CN104614611A - Method for detecting damaged element of receiving antenna array online - Google Patents

Abstract

The invention belongs to the online test method that the receiving antenna array in radar application is damaged array element, including establishes Test database, determines the autocorrelation matrix for receiving signal, determines virtual array to be measured, the processing of autocorrelation matrix column vectorization, column vector The sequence of element determines the number of Virtual array on each array element position in virtual array to be measured, determines that receiving antenna array whether there is impaired array element and determine impaired element number of array and corresponding array element coordinate bit. The invention, which has, not to be limited by the number size for being damaged array element in the scale of receiving antenna array and aerial array, in the case where same array scale, identical impaired element number of array, compared with support vector cassification method, time spent by the invention on-line checking is only the 1/50-1/4 of support vector cassification method, performance is obviously improved, and detection efficiency and reliability are higher. Thus the present invention has the features such as detection method is easy, and required detection time is short, detection efficiency and high reliablity.

Description

The online test method of the impaired array element of a kind of receiving antenna array

Technical field

The invention belongs to a kind of method detecting impaired element position in receiving antenna array, particularly a kind of relating to when there is impaired array element in the array run detects the method for its position.The method utilizes the information of the signal that other intact array elements receive in impaired array, detects the particular location of impaired array element under the receiving array damage situations of complexity exactly.

Background technology

Antenna array signals treatment technology has been widely used in the fields such as radar, astronomy, earthquake, sonar and radio communication now.Aerial array is made up of, for receiving spacing wave and processing to the received signal the multiple sensors be placed on different coordinate position, space.But in actual applications, the reasons such as severe physical environment and its shelf-life of sensor hardware can cause part array element in receiving antenna array to occur damaging, cannot normal received signal, claim these array elements that cannot normally work to be impaired array element.Along with the continuous increase of aerial array scale in radar application, the possibility that there is impaired array element also increases thereupon, cause array performance is seriously declined because impaired array element loses a lot of useful information, therefore detect impaired array element and replace impaired array element in time to ensureing that array performance is most important, especially true for modern sophisticated electronic war environment.

The impaired array element detection method generally adopted at present mainly contains far field beam inversefouriertransform method, genetic algorithm, matrix method, signal source reconstruction method, parameter model, support vector machine classifier method, Bayes's compressed sensing method and singular value decomposition method etc.But these methods above all have some limitations and defect: inversefouriertransform method, signal source reconstruction method, parameter model is applicable to the less situation of receiving array array element number, and when array scale is larger, be difficult to obtain far field beam, detect the hydraulic performance decline of impaired array element method, limit the application of these methods above in large scale array.Matrix method needs the equation solving concerned matrix, when the matrix solved is that (in matrix equation, matrix of coefficients is very sensitive to less error for ill-condition matrix, to such an extent as to there is very large difference with potential solution in the solution that very little disturbance can cause equation to be obtained, such matrix of coefficients is called ill-condition matrix) time, the method iterative process used there will be the very slow problem of convergence.May there is locally optimal solution when data volume is larger in the detection method based on genetic algorithm, cause the impaired array that detects and the impaired array of reality inconsistent (namely causing flase drop).But and take the disadvantages such as larger physical resource (machine internal memory and working time) when support vector cassification method exists process large scale training data and higher-order matrix number.

Summary of the invention

The object of the invention is the defect for existing in background technology, the online test method of the impaired array element of a kind of receiving antenna array of research and design, this detection method is not by scale and the impaired array number quantitative limitation of receiving antenna; Expend time in shorter under identical array damage situations to reach, under equal complex array damage situations, detect the objects such as more accurate.

Solution of the present invention is: first carry out relevant treatment according to the coordinate position of the array number of intact array, each array element, the direction vector that forms for the sense and wavelength thereof setting up Test database, determine the maximum number of Virtual array on each array element position and each array element position in covariance matrix, virtual array, to set up Test database; Then to array received to be measured to signal data carry out autocorrelation matrix, to determine virtual array to be measured and to determine the number of Virtual array on its each array element position, and then determine that whether receiving antenna array is by the number undermining impaired array element, determine the coordinate position of each impaired array element finally by the method for searching thoroughly; Namely the present invention realizes its goal of the invention with this.

Therefore the inventive method comprises:

Step 1. sets up Test database: first using first of the array left side array element as true origin, and to the array number by intact array, the coordinate position of each array element, the direction vector formed for the sense and wavelength thereof of setting up Test database carries out relevant treatment, determine covariance matrix, and then by the number of Virtual array on each array element position in this matrix determination virtual array and each array element position, what wherein number was maximum is Virtual array number corresponding in true origin, and by the array number of intact array, the coordinate position of each array element, the maximal value of the Virtual array number in virtual array on array element position is stored in processor,

Step 2. determines the autocorrelation matrix of Received signal strength: to array received to be measured to signal data carry out auto-correlation processing by following formula:

