JP3285503B2 - Radio direction detection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a radio wave searching method for measuring arrival directions of a plurality of incident waves from signals received by an array antenna.

[0002]

2. Description of the Related Art Conventional MUSIC (MULTIPLE)
For example, as a document showing an Sig. O. Schmi
dt, “Multiple Emitter Local
Tion and Signal Parameter
Estimation, "IEEE Trans. A
ntennas & Propag. , AP-34,3, p
p. 276-280 (March 1986).
With reference to FIG. 9, a conventional radio wave direction finding device using the MUSIC algorithm will be described.

In FIG. 9, reference numeral 1 denotes an array antenna, and a received signal is input to a MUSIC processor. 2-6
Shows a processing flow in the MUSIC processor. 2 is a step of inputting an output of the array antenna 1 to obtain a correlation between the array elements to obtain a correlation matrix R, 3 is a step of calculating an eigenvalue and an eigenvector of the correlation matrix R,
4 is a step of obtaining an eigenvector corresponding to a minimum eigenvalue corresponding to noise among eigenvalues obtained in step 3; 5 is an array antenna 1 for each direction to be measured.
Calculating the reciprocal of the square of the absolute value of the inner value of the steering vector having the information of the reception phase difference and the gain and the eigenvector obtained in step 4 as an azimuth evaluation function;
6 is a step of obtaining the arrival direction θ of the incident wave from the peak appearing in the azimuth evaluation function obtained in step 5.

Here, the processing of the MUSIC algorithm will be described in detail. The MUSIC algorithm is M
When a signal of K waves smaller than M is incident on the array antenna 1 having the element antennas, the eigenvector corresponding to the minimum eigenvalue corresponding to the noise of the correlation matrix of the received signal of the array antenna and the incident direction of the radio wave are corresponded. This is a direction finding algorithm using the property that the steering vectors to be orthogonal are orthogonal.

The array antenna 1 is a non-directional antenna, and receives signals of the m-th array element antenna by x _m ,
The noise n _m, k-th arrival wave s _k, the arrival direction theta _k
And At this time, the received signal column vector of the array antenna 1 is represented by the following equation.

[0006]

(Equation 1)

In the equation (1), a (θ _k ) is a steering vector corresponding to the arrival direction θ _k represented by the following equation.

[0008]

(Equation 2)

In the equation (3), g _m (θ _k ) is a complex number expressing the complex gain exp () of the arrival direction θ _k of the element antenna m as a reception phase difference. Further, d _m is the position vector, p _k antenna elements m is a unit vector in the theta _k direction. In equation (1), the received signal vector on the left side is X, the matrix composed of the steering vector is A,
If the incoming signal vector is S and the noise vector is N,
It can be rewritten as the following equation (4). X = AS + N (4)

[0010] correlation matrix R from equation (4) is obtained as X ^* X ^T. The above processing is the processing performed in step 2 of FIG. When the number of array antennas M> the number of arriving signals K, the smallest eigenvalue of the correlation matrix R corresponds to the receiver noise, and the eigenvector E _min and the steering vector a (θ _k ) corresponding to the radio wave arrival direction θ _k are orthogonal. There is nature. Therefore, in the MUSIC algorithm, equation (5)
Is used as an azimuth evaluation function.

[0011]

(Equation 3)

The eigenvalues and eigenvectors of the correlation matrix R are obtained in step 3 of FIG. 8, and in step 4, the smallest eigenvalue λ of the M eigenvalues obtained is obtained.
_The eigenvector E _min corresponding to _min is selected, and the processing shown in Expression (5) is performed in Step 5. From the bearing evaluation function shown in equation (5), the eigenvector E _min
And steering vector a corresponding to radio wave arrival direction θ
When (θ) is orthogonal, the azimuth evaluation function P (θ) is maximized, and this direction θ is determined as the arrival direction of the radio wave.
Step 6 in FIG. 1 is a step of estimating the arrival direction of the incident wave from the peak appearing in the azimuth evaluation function.

