCN2020088567-appb-000065
Figure PCTCN2020088567-appb-000066
Figure PCTCN2020088567-appb-000067
为虚拟域均匀面阵
Figure PCTCN2020088567-appb-000068
在x轴和y轴方向上的导引矢量,对应于来波方向为
Figure PCTCN2020088567-appb-000069
的信号源;
(5)在虚拟域均匀面阵
Figure PCTCN2020088567-appb-000070
中,分别沿x轴和y轴方向每隔一个阵元取一个大小为Y 1×Y 2的子阵列,则可以将虚拟域均匀面阵
Figure PCTCN2020088567-appb-000071
分割成L 1×L 2个互相部分重叠的均匀子阵列。将上述子阵列表示为
Figure PCTCN2020088567-appb-000072
g 1=1,2,…,L 1,g 2=1,2,…,L 2,根据子阵列
Figure PCTCN2020088567-appb-000073
对应虚拟域信号
Figure PCTCN2020088567-appb-000074
中相应位置元素,得到虚拟域子阵列
Figure PCTCN2020088567-appb-000075
的等价信号
Figure PCTCN2020088567-appb-000076
Figure PCTCN2020088567-appb-000077
其中,
Figure PCTCN2020088567-appb-000078
Figure PCTCN2020088567-appb-000079
Figure PCTCN2020088567-appb-000080
Figure PCTCN2020088567-appb-000081
为对应于
Figure PCTCN2020088567-appb-000082
方向的虚拟域子阵列
Figure PCTCN2020088567-appb-000083
在x轴和y轴上的导引矢量。经过上述操作,一共得到L 1×L 2个维度均为Y 1×Y 2的虚拟域子阵信号
Figure PCTCN2020088567-appb-000084
对这L 1×L 2个虚拟域子阵信号
Figure PCTCN2020088567-appb-000085
求平均值,得到一个虚拟域平滑信号
Figure PCTCN2020088567-appb-000086
Figure PCTCN2020088567-appb-000087
对该虚拟域信号
Figure PCTCN2020088567-appb-000088
求得其四阶自相关张量
Figure PCTCN2020088567-appb-000089
表示为:
Figure PCTCN2020088567-appb-000090
(6)对四阶自相关张量v进行CANDECOMP/PARACFAC分解以提取多维特征,得到结果表示如下:
Figure PCTCN2020088567-appb-000091
其中,
Figure PCTCN2020088567-appb-000092
Figure PCTCN2020088567-appb-000093
为CANDECOMP/PARACFAC分解得到的两组正交因子矢量,分别表示x轴和y轴方向上的空间信息,
Figure PCTCN2020088567-appb-000094
Figure PCTCN2020088567-appb-000095
Figure PCTCN2020088567-appb-000096
为因子矩阵;取
Figure PCTCN2020088567-appb-000097
张成的空间,记作
Figure PCTCN2020088567-appb-000098
Figure PCTCN2020088567-appb-000099
作为信号子空间,用一个张量
Figure PCTCN2020088567-appb-000100
表示该信号子空间,其中
Figure PCTCN2020088567-appb-000101
表示
Figure PCTCN2020088567-appb-000102
沿着第三维度的第k个切片,表示为:
Figure PCTCN2020088567-appb-000103
为了得到噪声子空间,对因子矩阵C x和C y求其正交补;C x的正交补记为
Figure PCTCN2020088567-appb-000104
C y的正交补记为
Figure PCTCN2020088567-appb-000105
Figure PCTCN2020088567-appb-000106
其中min(·)表示取最小值操作;则取
Figure PCTCN2020088567-appb-000107
Figure PCTCN2020088567-appb-000108
作为噪声子空间,用张量
Figure PCTCN2020088567-appb-000109
表示该噪声子空间,
Figure PCTCN2020088567-appb-000110
表示
Figure PCTCN2020088567-appb-000111
沿着第三维度的第h个切片,表示为:
Figure PCTCN2020088567-appb-000112
(7)定义用于谱峰搜索的二维波达方向
Figure PCTCN2020088567-appb-000113
Figure PCTCN2020088567-appb-000114
构造对应虚拟域均匀面阵
Figure PCTCN2020088567-appb-000115
的导引信息
Figure PCTCN2020088567-appb-000116
表示为:
Figure PCTCN2020088567-appb-000117
使用基于CANDECOMP/PARACFAC分解得到的噪声子空间构造张量空间谱函数
Figure PCTCN2020088567-appb-000118
表示如下:
Figure PCTCN2020088567-appb-000119
其中,<× {Q}>表示两个张量沿着第Q维度的模{Q}缩并操作,要求两个张量的第Q维度的大小相同;‖·‖ F表示Frobenius范数;
Figure PCTCN2020088567-appb-000120
Figure PCTCN2020088567-appb-000121
沿着第1,2维度的模{1,2}缩并操作得到一个矢量
Figure PCTCN2020088567-appb-000122
Figure PCTCN2020088567-appb-000123
得到空间谱函数
Figure PCTCN2020088567-appb-000124
之后,可以构造出对应二维搜索波达方向的空间谱,随后通过搜索谱峰所在位置对应的二维波达方向,即为入射信源的二维波达方向估计。
进一步地,步骤(1)所述的平面互质阵列结构可具体描述为:在平面坐标系xoy上构造一对稀疏均匀平面子阵列
Figure PCTCN2020088567-appb-000125
Figure PCTCN2020088567-appb-000126
其中
Figure PCTCN2020088567-appb-000127
包含2M x×2M y个天线阵元,在x轴方向上和y轴方向上的阵元间距分别为N xd和N yd,其在xoy上的位置坐标为{(N xdm x,N ydm y),m x=0,1,...,2M x-1,m y=0,1,...,2M y-1};
Figure PCTCN2020088567-appb-000128
包含N x×N y个天线阵元,在x轴方向上和y轴方向上的阵元间距分别为M xd和M yd,其在xoy上的位置坐标为{(M xdn x,M ydn y),n x=0,1,...,N x-1,n y=0,1,...,N y-1};这里,M x、N x以及M y、N y分别为一对互质整数,且M x<N x,M y<N y;将
Figure PCTCN2020088567-appb-000129
Figure PCTCN2020088567-appb-000130
按照(0,0)坐标处阵元重叠的方式进行子阵列组合,获得实际包含4M xM y+N xN y-1个物理天线阵元的互质面阵。
进一步地,步骤(3)所述的互相关张量
Figure PCTCN2020088567-appb-000131
可理想(无噪声场景)建模为:
Figure PCTCN2020088567-appb-000132
此时,
Figure PCTCN2020088567-appb-000133
Figure PCTCN2020088567-appb-000134
等价于沿着x轴的一个增广虚拟域,
Figure PCTCN2020088567-appb-000135
等价于沿着y轴的一个增广虚拟域,从而可以得到非均匀虚拟域面阵
Figure PCTCN2020088567-appb-000136
进一步地,步骤(5)所述的虚拟域均匀面阵
Figure PCTCN2020088567-appb-000137
的等价信号