$\begin{array}{c}{R}_{\mathrm{xx}}=E\left({\mathrm{XX}}^H\right)=b\left({\mathrm{\θ}}_s\right)R_{\mathrm{ss}}{b}^H\left({\mathrm{\θ}}_s\right)+{\mathrm{\σ}}_n^{2}I={\mathrm{\σ}}_s^{2}b\left({\mathrm{\θ}}_s\right)b^H\left({\mathrm{\θ}}_s\right)+{\mathrm{\σ}}_n^{2}I\\ ={\mathrm{\σ}}_s^2\left[\begin{array}{c}{e}^{-j2\mathrm{\π}{d}_0\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\\ e^{-j2\mathrm{\π}{d}_1\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\\ .\\ .\\ .\\ e^{-j2\mathrm{\π}{d}_{M-1}\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\end{array}\right]\·\left[\begin{array}{cccc}{e}^{j2\mathrm{\π}{d}_0\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& e^{j2\mathrm{\π}{d}_1\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& ...& e^{j2\mathrm{\π}{d}_{M-1}\mathrm{s\; in}{\mathrm{\θ}}_s/\mathrm{\λ}}\end{array}\right]+{\mathrm{\σ}}_n^{2}I\\ ={\mathrm{\σ}}_s^2\left[\begin{array}{cccc}{e}^{j2\mathrm{\π}(d_0-d_0)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& e^{j2\mathrm{\π}(d_1-d_0)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& ...& e^{j2\mathrm{\π}(d_{M-1}-d_0)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\\ e^{j2\mathrm{\π}(d_0-d_1)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& e^{j2\mathrm{\π}(d_1-d_1)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& ...& e^{j2\mathrm{\π}(d_{M-1}-d_1)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\\ .& .& & .\\ .& .& ...& .\\ .& .& & .\\ e^{j2\mathrm{\π}(d_0-d_{M-1})\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& e^{j2\mathrm{\π}(d_1-d_{M-1})\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& ...& e^{j2\mathrm{\π}(d_{M-1}-d_{M-1})\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\end{array}\right]+{\mathrm{\σ}}_n^{2}I\end{array}$

In formula: R _xx, R _ssbe respectively the autocorrelation matrix of array received signal to be measured and target source signal, E () is covariance matrix computing, and X is array received signal to be measured, [] ^hfor conjugate transpose operation, $b\left({\mathrm{\θ}}_s\right)={\left[\begin{array}{cccc}{e}^{-j2\mathrm{\π}{d}_0\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& e^{-j2\mathrm{\π}{d}_1\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& ,...,& e^{-j2\mathrm{\π}{d}_{M-1}\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\end{array}\right]}^T$ For the position angle of direction vector corresponding to this array to be measured, echo signal and wavelength are respectively θ _sand λ, d _ifor coordinate, the i=0 of array element in array to be measured, 1 ..., M-1 (M is array number), i is unit matrix, with be respectively the power of echo signal and noise;

Step 3. determines virtual array to be measured: by step 2 matrix R _xxbe D={v by the set expression of the difference of the array element coordinate position in each element successively after column vector process _k, v in formula _kfor the arbitrary element in set D, by each element v _kthe array composition difference array that corresponding array element is formed; Different element composition set D in D will be gathered _dis, composition set D _dismethod be: be as the criterion by first element in set D, the element occurred after all then removes with arbitrary element equivalence above, forms successively by different element V _fthe set D formed _dis, this set D _disin the array element corresponding to different element form virtual array to be measured;

Wherein: k is each element v in set D _kcoordinate (position) in coordinate axis, span are [0, M-1], and F is set D _disin each element V _fcoordinate (position) in coordinate axis, span are [-M+1, M-1];

Step 4. autocorrelation matrix column vector process: to step 2 gained autocorrelation matrix R _xxmerge by row, namely arrange cascade in order from the 1st row to M, form a M ²the column vector y of × 1 dimension:

$\begin{array}{c}y={[{\left[R_{\mathrm{xx}}\right]}_{\mathrm{1,1}},{\left[R_{\mathrm{xx}}\right]}_{\mathrm{2,1}},...,{\left[R_{\mathrm{xx}}\right]}_{M,1},{\left[R_{\mathrm{xx}}\right]}_{\mathrm{2,1}}...,{\left[R_{\mathrm{xx}}\right]}_{M,2},...,{\left[R_{\mathrm{xx}}\right]}_{p,q},...,{\left[R_{\mathrm{xx}}\right]}_{M,M}]}^T\\ ={\mathrm{\σ}}_s^2{[1,e^{j2\mathrm{\π}(d_0-d_1)\sin {\mathrm{\θ}}_s/\mathrm{\λ}},...,e^{j2\mathrm{\π}(d_0-d_{M-1})\sin {\mathrm{\θ}}_s/\mathrm{\λ}},...,e^{j2\mathrm{\π}(d_{M-1}-d_0)\sin {\mathrm{\θ}}_s/\mathrm{\λ}},...,e^{j2\mathrm{\π}(d_{M-1}-d_{M-1})\sin {\mathrm{\θ}}_s/\mathrm{\λ}}]}^T\\ +{\mathrm{\σ}}_n^2{[\mathrm{1,0},...,\mathrm{0,0,1},...,0,...,1,...,\mathrm{0,1}]}^T\end{array}$