In the MUSIC algorithm, it is ideal to find the azimuth evaluation function P (θ) for all directions θ. However, in practice, the azimuth evaluation function P (θ) is obtained only for a discrete θ sequence. Cannot calculate. The number of points in the θ column is defined as i. At this time, (all azimuths to be measured) / i
The orientation evaluation function can be calculated with a resolution of ## EQU1 ## and the larger the number of points i, the higher the resolution can be searched. However, as will be described later, the inner product operation of the matrix of Expression (5) requires The operation amount of Mi is required, and there is a problem that the operation amount increases accordingly.

[0014]

In the conventional radio wave direction finding apparatus using the MUSIC algorithm, there is a problem that the calculation for calculating the azimuth evaluation function is large and the measurement takes a long time.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and an object thereof is to provide a high-speed radio wave finding method with a small amount of calculation.

[0016]

SUMMARY OF THE INVENTION A radio azimuth detecting method according to the present invention correlates signals received from a plurality of element antennas arranged at substantially equal intervals on a circle, calculates an eigenvector, and calculates an evaluation function. In the configuration in which the direction of arrival of the received signal is calculated from the result, a steering vector having an element different from the elements of the other steering vectors is calculated in the steering vector having the phase characteristic and the amplitude characteristic for each direction of the element antenna as elements. An operation object extraction step extracted as an object, an FFT processing step of performing a fast Fourier transform on elements of the steering vector extracted as the operation object and eigenvector elements corresponding to the calculated receiver noise,
A multiplication step of multiplying the result obtained by the T processing, and an inverse FFT processing step of performing an inverse fast Fourier transform on the element of the multiplication result, wherein the value of the element obtained by the inverse FFT processing is An evaluation function is calculated by using a value obtained by performing an operation equivalent to another steering vector to which an operation equivalent to a steering vector extracted as an operation target can be applied.

Further, for another steering vector having an arrangement reverse to the sequence of the elements of the steering vector to be operated, the complex conjugate for calculating the complex conjugate number of the FFT processing result of the element of the steering vector to be operated is calculated. A number calculation step is provided, and the calculated complex conjugate number is used as an input to a subsequent multiplication step instead of the processing in the FFT processing step.

In the radio azimuth detecting method according to the present invention, received signals from a plurality of element antennas arranged at substantially equal intervals on a circle are correlated, an eigenvector is calculated, an evaluation function is calculated, and a result is calculated from the result. In the above-described configuration for determining the direction of arrival of a received signal, a steering vector element having a phase characteristic and an amplitude characteristic with respect to each direction of the element antenna as elements, and a steering vector element having an element different from other steering vector elements as an operation target. Extracting an operation target extraction step, extracting a required sub-matrix by dividing a matrix composed of elements of the steering vector extracted as an operation target into sub-matrices, and extracting the calculated receiver noise Addition and subtraction for sum and difference for required elements of corresponding eigenvectors And step comprises a multiplication step for multiplying the vector with the output of the extracted sub-matrix and the subtraction step as an element, and to calculate the evaluation function using the values of elements obtained by this multiplication step.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION

Embodiment 1 FIG. Hereinafter, an embodiment of the present invention will be described. In the present embodiment, the steering vector a
(Θ) ^H Matrix A ^H Focusing on the fact that multiplication of the eigenvector (hereinafter referred to as the steering vector matrix) and the eigenvector maintains the same relationship regardless of the order of the columns, the same result can be obtained. The following sub-matrix is configured. Here, H represents Hermitian transposition. Also, the Hermitian transpose of the steering vector a (θ) ^H Is simply referred to as a steering vector. Further, the operation of the sub-matrix and the eigenvector E _min is performed by performing a Fourier transform, multiplying the resultant, and then performing an inverse Fourier transform to reduce the amount of calculation. This basic method is shown in FIG. . Furthermore, this replacement or rearrangement is not indispensable, but only the calculation of a steering vector having an element different from that of another steering vector is required.