Figure PCTCN2020088567-appb-000138
中保存了虚拟域面阵的空间结构信息,然而,由于
Figure PCTCN2020088567-appb-000139
可以视作一个单快拍的虚拟域信号,其统计量往往存在秩亏问题。因此,基于二维空间平滑的思想对虚拟域信号
Figure PCTCN2020088567-appb-000140
进行处理,构造多个等效快拍虚拟域子阵信号,对这些虚拟域子阵信号进行求和平均后,求其四阶自相关张量。子阵列
Figure PCTCN2020088567-appb-000141
中阵元的位置表示为:
Figure PCTCN2020088567-appb-000142
通过子阵列
Figure PCTCN2020088567-appb-000143
对应选取虚拟域信号
Figure PCTCN2020088567-appb-000144
中相应位置元素,得到虚拟域子阵列
Figure PCTCN2020088567-appb-000145
的等价信号
Figure PCTCN2020088567-appb-000146
进一步地,步骤(6)所述的提取虚拟域四阶自相关张量的多维特征以实现信号与噪声子空间分类,除了通过CANDECOMP/PARACFAC分解,还可以通过高阶奇异值分解实现,具体表示为:
Figure PCTCN2020088567-appb-000147
其中,× Q表示张量与矩阵沿着第Q维度的模Q内积;
Figure PCTCN2020088567-appb-000148
表示包含高阶奇异值的核张量,
Figure PCTCN2020088567-appb-000149
Figure PCTCN2020088567-appb-000150
表示对应v四个维度的奇异矩阵。将D x的前K列和后Y 1-K列分开为信号子空间
Figure PCTCN2020088567-appb-000151
和噪声子空间
Figure PCTCN2020088567-appb-000152
类似地,将D y的前K列和后Y 2-K列分开为信号子空间
Figure PCTCN2020088567-appb-000153
和噪声子空间
Figure PCTCN2020088567-appb-000154
进一步地,步骤(7)中得到空间谱函数
Figure PCTCN2020088567-appb-000155
之后进行二维谱峰搜索的具体步骤为:以a°为步长逐渐分别增加
Figure PCTCN2020088567-appb-000156
的值,二维波达方向
Figure PCTCN2020088567-appb-000157
的搜索起点为(-90°,0°),终点为(90°,180°);每个
Figure PCTCN2020088567-appb-000158
可以对应计算出一个
Figure PCTCN2020088567-appb-000159
的空间谱值,从而可以构造出一个对应于
Figure PCTCN2020088567-appb-000160
的空间谱。空间谱中存在K个峰值,该K个峰值所对应的
Figure PCTCN2020088567-appb-000161
的值,即为信源的二维波达方向估计。
进一步地,步骤(7)中所述的张量空间谱构造还可以使用基于高阶奇异值分解得到的噪声子空间实现,表示为
Figure PCTCN2020088567-appb-000162
Figure PCTCN2020088567-appb-000163
其中,(·) H表示共轭转置操作。同样地,得到空间谱函数
Figure PCTCN2020088567-appb-000164
之后,按照上述的二维谱峰搜索过程,即可得到信源的二维波达方向估计。
本发明与现有技术相比具有以下优点:
(1)本发明通过张量表示平面互质实际接收信号,不同于传统方法将二维空间信息进行矢量化表征,并将快拍信息进行平均得到二阶统计量,本发明将各采样快拍信号在第三维度上叠加,并利用包含四维空间信息的二阶互相关张量进行空间谱估计,保留了平面互质阵列实际入射信号的多维空间结构信息;
(2)本发明通过虚拟域等价信号的张量统计量分析构建虚拟域信号的子空间分类思路,为张量空间谱的构造提供了理论基础,从而解决了平面互质阵列信号失配的问题,实现了符合奈奎斯特匹配条件的虚拟域张量空间谱构造;
(3)本发明采用张量CANDECOMP/PARACFAC分解和高阶奇异值分解的方式对虚拟域信号的四阶自相关张量进行多维特征提取,从而建立起虚拟域模型与信号、噪声子空间之间的联系,为实现高精度、高分辨度的张量空间谱提供了基础。
附图说明
图1是本发明的总体流程框图。
图2是本发明中平面互质阵列的结构示意图。
图3是本发明所推导增广虚拟域面阵结构示意图。
图4是本发明所构造张量空间谱示意图。
具体实施方式
以下参照附图,对本发明的技术方案作进一步的详细说明。
为了解决现有方法存在的信号多维空间结构信息丢失和空间谱分辨度、精度性能受限问题,本发明提供了一种基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法。通过对平面互质阵列接收张量信号进行统计分析,构造具有虚拟域面阵空间结构信息的虚拟域等价信号;基于虚拟域 信号张量统计量的多维特征分析手段,建立虚拟域模型与张量空间谱之间的联系,从而在虚拟域上实现符合奈奎斯特匹配条件的基于张量空间谱搜索的高分辨精确二维波达方向估计方法。参照图1,本发明的实现步骤如下:
步骤1:构建平面互质阵列。在接收端使用4M xM y+N xN y-1个物理天线阵元构建平面互质阵列,如图2所示:在平面坐标系xoy上构造一对稀疏均匀平面子阵列
Figure PCTCN2020088567-appb-000165
Figure PCTCN2020088567-appb-000166
其中
Figure PCTCN2020088567-appb-000167
包含2M x×2M y个天线阵元,在x轴方向上和y轴方向上的阵元间距分别为N xd和N yd,其在xoy上的位置坐标为{(N xdm x,N ydm y),m x=0,1,...,2M x-1,m y=0,1,...,2M y-1};
Figure PCTCN2020088567-appb-000168
包含N x×N u个天线阵元,在x轴方向上和y轴方向上的阵元间距分别为M xd和M yd,其在xoy上的位置坐标为{(M xdn x,M ydn y),n x=0,1,...,N x-1,n y=0,1,...,N y-1};这里,M x、N x以及M y、N y分别为一对互质整数,且M x<N x,M y<N y;单位间隔d取为入射窄带信号波长λ的一半,即d=λ/2;将
Figure PCTCN2020088567-appb-000169
Figure PCTCN2020088567-appb-000170
按照(0,0)坐标处阵元重叠的方式进行子阵列组合,获得实际包含4M xM y+N xN y-1个物理天线阵元的平面互质阵列;
步骤2:平面互质阵列接收信号张量建模。假设有K个来自
Figure PCTCN2020088567-appb-000171
方向的远场窄带非相干信号源,将平面互质阵列稀疏子阵列
Figure PCTCN2020088567-appb-000172
的各采样快拍信号在第三维度进行叠加,得到一个三维张量信号
Figure PCTCN2020088567-appb-000173
(L为采样快拍数),表示为:
Figure PCTCN2020088567-appb-000174
其中,s k=[s k,1,s k,2,…,s k,L] T为对应第k个入射信源的多快拍采样信号波形,[·] T表示转置操作,ο表示矢量外积,
Figure PCTCN2020088567-appb-000175
为与各信号源相互独立的噪声张量,
Figure PCTCN2020088567-appb-000176
Figure PCTCN2020088567-appb-000177
分别为
Figure PCTCN2020088567-appb-000178
在x轴和y轴方向上的导引矢量,对应于来波方向为
Figure PCTCN2020088567-appb-000179
的信号源,表示为:
Figure PCTCN2020088567-appb-000180
Figure PCTCN2020088567-appb-000181
其中,
Figure PCTCN2020088567-appb-000182
Figure PCTCN2020088567-appb-000183
分别表示稀疏子阵列
Figure PCTCN2020088567-appb-000184
在x轴和y轴方向上第i 1和i 2个物理天线阵元的实际位置,且
Figure PCTCN2020088567-appb-000185
Figure PCTCN2020088567-appb-000186
类似地,稀疏子阵列
Figure PCTCN2020088567-appb-000187
的接收信号可用另一个三维张量
Figure PCTCN2020088567-appb-000188