Then element set different in column vector y is become column vector

[R in formula _xx] _p,qfor autocorrelation matrix R _xxin the value of p capable q row, the span of p and q is 1,2 ..., M, [] ^trepresent transpose operation;

Step 5. column vector the sequence of element: by step 4 gained column vector sort as follows: from the element of first in column vector y, after this, then removing of all elements occurred in column vector y and above arbitrary element equivalence, thus form the column vector be made up of different element successively to make column vector in each arrangement of elements order and D _disin each arrangement of elements order corresponding;

Step 6. determines the number of Virtual array on each array element position in virtual array to be measured: according to the result in step 4, with column vector in first element be as the criterion, determine successively in column vector y with column vector in the element number of each equivalence, the element number of each equivalence corresponds respectively to the number of the Virtual array in virtual array to be measured on each array element position, the number of the Virtual array on the maximal value respective coordinates initial point wherein in Virtual array number;

Step 7. determines whether receiving antenna array exists impaired array element: the maximal value of the Virtual array number on array element position in maximal value in step 6 gained Virtual array number and step 1 Test database compared, if equal, array to be measured is intact array, exits detection; If be less than the maximal value of the Virtual array number in Test database on array element position, then judge that it exists impaired array element, proceed to step 8;

Step 8. determines impaired element number of array and corresponding array element coordinate position: the maximal value of the Virtual array number in comparison step 1 Test database on array element position and the difference of step 6 gained Virtual array number maximal value, to determine the number of impaired array element, and then determine the coordinate position of each impaired array element in the following manner:

Step 8.1. determines impaired array element coordinate position: all combinations of the different array elements being first radix by whole array elements (M) composition known for array elements coordinate position with the determined impaired element number of array of step 8, in each combination, the coordinate position of each array element is all corresponding identical with the coordinate position of each array element in step 1 Test database, and each combination performs successively in the following manner:

A. an optional combination is as impaired array element group (N number of) to be measured, and all the other array elements removed in whole array element in selected impaired array element group to be measured after all array element are formed new current array to be measured, then the coordinate position of array element each in current array is from left to right formed row vector successively, the row vector after this row vector and its negate number makes cross correlation process;

B. cross correlation process gained matrix in step a is carried out to column vector process, obtains column vector h, again successively by element composition Virtual array position set (w) different in this column vector, the method of its set (w) is: be as the criterion with first element in column vector h, then removing, to form current virtual array of the element occurred after all and above arbitrary element equivalence;

C. be as the criterion by first element in column vector h, determine successively to gather the element number of each equivalence in (w) with Virtual array position in column vector h, the element number of each equivalence corresponds respectively to the number of Virtual array on each array element position in current virtual array, now by the number corresponding comparison of the number of Virtual array on array element position each in current virtual array with Virtual array on each array element position in virtual array to be measured described in step 6, if not quite identical, then directly perform steps d; If completely the same, then using in step a this array element combination of selecting as just fixed impaired array element group, and store for subsequent use;

D. then respectively all the other different array elements are combined and press step a successively, b, c circulate execution, if all the other each group of results performed respectively all do not meet and to store standby condition described in step c, then above-mentioned steps c at the beginning of fixed impaired array element group be actual impaired array element group, the coordinate position respectively by array element in this actual impaired array element group is the coordinate position of corresponding array element in step 1 Test database; If once occur still meeting and to store standby condition described in step c in all the other each group of results performed respectively, then perform step 8.2;

Step 8.2: detect meeting the two groups of array element groups storing standby condition: any one array element finding out corresponding different coordinate position in two groups, then by the output situation of actual this array element of test, if no-output, then the array element combination at this array element place is actual impaired array element group; If there is output, then another array element group is actual impaired array element group; Coordinate position respectively by array element in actual impaired array element group is the coordinate position of corresponding array element in step 1 Test database.

Described in above-mentioned steps 1 and then by the number of Virtual array on each array element position in this matrix determination virtual array and each array element position, its method carries out vectorization process to this covariance matrix, determine the multiplicity of each identical element in vector, each multiplicity is the number of Virtual array on corresponding array element position.

The present invention has and does not limit by the number size of impaired array element in the scale of receiving antenna array and aerial array, at equal array scale, when identical impaired element number of array, the present invention is compared with support vector cassification method, support vector cassification method is at 4, 5, consuming timely when 10 array elements are impaired be respectively 1.0225s, 1.4072s, 22.1638s, the inventive method is then respectively 0.0203s, 0.2974s, 4.8251s, time spent by embodiment of the present invention is only the 1/50-1/4 of support vector cassification method, performance improves obviously, detection efficiency and reliability higher.Thus to have detection method easy in the present invention, and required detection time is short, detection efficiency and reliability high.

Accompanying drawing illustrates:

Fig. 1 is actual impaired receiving antenna array (in figure, black triangle is impaired array element, and white triangles shape is intact array element);

Fig. 2 is Virtual array number schematic diagram corresponding on each array element position in the specific embodiment of the invention; In figure: x-axis is the coordinate position of Virtual array corresponding on each array element position in actual array; Y-axis is the number of the Virtual array in virtual array on each array element position;

In Fig. 3: the impaired array element coordinate position schematic diagram in the impaired array of a. detected by the specific embodiment of the invention; B. be the coordinate position schematic diagram of impaired array element in actual impaired array.