FIG. 1 is a diagram for explaining a method for calculating an evaluation function of the radio wave direction finding apparatus according to the present invention. In the figure, reference numeral 11 denotes a step of extracting a steering vector having a different element from the other steering elements as a calculation target. In the present embodiment, a step of rearranging the order of the row vector (steering vector) by a predetermined method; Dividing the steering vector matrix into sub-matrices, 13
Performs a fast Fourier transform (FFT operation) on the sequence c _n obtained from the elements of the row vector of the matrix obtained in step 12 and the sequence e _n obtained from the elements of the eigenvector E _min corresponding to the minimum eigenvalue of the correlation matrix R. Performing 14; multiplying the sequence C (k) and E (k) obtained in step 13 to obtain a sequence Y (k);
15 is a step of performing an inverse fast Fourier transform (IFFT-Inverse FFT) operation of the sequence Y (k) obtained in the step 14; step permuting all order of the sequence y _n obtained as a result of applying even 15) to the other sub-matrix of a predetermined manner, 17 for each element of the sequence z _n which is rearranged in the step 16, This is the step of calculating the reciprocal of the square of the absolute value to obtain the azimuth evaluation function P (θ _n ).

Next, the operation of the first embodiment according to the present invention will be described. In the present invention, using equally spaced circular array as shown in FIG. 2, the number of Ki number of elements of the array are mentioned 2, i.e. 2, 4, 8, · · ·, 2 ⁿ And
It is also assumed that the number of orientations of the orientation evaluation function to be obtained is a power of two. Assuming that the number of element antennas is M, the number of evaluation direction points is i, and the element antennas are 0 in the x-axis direction as shown in FIG. 2, and 1, 2,.
M-1. The position vectors (starting from the center of the circle) are d ₀ , d ₁ ,..., D _M−1 . Evaluation direction unit vector is a position vector on the unit circle in the same plane at the same center as the circle is arranged element antennas (starting at the center of the circle), the x-axis positive direction is p _0, counterclockwise .., P _i−1 are sequentially set around p ₁ , p ₂ ,. In this case, the MUSIC algorithm, in order to obtain an evaluation function, performs an inner product operation between eigenvector E _{mi n} corresponding to the minimum eigenvalue lambda _min of the steering vector and correlation matrix R shown in equation (6).

[0022]

(Equation 4)

In the present invention, in order to perform the calculation of the equation (6), the calculation is performed according to steps 11 to 16 in FIG. First, before describing step 11, the features of the i × M matrix B composed of the steering vector a (θ) ^H in equation (6) will be described. The (n, m) component b _nm of this matrix is represented by the following equation (7).

[0024]

(Equation 5)

In equation (7), the amplitude is set to 1 for simplicity. Since p _n · d _m is the inner product expressed as follows. _{p n · d m = rcosα nm} (8)

[0026] Here, r is the radius of the circle to place the antenna, alpha _nm is an angle formed p _n and d _m, expressed as follows.

[0027]

(Equation 6)

At this time, considering the (n + i / M, m) component and the (n + i / M, 0) component,

[0029]

(Equation 7)

From the above equations (10) and (11), it can be seen that the matrix B has a row vector (numerical sequence) in which components circulate one by one for each i / M row. 0 ≦ n for simplicity
Considering the range of ≦ (i / M) −1, the following equation is established between the i / M−n row vector and the n row vector.

[0031]

(Equation 8)

From equations (12) and (13), 0 ≦ n ≦ (i /
It can be seen that among the row vectors in the range of (M) -1, there is a row vector (sequence) having an array in a reverse order to each other. Also, cosα _{n, m} has two properties as shown in equations (14) and (15).

[0033] _{cos (α n, m + iΔθ} ) = cosα n, m (14) cos (-α n, m) = cosα n, m (15)

From this property, there are at most i / 2 types of angles α _{n, m} having different cosine. Therefore, the angle α is newly expressed as follows. α _{n, m} = α _h = hΔθ (h = 0,1..., i / 2-1) (16)

Further components of the components of the matrix B is also corresponding to the alpha _h B to be expressed as b _h. At this time, the column of the steering vector matrix B when the number M of antennas is 4 and the number i of evaluation directions of the evaluation function p (θ _n ) is 16 is shown in FIG.
(A). In FIG. 4, the components are represented as b _h ⇒h, b ^* _h ⇒
Only the suffix was described by the method of h *. FIG. 4 (a)
Indicates that four columns from left to right are position vectors d ₀ , d ₁ ,
d _2, corresponding to d _3, a value eg M ₀ at that position, d ₀ the position of such 0,1,2,3, ..., is 2,1, etc. d ₀ and p _1,
d ₀ and p _2, d ₀ and p _3, d ₀ and p ₄ ... d ₀ and p _14, d
Shows the inner product of ₀ and p _15. 1, 2 etc. are the inner product values, and 1 *, 2 * etc. represent the complex conjugate values of 1, 2 etc.