表示为:
Figure PCTCN2020088567-appb-000189
其中,
Figure PCTCN2020088567-appb-000190
为与各信号源相互独立的噪声张量,
Figure PCTCN2020088567-appb-000191
Figure PCTCN2020088567-appb-000192
分别为稀疏子阵列
Figure PCTCN2020088567-appb-000193
在x轴和y轴方向上的导引矢量,对应于来波方向为
Figure PCTCN2020088567-appb-000194
的信号源,表示为:
Figure PCTCN2020088567-appb-000195
Figure PCTCN2020088567-appb-000196
其中,
Figure PCTCN2020088567-appb-000197
Figure PCTCN2020088567-appb-000198
分别表示稀疏子阵列
Figure PCTCN2020088567-appb-000199
在x轴和y轴方向上第i 3和i 4个物理天线阵元的实际位置,且
Figure PCTCN2020088567-appb-000200
计算子阵列
Figure PCTCN2020088567-appb-000201
Figure PCTCN2020088567-appb-000202
的接收张量信号x 1和x 2的互相关统计量,得到一个具有四维空间信息的二阶互相关张量
Figure PCTCN2020088567-appb-000203
表示为:
Figure PCTCN2020088567-appb-000204
这里,x 1(l)和x 2(l)分别表示x 1和x 2在第三维度(即快拍维度)方向上的第l个切片,(·) *表示共轭操作;
步骤3:推导基于平面互质阵列二阶互相关张量的虚拟域等价信号。平面互质阵列两个子阵列接收张量信号的二阶互相关张量
Figure PCTCN2020088567-appb-000205
可理想建模(无噪声场景)为:
Figure PCTCN2020088567-appb-000206
其中,
Figure PCTCN2020088567-appb-000207
表示第k个入射信号源的功率;此时,
Figure PCTCN2020088567-appb-000208
Figure PCTCN2020088567-appb-000209
等 价于沿着x轴的一个增广虚拟域,
Figure PCTCN2020088567-appb-000210
等价于沿着y轴的一个增广虚拟域,从而可以得到一个增广的非均匀虚拟域面阵
Figure PCTCN2020088567-appb-000211
如图3所示,其中各虚拟阵元的位置表示为:
Figure PCTCN2020088567-appb-000212
为了得到对应于增广虚拟域面阵
Figure PCTCN2020088567-appb-000213
的等价接收信号,将互相关张量
Figure PCTCN2020088567-appb-000214
中表征x轴方向空间信息的第1、3维度合并成一个维度,将表征y轴方向空间信息的第2、4维度合并成另一个维度。张量的维度合并可通过其PARAFAC分解的模展开操作实现,具体地,定义维度集合
Figure PCTCN2020088567-appb-000215
Figure PCTCN2020088567-appb-000216
则通过对互相关张量
Figure PCTCN2020088567-appb-000217
进行PARAFAC分解的模
Figure PCTCN2020088567-appb-000218
展开,可获得增广虚拟域面阵
Figure PCTCN2020088567-appb-000219
的等价接收信号
Figure PCTCN2020088567-appb-000220
表示为:
Figure PCTCN2020088567-appb-000221
其中,
Figure PCTCN2020088567-appb-000222
Figure PCTCN2020088567-appb-000223
Figure PCTCN2020088567-appb-000224
是增广虚拟域面阵
Figure PCTCN2020088567-appb-000225
在x轴和y轴方向上的导引矢量,对应于来波方向为
Figure PCTCN2020088567-appb-000226
的信号源;这里,
Figure PCTCN2020088567-appb-000227
表示克罗内克积;
步骤4:构造虚拟域均匀面阵的等价接收信号。增广虚拟域面阵
Figure PCTCN2020088567-appb-000228
中包含一个x轴分布为(-N x+1)d到(M xN x+M x-1)d、y轴分布为(-N y+1)d到(M yN y+M y-1)d的虚拟域均匀面阵
Figure PCTCN2020088567-appb-000229
中共有D x×D y个虚拟阵元,其中D x=M xN x+M x+N x-1,D y=M yN y+M y+N y-1;虚拟域均匀面阵
Figure PCTCN2020088567-appb-000230
的结构如图3中虚线框内所示,表示为:
Figure PCTCN2020088567-appb-000231
通过选取非均匀虚拟域面阵
Figure PCTCN2020088567-appb-000232
的等价接收信号V中与
Figure PCTCN2020088567-appb-000233
各虚拟阵元位置相对应的元素,可以获取虚拟域均匀面阵
Figure PCTCN2020088567-appb-000234
的等价接收信号
Figure PCTCN2020088567-appb-000235
Figure PCTCN2020088567-appb-000236
其中,
Figure PCTCN2020088567-appb-000237
Figure PCTCN2020088567-appb-000238
Figure PCTCN2020088567-appb-000239
为虚拟域均匀面阵
Figure PCTCN2020088567-appb-000240
在x轴和y轴方向上的导引矢量,对应于来波方向为
Figure PCTCN2020088567-appb-000241
的信号源;
步骤5:推导虚拟域平滑信号的四阶自相关张量。由上述步骤得到虚拟域均匀面阵
Figure PCTCN2020088567-appb-000242
的等价信号
Figure PCTCN2020088567-appb-000243
虚拟域信号
Figure PCTCN2020088567-appb-000244
中保存了虚拟域面阵的空间结构信息,然而,由于
Figure PCTCN2020088567-appb-000245
可以视作一个单快拍的虚拟域信号,其统计量往往存在秩亏问题。因此,基于二维空间平滑的思想对虚拟域信号
Figure PCTCN2020088567-appb-000246
进行处理,构造多个等效快拍虚拟域子阵信号,对这些虚拟域子阵信号进行求和平均后,求其四阶自相关张量。具体做法为,在虚拟域均匀面阵
Figure PCTCN2020088567-appb-000247
中,分别沿x轴和y轴方向每隔一个阵元取一个大小为Y 1×Y 2的子阵列,则可以将虚拟域均匀面阵
Figure PCTCN2020088567-appb-000248
分割成L 1×L 2个互相部分重叠的均匀子阵列,L 1、L 2、Y 1、Y 2之间满足以下关系:
Y 1+L 1-1=M xN x+M x+N x-1,
Y 2+L 2-1=M yN y+M y+N y-1.
将上述子阵列表示为
Figure PCTCN2020088567-appb-000249
g 1=1,2,…,L 1,g 2=1,2,…,L 2,则
Figure PCTCN2020088567-appb-000250
中阵元的位置表示为:
Figure PCTCN2020088567-appb-000251
根据子阵列
Figure PCTCN2020088567-appb-000252
对应虚拟域信号
Figure PCTCN2020088567-appb-000253
中相应位置元素,得到虚拟域子阵列
Figure PCTCN2020088567-appb-000254
的等价信号
Figure PCTCN2020088567-appb-000255
Figure PCTCN2020088567-appb-000256
其中,
Figure PCTCN2020088567-appb-000257
Figure PCTCN2020088567-appb-000258
Figure PCTCN2020088567-appb-000259
Figure PCTCN2020088567-appb-000260
为对应于
Figure PCTCN2020088567-appb-000261
方向的虚 拟域子阵列