Concrete implementation content:

The present embodiment be for the operation wavelength 10cm of aerial array, array element distance d=5cm be signal wavelength lambda half, in the aperture of 25 times of half-wavelengths (5cm), have the uniform linear array of M=25 omnidirectional's array element, carrier frequency f=3GHz.Signal arrival bearing θ ₀be 0 °, setting horizontal linear is x-axis, each array element from true origin distance from close to being far followed successively by d ₀, d ₁..., d ₂₄( i=0 ..., 24), the sequence number of actual impaired array element is that { 7,8,12,13,15,16,17,20,21,23}, this actual impaired array schematic diagram is as shown in Figure 2.

Step 1. sets up Test database: first using first of the array left side array element as true origin.Array, when not having impaired, determines that the set that the difference of any two array element coordinate positions in array forms is:

D _int＝{v ₀＝0d＝0,v ₁＝1d＝5,…，v ₂₄＝24d＝120,v _-1＝-1d＝-5,v ₀＝0d＝0,…,v ₂₃＝23d＝115,

v _-1＝-1d＝-5,…,v ₂₂＝22d＝110,…,v ₀＝-24d＝-120,v ₀＝-23d＝-115,…,v ₀＝0d＝0}

Then, determine the number of Virtual array on each array element position of virtual array under present case, process is as follows:

A. determine that in virtual array, each array element coordinate position is

{v _-24＝-24d＝-120,v _-23＝-23d＝-115,…,v _-2＝-2d＝-10,

v _-1＝-1d＝-5,v ₀＝0d＝0,v ₁＝1d＝5,…,v ₂₄＝24d＝120}

B. the number of Virtual array on each array element position in virtual array is determined,

{1,2,3,…,24,25,24,…,3,2,1}

C. by array number 25, array element distance d=5cm, the maximal value 25 of the Virtual array number in virtual array on array element position is stored in processor;

Step 2. determines the autocorrelation matrix of Received signal strength: to array received to be measured to signal data carry out auto-correlation computation process, determine autocorrelation matrix R _xxfor:

$R_{\mathrm{xx}}=\left[\begin{array}{ccccc}5.0618& ...& -0.9640-1.3453j& ...& 0.3464-1.0272j\\ -1.7223+0.7969j& ...& -0.2591+0.7133j& ...& 0.5212+0.1465j\\ .& .& .& .& .\\ .& .& .& .& .\\ .& .& .& .& .\\ -2.2027-0.2719j& ...& 0.1488+1.5245j& ...& -0.0648-0.2305j\\ 0.3464+1.0272j& ...& -1.4074-1.9790j& ...& 5.0618\end{array}\right]$

Step 3. determines virtual array to be measured: by step 2 matrix R _xxafter column vector process by the set expression of the difference of the array element coordinate position in each element be successively

D＝{v ₀＝0d＝0,v ₁＝1d＝5,…,v ₂₄＝24d＝120,v _-1＝-1d＝-5,v ₀＝0d＝0,…,v ₂₃＝23d＝115,

v _-1＝-1d＝-5,…,v ₂₂＝22d＝110,…,v _-24＝-24d＝-120,v _-23＝-23d＝-115,…,v ₀＝0d＝0}

Be as the criterion by first element in set D, then the removing of the element occurred after all and arbitrary element equivalence above, form the set D be made up of different element successively _dis

D _dis＝{0,1d＝5,2d＝10,...,24d＝120,-1d＝-5,-2d＝-10,...,-24d＝-120}

This set D _disin the array element corresponding to different element form current virtual array;

Step 4. autocorrelation matrix column vector: to gained autocorrelation matrix R in step 2 _xxcarry out the column vector that column vector obtains one 625 × 1 dimension

y＝[5.0618,-1.7223+0.7969j,…,-2.2027-0.2719j,0.3464+1.0272j,…,

-0.9640-1.3453j,…,-1.4074-1.9790j,…,-0.0648-0.2305j,5.0618] ^T

Step 5. column vector the sequence of element: from the element of first in column vector y, after this, then the removing of all elements occurred in vectorial y and arbitrary element equivalence above, forms the vector be made up of different element successively

$\begin{array}{c}\tilde{y}=[5.0618,-1.7223+0.7969j,...,-2.2027-0.2719j,0.3464+1.0272j,...\\ {0.5212+0.1465j,0.2584-0.9844j]}^T\end{array}$

To make column vector in each arrangement of elements order and D _disin each arrangement of elements order corresponding;

Step 6. determines the number of Virtual array on each array element position in virtual array to be measured: according to the result in step 4, by column vector in first element be as the criterion, determine successively in column vector y with column vector in the element number of each equivalence and then determine the number of the Virtual array in virtual array to be measured on each array element position and Virtual array number maximal value for in c element, by calculate determine

In y with the 1st element ${\left[\tilde{y}\right]}_0=5.0618$ Equivalent element number is 15, ${\left[{\tilde{f}}_{\mathrm{dam}}\right]}_0=15;$

In y with the 2nd element ${\left[\tilde{y}\right]}_1=-1.7223+0.7969j$ Equivalent element number is 1, ${\left[{\tilde{f}}_{\mathrm{dam}}\right]}_1=1;$

.