FIG. 4 shows that a row vector in which components circulate appears at every i / M (= 4) row as shown in equations (10) and (11). Therefore, in step 1, the order of the rows of the matrix B is rearranged by a method as shown in FIG. The meaning of FIG. 3 indicates a procedure for performing a process of changing the order before the rearrangement in FIG. 4A to the order after the rearrangement in FIG. That is, in FIG.
Means that the first and second digits of the i / M base number are considered, and each i / M row is a subgroup. Step 3
3 shifts one of the inner products by one position vector d (four in the direction p) in the above specific example.

The row vector of the matrix B before rearrangement is the 0th, 1st,..., I−1th from the top. FIG.
In the flow shown in (a), d and p are initialized in step 21, a loop using the direction p in steps 22 to 24 as a variable, and a position vector d in 22 to 26 (for four of p)
Is a variable, and n is given at 22. As a result, according to the processing order, the value of n is 0,
1, 2,..., I−1. On the other hand, the flow shown in FIG.
The loop has been replaced. As a result, the value of n is 0, i / M, 2i / M,.
Are arranged in that order. In step 11 in FIG. 1, the matrix B is rearranged according to the processing order of the flow in FIG. 3B so that the n-th row vector of the matrix B is arranged.

As a result of rearrangement as shown in FIG.
The matrix shown in FIG. 4A is a matrix as shown in FIG. It is assumed that the rearranged matrix is C. From the rearranged matrix of FIG. 4, the matrix C is an i / M × 1 matrix of an M × M submatrix in which the elements of the row vector are circulating. In step 12 of FIG. 1, this sub-matrix is represented by the formula (1)
As shown in 7), the data is divided as D ₀ , D ₁ ,..., _{Di / M-1} . In addition, the components of the sub-matrix _Dq are represented as Expression (18) in consideration of the fact that they are cyclic sequences,
To extract the sequence c _n. Also, it represents the eigenvectors E _min in equation (6) as in equation (19) to obtain a sequence e _n.

[0039]

(Equation 9)

At this time, the product of the matrix C and the eigenvector E _min is as shown in equation (20).

[0041]

(Equation 10)

[0042] Among the operations of equation (20), D _q E _min is as the equation (18) and (19) formula (21).

[0043]

[Equation 11]

In equation (21), () modM indicates that a number modulo M is given. From equation (21), D _q
It can be seen that E _min is given by a circular convolution operation. Of course, the product of the other sub-matrices D and E _min is also given as a circular convolution operation. That is, by performing row rearrangement on the steering vector matrix B as shown in FIG. 4, the evaluation function derivation operation of Expression (6) can be performed by a circular convolution operation.

This circular convolution operation is performed by FFT,
It can be calculated as a product in the frequency domain. That is, the convolution operation and the FFT have a relationship represented by Expression (22).

[0046]

(Equation 12)

In equation (22), ⇔ represents a Fourier transform pair, and the following relationship is assumed. _{y n ⇔Y (k), e} n ⇔E (k), C n ⇔C (k) (23)

Using the property of equation (22), a circular convolution operation is performed as steps 12 to 15 in FIG. In step 12, the steering vector matrix C is divided into sub-matrices. In step 13, the FFT processing is performed on the sequence c _n and the sequence e _n obtained from the elements of the sub-matrix and the eigenvector, and C (k) and E
(K) is obtained. In step 14, C (k) and E (k)
To obtain Y (k). In step 15, obtain the sequence y _n performs IFFT processing on the sequence Y (k).
Then, in step 16, creating a bound sequence in the order of the original sub-matrix of the sequence y _n obtained from the result of the processing in step 15 from the other sub-matrix in step 13, and shows the order of the sequence in Figure 3 Rearrangement is performed in the reverse direction of the method to obtain the inner product operation result z _n of Expression (6).