Figure PCTCN2020088567-appb-000262
在x轴和y轴上的导引矢量。经过上述操作,一共得到L 1×L 2个维度均为Y 1×Y 2的虚拟域子阵信号
Figure PCTCN2020088567-appb-000263
对这L 1×L 2个虚拟域子阵信号
Figure PCTCN2020088567-appb-000264
求平均值,得到一个虚拟域平滑信号
Figure PCTCN2020088567-appb-000265
Figure PCTCN2020088567-appb-000266
对该虚拟域信号
Figure PCTCN2020088567-appb-000267
求得其四阶自相关张量
Figure PCTCN2020088567-appb-000268
表示为:
Figure PCTCN2020088567-appb-000269
步骤6:基于虚拟域四阶自相关张量的多维特征提取实现信号与噪声子空间分类。为了构建基于子空间分类思想的张量空间谱,对四阶自相关张量v进行CANDECOMP/PARACFAC分解以提取多维特征,得到结果表示如下:
Figure PCTCN2020088567-appb-000270
其中,
Figure PCTCN2020088567-appb-000271
Figure PCTCN2020088567-appb-000272
为CANDECOMP/PARACFAC分解得到的两组正交因子矢量,分别表示x轴和y轴方向的空间信息,
Figure PCTCN2020088567-appb-000273
Figure PCTCN2020088567-appb-000274
为因子矩阵;取
Figure PCTCN2020088567-appb-000275
Figure PCTCN2020088567-appb-000276
张成的空间,记作
Figure PCTCN2020088567-appb-000277
Figure PCTCN2020088567-appb-000278
作为信号子空间,用一个张量
Figure PCTCN2020088567-appb-000279
表示该信号子空间,其中
Figure PCTCN2020088567-appb-000280
表示
Figure PCTCN2020088567-appb-000281
沿着第三维度的第k个切片,表示为:
Figure PCTCN2020088567-appb-000282
为了得到噪声子空间,需要对因子矩阵C x和C y求其正交补;C x的正交补记为
Figure PCTCN2020088567-appb-000283
C y的正交补记为
Figure PCTCN2020088567-appb-000284
Figure PCTCN2020088567-appb-000285
其中min(·)表示取最小值操作;则取
Figure PCTCN2020088567-appb-000286
Figure PCTCN2020088567-appb-000287
作为噪声子空间,用张量
Figure PCTCN2020088567-appb-000288
表示该噪声子空间,
Figure PCTCN2020088567-appb-000289
表示
Figure PCTCN2020088567-appb-000290
沿着第三维度的第h个切片,表示为:
Figure PCTCN2020088567-appb-000291
除了使用张量分解提取虚拟域自相关张量的多维特征,还可以通过高阶奇 异值分解,具体表示为:
Figure PCTCN2020088567-appb-000292
其中,× Q表示张量与矩阵沿着第Q维度的模Q内积;
Figure PCTCN2020088567-appb-000293
表示包含高阶奇异值的核张量,
Figure PCTCN2020088567-appb-000294
Figure PCTCN2020088567-appb-000295
表示对应v四个维度的奇异矩阵。将D x的前K列和后Y 1-K列分开为信号子空间
Figure PCTCN2020088567-appb-000296
和噪声子空间
Figure PCTCN2020088567-appb-000297
类似地,将D y的前K列和后Y 2-K列分开为信号子空间
Figure PCTCN2020088567-appb-000298
和噪声子空间
Figure PCTCN2020088567-appb-000299
步骤7:基于虚拟域张量空间谱搜索的高分辨精确二维波达方向估计。定义用于谱峰搜索的二维波达方向
Figure PCTCN2020088567-appb-000300
构造对应虚拟域均匀面阵
Figure PCTCN2020088567-appb-000301
的导引信息
Figure PCTCN2020088567-appb-000302
表示为:
Figure PCTCN2020088567-appb-000303
使用基于CANDECOMP/PARACFAC分解得到的噪声子空间构造张量空间谱函数
Figure PCTCN2020088567-appb-000304
表示如下:
Figure PCTCN2020088567-appb-000305
其中,<× {Q}>表示两个张量沿着第Q维度的模{Q}缩并操作,要求两个张量的第Q维度的大小相同;‖·‖ F表示Frobenius范数;
Figure PCTCN2020088567-appb-000306
Figure PCTCN2020088567-appb-000307
沿着第1,2维度的模{1,2}缩并操作得到一个矢量
Figure PCTCN2020088567-appb-000308
Figure PCTCN2020088567-appb-000309
得到空间谱函数
Figure PCTCN2020088567-appb-000310
之后,通过二维谱峰搜索得到二维波达方向估计结果,具体步骤为:以搜索步长a°逐渐分别增加
Figure PCTCN2020088567-appb-000311
的值,二维波达方向
Figure PCTCN2020088567-appb-000312
的搜索起点为(-90°,0°),终点为(90°,180°);每个
Figure PCTCN2020088567-appb-000313
可以对应计算出一个
Figure PCTCN2020088567-appb-000314
的空间谱值,从而可以构造出一个对应于
Figure PCTCN2020088567-appb-000315
Figure PCTCN2020088567-appb-000316
的空间谱。空间谱中存在K个峰值,该K个峰值位置所对应的
Figure PCTCN2020088567-appb-000317
的值,即为信源的二维波达方向估计。
基于高阶奇异值分解得到的噪声子空间构造张量空间谱函数
Figure PCTCN2020088567-appb-000318
表示为:
Figure PCTCN2020088567-appb-000319
其中,(·) H表示共轭转置操作。同样地,得到空间谱函数
Figure PCTCN2020088567-appb-000320
之后,按照上述的二维谱峰搜索方式,即可得到信源的二维波达方向估计。
下面结合仿真实例对本发明的效果做进一步的描述。
仿真实例:采用互质阵列接收入射信号,其参数选取为M x=2,M y=2,N x=3,N y=3,即架构的互质阵列共包含4M xM y+N xN y-1=24个物理阵元。假定入射窄带信号个数为1,且入射方向方位角和仰角分别为[45°,50°];采用L=500个采样快拍及10dB的输入信噪比进行仿真实验。
本发明所提出的基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法的空间谱如图4所示。可以看出,本发明所提方法能够有效地构造出二维空间谱,其中对应入射信源的二维波达方向位置存在一个精尖的谱峰,该谱峰所对应的x轴和y轴的值即为入射信源的俯仰角和方位角。
综上所述,本发明充分考虑了平面互质阵列信号的多维结构信息,利用张量信号建模,构造具有虚拟域面阵空间结构信息的虚拟域等价信号,并通过分析其张量统计特性,构建起基于虚拟域自相关张量多维特征提取的子空间分类思路,建立起平面互质阵列虚拟域模型与张量空间谱之间的联系,解决了平面互质阵列的信号失配问题;同时,本发明通过利用张量分解和高阶奇异值分解两种张量特征提取手段,提出了高精度、高分辨度张量空间谱的构造机理,相较于现有方法,在空间谱的分辨度和二维波达方向估计精度性能上取得了突破。
以上所述仅是本发明的优选实施方式,虽然本发明已以较佳实施例披露如上,然而并非用以限定本发明。任何熟悉本领域的技术人员,在不脱离本发明技术方案范围情况下,都可利用上述揭示的方法和技术内容对本发明技术方案做出许多可能的变动和修饰,或修改为等同变化的等效实施例。因此,凡是未脱离本发明技术方案的内容,依据本发明的技术实质对以上实施例所做的任何的简单修改、等同变化及修饰,均仍属于本发明技术方案保护的范围内。