.

.

In y with the 23rd element ${\left[\tilde{y}\right]}_{22}=0.5433-0.1597j$ Equivalent element number is 1, ${\left[{\tilde{f}}_{\mathrm{dam}}\right]}_{22}=1;$

.

.

.

In y with the 48th element ${\left[\tilde{y}\right]}_{47}=0.5212+0.1456j$ Equivalent element number is 1, ${\left[{\tilde{f}}_{\mathrm{dam}}\right]}_{47}=1;$

In y with the 49th element ${\left[\tilde{y}\right]}_{48}=0.2584-0.9844j$ Equivalent element number is 1, ${\left[{\tilde{f}}_{\mathrm{dam}}\right]}_{48}=1;$ Namely determine in column vector y with column vector in the element number of each equivalence

${\tilde{f}}_{\mathrm{dam}}=[\mathrm{15,1},...,1,...,\mathrm{1,1}]$

And the maximal value of Virtual array number in virtual array on array element position

Step 7. determines whether receiving antenna array exists impaired array element: step 6 gained Virtual array number maximal value 15 is less than the Virtual array number maximal value 25 in step 1 Test database on array element position, determines that current array exists impaired array element, performs step 8;

Step 8. determines impaired element number of array and corresponding array element coordinate position: determine that the number of impaired array element to be measured is 10 (being N=25-15=10) by the Virtual array number maximal value 25 in comparison step 1 Test database and the difference of step 6 gained Virtual array number maximal value 15, and then determine the coordinate position of each impaired array element in the following manner:

Step 8.1. determines impaired array element coordinate position: first, and be all different array element combinations of benchmark by determined 10 the impaired element number of array of whole 25 array elements composition step 8 known for array elements coordinate position, the array element coordinate position in each combination is respectively:

{{d ₀,d ₁₁,d ₁₂,d ₁₃,…,d ₂₂,d ₂₃,d ₂₄},{d ₀,d ₁,d ₁₂,d ₁₃,…,d ₂₂,d ₂₃,d ₂₄},...,{d ₀,d ₁,d ₂,d ₃,…,d ₁₃,d ₁₄,d ₁₅}}

In each combination the coordinate position of each array element all with the coordinate position (d of each array element in step 1 Test database ₀, d ₁..., d ₂₄(0,5 ..., 120})) corresponding identical, each combination performs according to following way of example:

A. first, an optional combination, the coordinate position d of each array element in its combination ₁, d ₂, d ₃, d ₄, d ₅, d ₆, d ₇, d ₈, d ₉, d ₁₀(5,10,15,20,25,30,35,40,45,50}) as impaired array element group, all the other array elements removed in whole array element in selected impaired array element group after all array element are formed new current array to be measured, and the coordinate position having remained array element is respectively: d ₀, d ₁₁, d ₁₂, d ₁₃..., d ₂₄(0,55,60,75 ..., 120}).Then the coordinate position of array element each in current array is from left to right formed row vector [d successively ₀, d ₁₁, d ₁₂, d ₁₃..., d ₂₄] ([0,55,60,75 ..., 120]), by the row vector-[d after this vector and its negate number ₀, d ₁₁, d ₁₂, d ₁₃..., d ₂₄] ([-0 ,-55 ,-60 ,-75 ... ,-120]) make cross correlation process; Obtain matrix:

B. to cross correlation process gained matrix R in step a _ddcarry out column vector process, obtain column vector: h=[h ₀=0d=0, h ₁₁=11d=55 ..., h ₀=0d=0 ..., h _-1=-1d=-5, h ₀=0d=0],

Successively element different in this column vector is formed successively the set of Virtual array position again:

w＝{0d＝0,11d＝55,12d＝60,...,24d＝120,-11d＝-55,...,-23d＝-115,-24d＝-120}；

C. by column vector: h=[h ₀=0d=0, h ₁₁=11d=55 ..., h ₀=0d=0 ..., h _-1=-1d=-5, h ₀=0d=0] in first element (h ₀=0d=0) be as the criterion, determine column vector h=[h successively ₀=0d=0, h ₁₁=11d=55 ..., h ₀=0d=0 ..., h _-1=-1d=-5, h ₀=0d=0] in gather (w={0d=0,11d=55,12d=60 with Virtual array position, ..., 24d=120 ,-11d=-55, ... ,-23d=-115 ,-24d=-120}) in the element number { 15 of each equivalence, 4,3 ... 1,1 ... 1,1}, the element number of each equivalence corresponds respectively to the number of the Virtual array in current virtual array on each array element position; With the Virtual array number unit number on array element position each in virtual array to be measured described in step 6 inconsistent, perform steps d;

D. then respectively all the other different array elements are combined and press step a successively, b, c circulate execution;