In step 17, the reciprocal of the absolute value of z _n is calculated to obtain the azimuth evaluation function P (θ _n ). By performing the calculation using the FFT in this manner, the amount of calculation can be reduced as compared with a case where a simple convolution calculation is performed. As mentioned earlier, the rearrangement of the order was made to make it easier to understand the progress of the expression, and the gist of the present invention is to reduce the amount of computation,
For that purpose, it is clear that only the elements of the steering vector having elements different from the others in the steering vector matrix are to be calculated, and FFT, multiplication and IFFT may be performed. For this purpose, it is desirable that the direction i is an integral multiple of the antenna M, preferably 2 n times, and M is 2 n times.

Comparing the amount of operation with the number of complex multiplications, FIG.
The amount of calculation required for the inner product calculation of equation (6) performed in step 5 of above is Mi. On the other hand, the calculation amount required for steps 13 to 15 in FIG. 1 is given by equation (24).
Also, the FFT for the elements of the steering vector matrix
Since the processing can be calculated before the measurement, the calculation amount is subtracted from the calculation amount, and the calculation amount is given by Expression (25), and the calculation amount is further reduced.

[0051]

(Equation 13)

This is smaller than Mi. Number of antenna elements M = 4 Specific examples for simplicity of explanation in the first embodiment, although the evaluation direction i = 16 to be the direction of the evaluation, as another example, M = 8, the direction finely And i =
Considering the case of 1024, the ratio of Mi to the value of equation (24) is 2.4, and the amount of calculation is reduced to about 2.4. Also, the ratio of Mi to the value of equation (25) is 3.2, and the FFT for the elements of the steering vector matrix is
If the processing is calculated before performing the measurement, the amount of calculation is 3.
It is reduced to about half.

Embodiment 2 Hereinafter, an embodiment of the present invention will be described. FIG. 5 is a diagram for explaining a method of calculating an evaluation function in the radio wave direction detecting method according to the present invention. In the figure,
1 denote the same or corresponding parts, and a description thereof will be omitted.

In the second embodiment, the amount of calculation is further reduced by omitting the FFT operation of the row vector c ′ _n having the reverse sequence to the sequence of numbers c _{n in} the processing of the first embodiment.

As shown in Expressions (12) and (13), there are sub-matrices in which the component elements of the row vector are the same and arranged in reverse order. By utilizing this, the amount of calculation can be further reduced. First, to express the sub-matrix D _q of the formula (18) as follows.

[0056]

[Equation 14]

In the equation (26), the element in {} is the column m
Is a function sequence. At this time, _{Di / Mq} is expressed by the following equation based on the relations expressed by equations (10) to (13).

[0058]

(Equation 15)

Focusing on the elements of the row vectors of the sub-matrices of Expressions (26) and (27), it can be seen that the sub-matrices are sequences in reverse order. Here, if it is assumed that _cm = c ' ^* _{(-m) modM} , Equation (27) becomes Equation (28).

[0060]

(Equation 16)

From the nature of equation (29), equation (30) is related, so equation (28) becomes equation (31).

[0062]

[Equation 17]

From the above, the product of the sub-matrix D _{i / Mq} and the eigenvector E _min is a convolution operation as shown in equations (32) and (33).

[0064]

(Equation 18)

This convolution operation can be performed as a product in the frequency domain by using FFT as in the case of FIG. Further, from the relationship of c ′ _n ⇔C ^* (k), the FFT processing of the sequence c ′ _{n in} reverse order is given by the complex conjugate of the FFT processing result C (k) of the sequence c _n . First, a process shown in FIG. 1 for the sub-matrix having a sequence c _n as an element, then, by performing the processing for the sub-matrix with a reverse order of the elements, FFT for sequence c _'n
Processing can be omitted. Obviously, even if the processing for the sub-matrix having reverse elements previously performed, the sequence c _n of the sub-matrix having a turn sequence c _n as an element
Can be omitted.

The operation of the second embodiment will be described with reference to FIG. In step 18, the complex conjugate C ^* (k) of C (k) calculated in advance is obtained. In step 14, the F
Multiply the FT processing result E (k) by C ^* (k) to obtain a sequence Y
(K) is obtained. In step 15, IFFT processing is performed on Y (k) obtained in step 14 to obtain a sequence.
In step 19, the reordering of the sequence in accordance with Equation (32), obtain the sequence y _n.