Claims (7)

  1. 一种基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,其特征在于,包含以下步骤:
    (1)接收端使用4M xM y+N xN y-1个物理天线阵元,按照平面互质阵列的结构进行架构;其中,M x、N x以及M y、N y分别为一对互质整数,且M x<N x,M y<N y;该平面互质阵列可分解为两个稀疏均匀子阵列
    Figure PCTCN2020088567-appb-100001
    Figure PCTCN2020088567-appb-100002
    (2)假设有K个来自
    Figure PCTCN2020088567-appb-100003
    方向的远场窄带非相干信号源,将平面互质阵列稀疏子阵列
    Figure PCTCN2020088567-appb-100004
    的接收信号用一个三维张量信号
    Figure PCTCN2020088567-appb-100005
    (L为采样快拍数)表示为:
    Figure PCTCN2020088567-appb-100006
    其中,s k=[s k,1,s k,2,…,s k,L] T为对应第k个入射信源的多快拍采样信号波形,
    Figure PCTCN2020088567-appb-100007
    表示转置操作,ο表示矢量外积,
    Figure PCTCN2020088567-appb-100008
    为与各信号源相互独立的噪声张量,
    Figure PCTCN2020088567-appb-100009
    Figure PCTCN2020088567-appb-100010
    分别为
    Figure PCTCN2020088567-appb-100011
    在x轴和y轴方向上的导引矢量,对应于来波方向为
    Figure PCTCN2020088567-appb-100012
    的信号源,表示为:
    Figure PCTCN2020088567-appb-100013
    Figure PCTCN2020088567-appb-100014
    其中,
    Figure PCTCN2020088567-appb-100015
    Figure PCTCN2020088567-appb-100016
    分别表示稀疏子阵列
    Figure PCTCN2020088567-appb-100017
    在x轴和y轴方向上第i 1和i 2个物理天线阵元的实际位置,且
    Figure PCTCN2020088567-appb-100018
    Figure PCTCN2020088567-appb-100019
    稀疏子阵列
    Figure PCTCN2020088567-appb-100020
    的接收信号可用另一个三维张量
    Figure PCTCN2020088567-appb-100021
    表示为:
    Figure PCTCN2020088567-appb-100022
    其中,
    Figure PCTCN2020088567-appb-100023
    为与各信号源相互独立的噪声张量,
    Figure PCTCN2020088567-appb-100024
    Figure PCTCN2020088567-appb-100025
    分别为稀疏子阵列
    Figure PCTCN2020088567-appb-100026
    在x轴和y轴方向上的导引矢量,对应于来波方向为
    Figure PCTCN2020088567-appb-100027
    的信号 源,表示为:
    Figure PCTCN2020088567-appb-100028
    Figure PCTCN2020088567-appb-100029
    其中,
    Figure PCTCN2020088567-appb-100030
    Figure PCTCN2020088567-appb-100031
    分别表示稀疏子阵列
    Figure PCTCN2020088567-appb-100032
    在x轴和y轴方向上第i 3和i 4个物理天线阵元的实际位置,且
    Figure PCTCN2020088567-appb-100033
    计算子阵列
    Figure PCTCN2020088567-appb-100034
    Figure PCTCN2020088567-appb-100035
    的接收张量信号x 1和x 2的二阶互相关张量
    Figure PCTCN2020088567-appb-100036
    Figure PCTCN2020088567-appb-100037
    表示为:
    Figure PCTCN2020088567-appb-100038
    这里,x 1(l)和x 2(l)分别表示x 1和x 2在第三维度(即快拍维度)方向上的第l个切片,(·) *表示共轭操作;
    (3)由互相关张量
    Figure PCTCN2020088567-appb-100039
    得到一个增广的非均匀虚拟域面阵
    Figure PCTCN2020088567-appb-100040
    其中各虚拟阵元的位置表示为:
    Figure PCTCN2020088567-appb-100041
    其中,单位间隔d取为入射窄带信号波长λ的一半,即d=λ/2。定义维度集合
    Figure PCTCN2020088567-appb-100042
    Figure PCTCN2020088567-appb-100043
    则通过对互相关张量
    Figure PCTCN2020088567-appb-100044
    的理想值
    Figure PCTCN2020088567-appb-100045
    (无噪声场景)进行PARAFAC分解的模
    Figure PCTCN2020088567-appb-100046
    展开,可获得增广虚拟域面阵
    Figure PCTCN2020088567-appb-100047
    的等价接收信号
    Figure PCTCN2020088567-appb-100048
    的理想表示为:
    Figure PCTCN2020088567-appb-100049
    其中,
    Figure PCTCN2020088567-appb-100050
    Figure PCTCN2020088567-appb-100051
    Figure PCTCN2020088567-appb-100052
    是增广虚拟域面阵
    Figure PCTCN2020088567-appb-100053
    在x轴和y轴方向上的导引矢量,对应于来波方向为
    Figure PCTCN2020088567-appb-100054
    的信号源;
    Figure PCTCN2020088567-appb-100055
    表示第k个入射信号源的功率;
    Figure PCTCN2020088567-appb-100056
    表示克罗内克积;张量下标表示张量的PARAFAC分解的模展开操作;
    (4)
    Figure PCTCN2020088567-appb-100057
    中包含一个x轴分布为(-N x+1)d到(M xN x+M x-1)d、y轴分布为(-N y+1)d到(M yN y+M y-1)d的虚拟域均匀面阵
    Figure PCTCN2020088567-appb-100058
    中共有D x×D y个虚拟 阵元,其中D x=M xN x+M x+N x-1,D y=M yN y+M y+N y-1,
    Figure PCTCN2020088567-appb-100059
    表示为:
    Figure PCTCN2020088567-appb-100060
    通过选取虚拟域等价接收信号V中与
    Figure PCTCN2020088567-appb-100061
    各虚拟阵元位置相对应的元素,获取虚拟域均匀面阵
    Figure PCTCN2020088567-appb-100062