When performing the combination of m group array element wherein,

A. d will be positioned at ₀, d ₁..., d ₂₄(0,5 ..., 120}) on all 25 array elements compositions linear array in remove 10 and lay respectively at d ₇, d ₈, d ₁₂, d ₁₃, d ₁₅, d ₁₆, d ₁₇, d ₂₀, d ₂₁, d ₂₃array element on (35,40,60,75,80,75,100,105,115), namely these 10 array elements are impaired array element, remain the coordinate position of array element respectively: d ₀, d ₁, d ₂, d ₃..., d ₂₄(0,5,10,15 ..., 120).Then the coordinate position of array element each in current array is from left to right formed row vector [d successively ₀, d ₁, d ₂, d ₃..., d ₂₄] ([0,5,10,15 ..., 120]), by the row vector-[d after this vector and its negate number ₀, d ₁, d ₂, d ₃..., d ₂₄] ([-0 ,-5 ,-10 ,-15 ... ,-120]) make cross correlation process; Obtain matrix:

B. to cross correlation process gained matrix R in step a _ddcarry out column vector process, obtain column vector ({ h ₀=0d=0, h ₁=1d=5 ..., h ₀=0d=0 ..., h _-2=-2d=-10, h ₀=0d=0}), more successively element different in this column vector is formed successively the set of Virtual array position:

w＝{0d＝0,1d＝5,2d＝10,...,24d＝120,-1d＝-5,...,-23d＝-115,-24d＝-120}，

C. by column vector ({ h ₀=0d=0, h ₁=1d=5 ..., h ₀=0d=0 ..., h _-2=-2d=-10, h ₀=0d=0}) in first element (h ₀=0d=0) be as the criterion, determine column vector ({ h successively ₀=0d=0, h ₁=1d=5 ..., h ₀=0d=0 ..., h _-2=-2d=-10, h ₀=0d=0}) in gather with Virtual array position (w={0d=0,1d=5,2d=10 ..., 24d=120 ,-1d=-5 ...,-23d=-115 ,-24d=-120}) in the element number { 15,9 of each equivalence, 7 ..., 1,1 ..., 1,1}, the element number of each equivalence corresponds respectively to the number of the Virtual array in current virtual array on each array element position, now with the Virtual array number unit number on each array element position in virtual array to be measured described in step 6 unanimously, then using in this combination step a the array element in this array element combination of selecting as just fixed impaired array element, and store for subsequent use;

To search the combination of all possible array element successively thoroughly, finally occur that two groups meet the condition of storing for future use,

Wherein the array element combination coordinate position of first group is:

d ₇,d ₈,d ₁₂,d ₁₃,d ₁₅,d ₁₆,d ₁₇,d ₂₀,d ₂₁,d ₂₃({35,40,60,65,75,80,85,100,105,115})

The array element combination coordinate position of second group is:

d ₁,d ₃,d ₄,d ₇,d ₈,d ₉,d ₁₁,d ₁₂,d ₁₆,d ₁₇({5,15,20,35,40,45,55,60,80,85})

Perform step 8.2;

Step 8.2: detect meeting the two groups of array elements combination storing standby condition: finding out differently in two groups of impaired array element coordinate positions of step 8.1 has d ₁, d ₃, d ₄, d ₉, d ₁₁, d ₂₀, d ₂₁, d ₂₃(5,15,20,45,55,100,105,115}), test in second group of array element combination herein and be positioned at d ₁the output situation of ({ 5}) upper array element, find that this array element has output data, what therefore determine first group of array element combination is actual impaired array element, and the coordinate position respectively by array element in the reality impaired array element group of these 10 impaired array elements is the coordinate position of corresponding array element in step 1 Test database: { d ₇, d ₈, d ₁₂, d ₁₃, d ₁₅, d ₁₆, d ₁₇, d ₂₀, d ₂₁, d ₂₃, be also { 35,40,60,65,75,80,85,100,105,115}.

According to the signal that impaired array received arrives, the method utilizing the present invention to propose determines the virtual array that impaired array is corresponding; Virtual array number comparison diagram in the virtual array that Virtual array number in virtual array on each array element position is corresponding with intact array on each array element position as shown in Figure 2; In the impaired array element coordinate position that the method proposed by the present invention is detected and the impaired array of reality, the schematic diagram of impaired array element coordinate position as shown in Figure 3.

Present embodiment is by Matlab simulation run, and be all the uniform linear array of array number 25, when actual impaired element number of array is equal, the present invention detects to expend time in the support vector cassification method in background technology and compares:

	4 array elements are impaired	5 array elements are impaired	10 array elements are impaired
				The present invention expends time in (/s)	0.0203s	0.2974s	4.8251s
Support vector cassification method (/s)	1.0225s	1.4072s	22.1638s

Can find out from above-mentioned list, under identical homogenous linear battle array and equal impaired condition, the time that the present invention expends and the time-consuming contrast of support vector cassification method in the background technology generally adopted, can find out that the time spent by embodiment of the present invention is only the 1/50-1/4 of support vector cassification method, performance improves obviously, detection efficiency and reliability higher.

Claims (2)

1. an online test method for the impaired array element of receiving antenna array, comprising:

Step 1. sets up Test database: first using first of the array left side array element as true origin, and to the array number by intact array, the coordinate position of each array element, the direction vector formed for the sense and wavelength thereof of setting up Test database carries out relevant treatment, determine covariance matrix, and then by the number of Virtual array on each array element position in this matrix determination virtual array and each array element position, what wherein number was maximum is Virtual array number corresponding in true origin, and by the array number of intact array, the coordinate position of each array element, the maximal value of the Virtual array number in virtual array on array element position is stored in processor,

Step 2. determines the autocorrelation matrix of Received signal strength: to array received to be measured to signal data carry out auto-correlation processing by following formula:

$\begin{array}{c}{R}_{\mathrm{xx}}=E\left({\mathrm{XX}}^H\right)=b\left({\mathrm{\θ}}_s\right)R_{\mathrm{ss}}{b}^H\left({\mathrm{\θ}}_s\right)+{\mathrm{\σ}}_n^{2}I={\mathrm{\σ}}_s^{2}b\left({\mathrm{\θ}}_s\right)b^H\left({\mathrm{\θ}}_s\right)+{\mathrm{\σ}}_n^{2}I\\ ={\mathrm{\σ}}_s^2\left[\begin{array}{c}{e}^{-j2\mathrm{\π}{d}_0\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\\ e^{-j2\mathrm{\π}{d}_1\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\\ .\\ .\\ .\\ e^{-j2\mathrm{\π}{d}_{M-1}\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\end{array}\right]\·\left[\begin{array}{cccc}{e}^{j2\mathrm{\π}{d}_0\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& e^{j2\mathrm{\π}{d}_1\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& ...& e^{j2\mathrm{\π}{d}_{M-1}\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\end{array}\right]+{\mathrm{\σ}}_s^{2}I\\ ={\mathrm{\σ}}_s^2\left[\begin{array}{cccc}{e}^{j2\mathrm{\π}(d_0-d_0)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& e^{j2\mathrm{\π}(d_1-d_0)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& ...& e^{j2\mathrm{\π}(d_{M-1}-d_0)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\\ e^{j2\mathrm{\π}(d_0-d_1)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& e^{j2\mathrm{\π}(d_1-d_1)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& ...& e^{j2\mathrm{\π}(d_{M-1}-d_1)\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\\ .& .& & .\\ .& .& ...& .\\ .& .& & .\\ e^{j2\mathrm{\π}(d_0-d_{M-1})\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& e^{j2\mathrm{\π}(d_1-d_{M-1})\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& ...& e^{j2\mathrm{\π}(d_{M-1}-d_{M-1})\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\end{array}\right]+{\mathrm{\σ}}_s^{2}I\end{array}$

In formula: R _xx, R _ssbe respectively the autocorrelation matrix of array received signal to be measured and target source signal, E () is covariance matrix computing, and X is array received signal to be measured, [] ^hfor conjugate transpose operation, $b\left({\mathrm{\θ}}_s\right)={\left[\begin{array}{cccc}{e}^{-j2\mathrm{\π}{d}_0\sin {\mathrm{\θ}}_s/\mathrm{\λ}}& e^{-j2\mathrm{\π}{d}_1\sin {\mathrm{\θ}}_s/\mathrm{\λ}},& ...,& e^{-j2\mathrm{\π}{d}_{M-1}\sin {\mathrm{\θ}}_s/\mathrm{\λ}}\end{array}\right]}^T$ For the position angle of direction vector corresponding to this array to be measured, echo signal and wavelength are respectively θ _sand λ, d _ifor coordinate, the i=0 of array element in array to be measured, 1 ..., M-1, M are array number, i is unit matrix, with be respectively the power of echo signal and noise;

Step 3. determines virtual array to be measured: by step 2 matrix R _xxbe D={v by the set expression of the difference of the array element coordinate position in each element successively after column vector process _k, v in formula _kfor the arbitrary element in set D, by each element v _kthe array composition difference array that corresponding array element is formed; Different element composition set D in D will be gathered _dis, composition set D _dismethod be: be as the criterion by first element in set D, the element occurred after all then removes with arbitrary element equivalence above, forms successively by different element V _fthe set D formed _dis, this set D _disin the array element corresponding to different element form virtual array to be measured;

Wherein: k is each element v in set D _kcoordinate position in coordinate axis, span are [0, M-1], and F is set D _disin each element V _fcoordinate position in coordinate axis, span are [-M+1, M-1];

Step 4. autocorrelation matrix column vector process: to step 2 gained autocorrelation matrix R _xxmerge by row, namely arrange cascade in order from the 1st row to M, form a M ²the column vector y of × 1 dimension:

$\begin{array}{c}y={[{\left[R_{\mathrm{xx}}\right]}_{\mathrm{1,1}},{\left[R_{\mathrm{xx}}\right]}_{\mathrm{2,1}},...,{\left[R_{\mathrm{xx}}\right]}_{M,1},{\left[R_{\mathrm{xx}}\right]}_{\mathrm{2,1}},...,{\left[R_{\mathrm{xx}}\right]}_{M,2},{\left[R_{\mathrm{xx}}\right]}_{p,q},..,{\left[R_{\mathrm{xx}}\right]}_{M,M}]}^T\\ ={\mathrm{\σ}}_s^2{[1,e^{j2\mathrm{\π}(d_0-d_1)\sin {\mathrm{\θ}}_s/\mathrm{\λ}},...,e^{j2\mathrm{\π}(d_0-d_{M-1})\sin {\mathrm{\θ}}_s/\mathrm{\λ}},...,e^{j2\mathrm{\π}(d_{M-1}-d_0)\sin {\mathrm{\θ}}_s/\mathrm{\λ}},...,e^{j2\mathrm{\π}(d_{m-1}-d_{M-1})\sin {\mathrm{\θ}}_s/\mathrm{\λ}}]}^T\\ +{\mathrm{\σ}}_n^2{[\mathrm{1,0},...,0,0,1,...,0,...,1,...,0,1]}^T\end{array}$

Then element set different in column vector y is become column vector

In formula for autocorrelation matrix R _xxin the value of p capable q row, the span of p and q is 1,2 ..., M, [] ^trepresent transpose operation;

Step 5. column vector the sequence of element: by step 4 gained column vector sort as follows: from the element of first in column vector y, after this, then removing of all elements occurred in column vector y and above arbitrary element equivalence, thus form the column vector be made up of different element successively to make column vector in each arrangement of elements order and D _disin each arrangement of elements order corresponding;

Step 6. determines the number of Virtual array on each array element position in virtual array to be measured: according to the result in step 4, with column vector in first element be as the criterion, determine successively in column vector y with column vector in the element number of each equivalence, the element number of each equivalence corresponds respectively to the number of the Virtual array in virtual array to be measured on each array element position, the number of the Virtual array on the maximal value respective coordinates initial point wherein in Virtual array number;

Step 7. determines whether receiving antenna array exists impaired array element: the maximal value of the Virtual array number on array element position in maximal value in step 6 gained Virtual array number and step 1 Test database compared, if equal, array to be measured is intact array, exits detection; If be less than the maximal value of the Virtual array number in Test database on array element position, then judge that it exists impaired array element, proceed to step 8;

Step 8. determines impaired element number of array and corresponding array element coordinate position: the maximal value of the Virtual array number in comparison step 1 Test database on array element position and the difference of step 6 gained Virtual array number maximal value, to determine the number of impaired array element, and then determine the coordinate position of each impaired array element in the following manner:

Step 8.1. determines impaired array element coordinate position: all combinations of the different array elements being first radix by whole array element compositions known for array elements coordinate position with the determined impaired element number of array of step 8, in each combination, the coordinate position of each array element is all corresponding identical with the coordinate position of each array element in step 1 Test database, and each combination performs successively in the following manner:

A. an optional combination is as impaired array element group to be measured, and all the other array elements removed in whole array element in selected impaired array element group to be measured after all array element are formed new current array to be measured, then the coordinate position of array element each in current array is from left to right formed row vector successively, the row vector after this row vector and its negate number makes cross correlation process;

B. cross correlation process gained matrix in step a is carried out to column vector process, obtains column vector h, again successively by element composition Virtual array position set different in this column vector, the method of its set is: be as the criterion with first element in column vector h, then removing, to form current virtual array of the element occurred after all and above arbitrary element equivalence;

C. be as the criterion by first element in column vector h, determine the element number of each equivalence in gathering with Virtual array position in column vector h successively, the element number of each equivalence corresponds respectively to the number of Virtual array on each array element position in current virtual array, now by the number corresponding comparison of the number of Virtual array on array element position each in current virtual array with Virtual array on each array element position in virtual array to be measured described in step 6, if not quite identical, then directly perform steps d; If completely the same, then using in step a this array element combination of selecting as just fixed impaired array element group, and store for subsequent use;

D. then respectively all the other different array elements are combined and press step a successively, b, c circulate execution, if all the other each group of results performed respectively all do not meet and to store standby condition described in step c, then above-mentioned steps c at the beginning of fixed impaired array element group be actual impaired array element group, the coordinate position respectively by array element in this actual impaired array element group is the coordinate position of corresponding array element in step 1 Test database; If once occur still meeting and to store standby condition described in step c in all the other each group of results performed respectively, then perform step 8.2;

Step 8.2: detect meeting the two groups of array element groups storing standby condition: any one array element finding out corresponding different coordinate position in two groups, then by the output situation of actual this array element of test, if no-output, then the array element combination at this array element place is actual impaired array element group; If there is output, then another array element group is actual impaired array element group; Coordinate position respectively by array element in actual impaired array element group is the coordinate position of corresponding array element in step 1 Test database.

2. by the online test method of the impaired array element of receiving antenna array described in claim 1, it is characterized in that described and then by the number of Virtual array on each array element position in this matrix determination virtual array and each array element position in step 1, its method carries out vectorization process to this covariance matrix, determine the multiplicity of each identical element in vector, each multiplicity is the number of Virtual array on corresponding array element position.