    的等价接收信号
    Figure PCTCN2020088567-appb-100063
    将其表示为:
    Figure PCTCN2020088567-appb-100064
    其中,
    Figure PCTCN2020088567-appb-100065
    Figure PCTCN2020088567-appb-100066
    Figure PCTCN2020088567-appb-100067
    Figure PCTCN2020088567-appb-100068
    为虚拟域均匀面阵
    Figure PCTCN2020088567-appb-100069
    在x轴和y轴方向上的导引矢量,对应于来波方向为
    Figure PCTCN2020088567-appb-100070
    的信号源;
    (5)在虚拟域均匀面阵
    Figure PCTCN2020088567-appb-100071
    中,分别沿x轴和y轴方向每隔一个阵元取一个大小为Y 1×Y 2的子阵列,则可以将虚拟域均匀面阵
    Figure PCTCN2020088567-appb-100072
    分割成L 1×L 2个互相部分重叠的均匀子阵列;将上述子阵列表示为
    Figure PCTCN2020088567-appb-100073
    根据子阵列
    Figure PCTCN2020088567-appb-100074
    对应虚拟域信号
    Figure PCTCN2020088567-appb-100075
    中相应位置元素,得到虚拟域子阵列
    Figure PCTCN2020088567-appb-100076
    的等价信号
    Figure PCTCN2020088567-appb-100077
    Figure PCTCN2020088567-appb-100078
    其中,
    Figure PCTCN2020088567-appb-100079
    Figure PCTCN2020088567-appb-100080
    Figure PCTCN2020088567-appb-100081
    Figure PCTCN2020088567-appb-100082
    为对应于
    Figure PCTCN2020088567-appb-100083
    方向的虚拟域子阵列
    Figure PCTCN2020088567-appb-100084
    在x轴和y轴上的导引矢量;经过上述操作,一共得到L 1×L 2个维度均为Y 1×Y 2的虚拟域子阵信号
    Figure PCTCN2020088567-appb-100085
    对这L 1×L 2个虚拟域子阵信号
    Figure PCTCN2020088567-appb-100086
    求平均值,得到一个虚拟域平滑信号
    Figure PCTCN2020088567-appb-100087
    Figure PCTCN2020088567-appb-100088
    对该虚拟域信号
    Figure PCTCN2020088567-appb-100089
    求得其四阶自相关张量
    Figure PCTCN2020088567-appb-100090
    表示为:
    Figure PCTCN2020088567-appb-100091
    (6)对四阶自相关张量
    Figure PCTCN2020088567-appb-100092
    进行CANDECOMP/PARACFAC分解以提取多维特征,得到结果表示如下:
    Figure PCTCN2020088567-appb-100093
    其中,
    Figure PCTCN2020088567-appb-100094
    Figure PCTCN2020088567-appb-100095
    为CANDECOMP/PARACFAC分解得到的两组正交因子矢量,分别表示x轴和y轴方向上的空间信息,
    Figure PCTCN2020088567-appb-100096
    Figure PCTCN2020088567-appb-100097
    Figure PCTCN2020088567-appb-100098
    为因子矩阵;取
    Figure PCTCN2020088567-appb-100099
    张成的空间,记作
    Figure PCTCN2020088567-appb-100100
    Figure PCTCN2020088567-appb-100101
    作为信号子空间,用一个张量
    Figure PCTCN2020088567-appb-100102
    表示该信号子空间,其中
    Figure PCTCN2020088567-appb-100103
    表示
    Figure PCTCN2020088567-appb-100104
    沿着第三维度的第k个切片,表示为:
    Figure PCTCN2020088567-appb-100105
    为了得到噪声子空间,对因子矩阵C x和C y求其正交补;C x的正交补记为
    Figure PCTCN2020088567-appb-100106
    C y的正交补记为
    Figure PCTCN2020088567-appb-100107
    Figure PCTCN2020088567-appb-100108
    则取
    Figure PCTCN2020088567-appb-100109
    作为噪声子空间,用张量
    Figure PCTCN2020088567-appb-100110
    表示该噪声子空间,
    Figure PCTCN2020088567-appb-100111
    表示
    Figure PCTCN2020088567-appb-100112
    沿着第三维度的第h个切片,表示为:
    Figure PCTCN2020088567-appb-100113
    (7)定义用于谱峰搜索的二维波达方向
    Figure PCTCN2020088567-appb-100114
    Figure PCTCN2020088567-appb-100115
    构造对应虚拟域均匀面阵
    Figure PCTCN2020088567-appb-100116
    的导引信息
    Figure PCTCN2020088567-appb-100117
    表示为:
    Figure PCTCN2020088567-appb-100118
    使用基于CANDECOMP/PARACFAC分解得到的噪声子空间构造张量空间谱函数
    Figure PCTCN2020088567-appb-100119
    表示如下:
    Figure PCTCN2020088567-appb-100120
    其中,<× {Q}>表示两个张量沿着第Q维度的模{Q}缩并操作,要求两个张量的第Q维度的大小相同;‖·‖ F表示Frobenius范数;
    Figure PCTCN2020088567-appb-100121
    Figure PCTCN2020088567-appb-100122
    沿着第1,2维度的模{1,2}缩并操作得到一个矢量
    Figure PCTCN2020088567-appb-100123
    Figure PCTCN2020088567-appb-100124
    得到空间谱函数
    Figure PCTCN2020088567-appb-100125
    之后,可以构造出对应二维搜索波达方向的空间谱,随后通过搜索谱峰所在位置对应的二维波达方向,即为入射信源的二维波达方向估计。
  2. 根据权利要求1所述的基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,其特征在于,步骤(1)所述的平面互质阵列结构可具体描述为:在平面坐标系xoy上构造一对稀疏均匀平面子阵列
    Figure PCTCN2020088567-appb-100126
    Figure PCTCN2020088567-appb-100127
    其中
    Figure PCTCN2020088567-appb-100128
    包含2M x×2M y个天线阵元,在x轴方向上和y轴方向上的阵元间距分别为N xd和N yd,其在xoy上的位置坐标为{(N xdm x,N ydm y),m x=0,1,...,2M x-1,m y=0,1,...,2M y-1};
    Figure PCTCN2020088567-appb-100129
    包含N x×N y个天线阵元,在x轴方向上和y轴方向上的阵元间距分别为M xd和M yd,其在xoy上的位置坐标为{(M xdn x,M ydn y),n x=0,1,...,N x-1,n y=0,1,...,N y-1};这里,M x、N x以及M y、N y分别为一对互质整数,且M x<N x,M y<N y;将
    Figure PCTCN2020088567-appb-100130
    Figure PCTCN2020088567-appb-100131
    按照(0,0)坐标处阵元重叠的方式进行子阵列组合,获得实际包含4M xM y+N xN y-1个物理天线阵元的互质面阵。
  3. 根据权利要求1所述的基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,其特征在于,步骤(3)所述的互相关张量
    Figure PCTCN2020088567-appb-100132
    可理想(无噪声场景)建模为:
    Figure PCTCN2020088567-appb-100133
    此时,
    Figure PCTCN2020088567-appb-100134
    Figure PCTCN2020088567-appb-100135
    等价于沿着x轴的一个增广虚拟域,
    Figure PCTCN2020088567-appb-100136
    等价于沿着y轴的一个增广虚拟域,从而可以得到非均匀虚拟域面阵
    Figure PCTCN2020088567-appb-100137
  4. 根据权利要求1所述的基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,其特征在于,步骤(5)所述的虚拟域均匀面阵
    Figure PCTCN2020088567-appb-100138
    的等价信号
    Figure PCTCN2020088567-appb-100139
    中保存了虚拟域面阵的空间结构信息,然而,由于
    Figure PCTCN2020088567-appb-100140
    可以视作一个单快拍的虚拟域信号,其统计量往往存在秩亏问题;因此,基于二维空间平滑的思想对虚拟域信号
    Figure PCTCN2020088567-appb-100141
    进行处理,构造多个等效快拍虚拟域子阵信号,对这些虚拟 域子阵信号进行求和平均后,求其四阶自相关张量;子阵列
    Figure PCTCN2020088567-appb-100142
    中阵元的位置表示为:
    Figure PCTCN2020088567-appb-100143
    通过子阵列
    Figure PCTCN2020088567-appb-100144
    对应选取虚拟域信号
    Figure PCTCN2020088567-appb-100145
    中相应位置元素,得到虚拟域子阵列
    Figure PCTCN2020088567-appb-100146
    的等价信号
    Figure PCTCN2020088567-appb-100147
  5. 根据权利要求1所述的基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,其特征在于,步骤(6)所述的提取虚拟域四阶自相关张量的多维特征以实现信号与噪声子空间分类,除了通过CANDECOMP/PARACFAC分解,还可以通过高阶奇异值分解实现,具体表示为:
    Figure PCTCN2020088567-appb-100148
    其中,× Q表示张量与矩阵沿着第Q维度的模Q内积;
    Figure PCTCN2020088567-appb-100149
    表示包含高阶奇异值的核张量,
    Figure PCTCN2020088567-appb-100150
    Figure PCTCN2020088567-appb-100151
    表示对应v四个维度的奇异矩阵;将D x的前K列和后Y 1-K列分开为信号子空间
    Figure PCTCN2020088567-appb-100152
    和噪声子空间
    Figure PCTCN2020088567-appb-100153
    将D y的前K列和后Y 2-K列分开为信号子空间
    Figure PCTCN2020088567-appb-100154
    和噪声子空间
    Figure PCTCN2020088567-appb-100155
  6. 根据权利要求1所述的基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,其特征在于,步骤(7)中得到空间谱函数
    Figure PCTCN2020088567-appb-100156
    之后进行二维谱峰搜索的具体步骤为:以a°为步长逐渐分别增加
    Figure PCTCN2020088567-appb-100157
    的值,二维波达方向
    Figure PCTCN2020088567-appb-100158
    的搜索起点为(-90°,0°),终点为(90°,180°);每个
    Figure PCTCN2020088567-appb-100159
    可以对应计算出一个
    Figure PCTCN2020088567-appb-100160
    的空间谱值,从而可以构造出一个对应于
    Figure PCTCN2020088567-appb-100161
    Figure PCTCN2020088567-appb-100162
    的空间谱。空间谱中存在K个峰值,该K个峰值所对应的
    Figure PCTCN2020088567-appb-100163
    的值,即为信源的二维波达方向估计。
  7. 根据权利要求5所述的基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法,其特征在于,步骤(7)中所述的张量空间谱构造还可以使用基于高阶奇异值分解得到的噪声子空间实现,表示为
    Figure PCTCN2020088567-appb-100164
    Figure PCTCN2020088567-appb-100165
    其中,(·) H表示共轭转置操作;得到空间谱函数
    Figure PCTCN2020088567-appb-100166
    之后,按照二维谱峰搜索过程,即可得到信源的二维波达方向估计。
PCT/CN2020/088567 2020-05-03 2020-05-03 基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法 WO2021068494A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
PCT/CN2020/088567 WO2021068494A1 (zh) 2020-05-03 2020-05-03 基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法
JP2021541668A JP7044289B2 (ja) 2020-05-03 2020-05-03 コプライム平面アレーのバーチャルドメインテンソル空間スペクトル検索に基づく高解像の正確な二次元到来方向推定方法
US17/395,478 US11300648B2 (en) 2020-05-03 2021-08-06 High-resolution, accurate, two-dimensional direction-of-arrival estimation method based on coarray tensor spatial spectrum searching with co-prime planar array

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/088567 WO2021068494A1 (zh) 2020-05-03 2020-05-03 基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/395,478 Continuation US11300648B2 (en) 2020-05-03 2021-08-06 High-resolution, accurate, two-dimensional direction-of-arrival estimation method based on coarray tensor spatial spectrum searching with co-prime planar array

Publications (1)

Publication Number Publication Date
WO2021068494A1 true WO2021068494A1 (zh) 2021-04-15

Family

ID=75437799

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/088567 WO2021068494A1 (zh) 2020-05-03 2020-05-03 基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法

Country Status (3)

Country Link
US (1) US11300648B2 (zh)
JP (1) JP7044289B2 (zh)
WO (1) WO2021068494A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113552532B (zh) * 2021-07-09 2022-03-22 浙江大学 基于耦合张量分解的l型互质阵列波达方向估计方法
CN114844544B (zh) * 2022-04-28 2024-05-14 中国人民解放军国防科技大学 一种基于低管秩张量分解的互质阵列波束成形方法、系统及介质

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110927661A (zh) * 2019-11-22 2020-03-27 重庆邮电大学 基于music算法的单基地展开互质阵列mimo雷达doa估计方法

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2883980B1 (fr) * 2005-04-01 2009-04-17 Thales Sa Procede et dispositif de goniometrie a haute resolution a un ordre pair arbitraire
JP2017116425A (ja) * 2015-12-24 2017-06-29 学校法人東京電機大学 Mimoレーダシステム、および信号処理装置
CN107037392B (zh) * 2017-03-01 2020-08-07 浙江大学 一种基于压缩感知的自由度增加型互质阵列波达方向估计方法
US11315032B2 (en) * 2017-04-05 2022-04-26 Yahoo Assets Llc Method and system for recommending content items to a user based on tensor factorization
JP7362727B2 (ja) * 2018-05-15 2023-10-17 ライトマター インコーポレイテッド フォトニック処理デバイス及び方法
CN109471086B (zh) * 2018-10-18 2020-11-24 浙江大学 基于多采样快拍和集阵列信号离散傅里叶变换的互质mimo雷达波达方向估计方法

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110927661A (zh) * 2019-11-22 2020-03-27 重庆邮电大学 基于music算法的单基地展开互质阵列mimo雷达doa估计方法

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
FAN JIN-YU, GU HONG, SU WEI-MIN, CHEN JIN-LI: "Co-prime MIMO Radar Multi-parameter Estimation Based on Tensor Decomposition", JOURNAL OF ELECTRONICS & INFORMATION TECHNOLOGY, KEXUE CHUBANSHE, vol. 37, no. 4, 1 April 2015 (2015-04-01), Kexue Chubanshe, pages 933 - 938, XP055800308, ISSN: 1009-5896, DOI: 10.11999/JEIT140826 *
GUI YUFENG, RAO WEI: "Two-dimensional coprime vector-sensor array signal processing based on tensor decompositions", JOURNAL OF NANCHANG INSTITUTE OF TECHNOLOGY, vol. 38, no. 4, 1 August 2019 (2019-08-01), pages 83 - 91, XP055800302 *
LIU CHUN-LIN; VAIDYANATHAN P. P.: "Tensor MUSIC in multidimensional sparse arrays", 2015 49TH ASILOMAR CONFERENCE ON SIGNALS, SYSTEMS AND COMPUTERS, IEEE, 8 November 2015 (2015-11-08), pages 1783 - 1787, XP032874535, DOI: 10.1109/ACSSC.2015.7421458 *
RAO WEI, LI DAN, ZHANG JIAN QIU: "A Tensor-Based Approach to L-Shaped Arrays Processing With Enhanced Degrees of Freedom", IEEE SIGNAL PROCESSING LETTERS., IEEE SERVICE CENTER, PISCATAWAY, NJ., US, vol. 25, no. 2, 1 February 2018 (2018-02-01), US, pages 1 - 5, XP055800309, ISSN: 1070-9908, DOI: 10.1109/LSP.2017.2783370 *
SHI JUNPENG, HU GUOPING, ZHANG XIAOFEI, SUN FENGGANG, ZHOU HAO: "Sparsity-Based Two-Dimensional DOA Estimation for Coprime Array: From Sum–Difference Coarray Viewpoint", IEEE TRANSACTIONS ON SIGNAL PROCESSING., IEEE SERVICE CENTER, NEW YORK, NY., US, vol. 65, no. 21, 1 November 2017 (2017-11-01), US, pages 5591 - 5604, XP055800311, ISSN: 1053-587X, DOI: 10.1109/TSP.2017.2739105 *

Also Published As

Publication number Publication date
JP7044289B2 (ja) 2022-03-30
JP2022511994A (ja) 2022-02-01
US11300648B2 (en) 2022-04-12
US20210364591A1 (en) 2021-11-25

Similar Documents

Publication Publication Date Title
CN111610486B (zh) 基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法
CN109932680B (zh) 一种基于平移互质阵列的非圆信号波达方向估计方法
CN108872929B (zh) 基于内插虚拟阵列协方差矩阵子空间旋转不变性的互质阵列波达方向估计方法
CN104898085B (zh) 一种极化敏感阵列参数估计的降维music算法
Chen et al. ESPRIT-like two-dimensional direction finding for mixed circular and strictly noncircular sources based on joint diagonalization
CN109655799B (zh) 基于iaa的协方差矩阵向量化的非均匀稀疏阵列测向方法
CN107037392B (zh) 一种基于压缩感知的自由度增加型互质阵列波达方向估计方法
CN102608565B (zh) 一种基于均匀圆阵列的波达方向估计方法
WO2021068495A1 (zh) 基于平面互质阵列块采样张量信号构造的自由度增强型空间谱估计方法
CN111610485B (zh) 基于平面互质阵列块采样张量信号构造的自由度增强型空间谱估计方法
WO2021068496A1 (zh) 基于结构化虚拟域张量信号处理的互质面阵二维波达方向估计方法
WO2022151511A1 (zh) 基于互相关张量的三维互质立方阵列波达方向估计方法
CN113673317B (zh) 基于原子范数最小化可降维的二维离格doa估计方法
CN107907855A (zh) 一种互素阵列转化为均匀线阵的doa估计方法及装置
CN107092004A (zh) 基于信号子空间旋转不变性的互质阵列波达方向估计方法
CN113552532B (zh) 基于耦合张量分解的l型互质阵列波达方向估计方法
WO2021068494A1 (zh) 基于平面互质阵列虚拟域张量空间谱搜索的高分辨精确二维波达方向估计方法
CN112731275B (zh) 一种基于零化插值的互质阵部分极化信号参数估计方法
CN104515969A (zh) 一种基于六角形阵列的相干信号二维doa估计方法
CN111965591A (zh) 一种基于四阶累积量矢量化dft的测向估计方法
CN109521393A (zh) 一种基于信号子空间旋转特性的波达方向估计算法
WO2023137813A1 (zh) 基于最优结构化虚拟域张量填充的超分辨互质面阵空间谱估计方法
CN109946663B (zh) 一种线性复杂度的Massive MIMO目标空间方位估计方法和装置
CN114200388A (zh) 基于四阶采样协方差张量去噪的子阵分置式l型互质阵列波达方向估计方法
CN109490821A (zh) 一种基于music算法的降维圆和非圆混合信号doa估计方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20874553

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021541668

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20874553

Country of ref document: EP

Kind code of ref document: A1