[0067] Treatment as in FIG. 5, c 'Fourier transform of _n, since obtained by taking the complex conjugate of c _n of Fourier transform C (k), c' Fourier transform for _n be omitted And the amount of calculation can be further reduced.

Embodiment 3 Hereinafter, an embodiment of the present invention will be described. In the third embodiment, by focusing on the nature of the sub-matrix D _q after permutation, to reduce the amount of calculation. FIG. 6 is a basic operation flow chart for explaining processing in the present embodiment. The eigenvector is divided into two parts, and the multiplication of the steering vector with the sub-matrix is replaced with an arithmetic expression that takes addition and subtraction into account, thereby reducing the number of multiplication operations. Here, the matrix shown in FIG. 4 will be described as an example. First, 4 after rearrangement of FIG.
One sub-matrix of × 4 to A _4. In step 41 of FIG. 6, the image is further divided into sub-matrices. That is, this sub-matrix can be divided into smaller sub-matrices as shown in equation (34).

[0069]

[Equation 19]

[0070] As can be seen from equation (34), also A2 becomes Izure sub-matrix of diagonal elements, and non-diagonal elements, the A ₂ and the complex conjugate A ₂ ^*. Step 4
2 by dividing the sub-matrix A ₂ in the real part A _2r and the imaginary part A _2i in, performing the product of A ₄ and eigenvectors E _{mi n} as follows. That is, in step 44, the process is performed by dividing the equation (35) into a real part and an imaginary part.

[0071]

(Equation 20)

From equation (35), the product of the first term and the eigenvector in ｛｝ is as shown in FIG. 7, and the product of the second term and the eigenvector in ｛｝ is as shown in FIG. can get. In these figures, e _{0 to} e ₄ are elements of the eigenvector E _min as shown in Expression (36). Also,
The calculation of the upper half and the lower half of FIG. 7 is the same, and the calculation results of the upper half and the lower half of FIG. E _min = (e ₀ , e ₁ , e ₂ , e ₃ ) ^T (36)

Comparing the amount of calculation in the above-described method with the number of complex multiplications, when simply multiplying one submatrix with the eigenvector, the number of elements of the eigenvector is set to N and N ² Complex multiplication is required.
On the other hand, when the calculation is performed by the above-described method, the load of multiplication of a complex number and a real number is １ ／ of that of multiplication of a complex number.
, The number of multiplications at the top of FIG. 7 is (N / 2)
It is ² , and the lower part does not need to perform the calculation because the calculation result is equal to the upper part. Further matrices A _2r is considered to be construed in a real, computational load needed to perform the processing of FIG. 7 is a ^{(N / 2) 2/2} . The calculation load of FIG. 8 is equal to the calculation load of FIG. 7 ignoring the process of inverting the sign. Therefore, the calculation load of the whole ^{(N / 2) 2 = N} 2/4
Becomes From this result, the amount of calculation can be reduced to about 1/4 compared to simply performing the inner product calculation.

[0074]

As described above, according to the present invention, the calculation of the azimuth evaluation function of the radio wave direction finding apparatus is performed by extracting only the steering vectors having different elements from the others, subjecting them to FFT processing, multiplying them, and Since processing is performed, there is an effect that a high-speed radio wave direction detection method with a small amount of calculation can be obtained.

Further, for the elements of the steering vector having an arrangement reverse to that of a certain steering vector element, the number of complex conjugations of the already calculated FFT processing result is calculated. There is an effect.

Further, the azimuth evaluation function is calculated by extracting only the elements of the steering vector having different elements from the others, dividing the matrix composed of the extracted elements of the steering vector into predetermined sub-matrices, A predetermined element of an eigenvector corresponding to the minimum eigenvalue of the correlation matrix is added / subtracted, a multiplication of the submatrix and a vector having the result of the addition / subtraction as an element is performed, and an evaluation function is obtained from the multiplication result. So,
This has the effect of further reducing the amount of calculation.

[Brief description of the drawings]

FIG. 1 is a diagram showing a method of calculating an azimuth evaluation function according to the present invention.

FIG. 2 is a diagram showing an arrangement of an array antenna and an evaluation direction unit vector according to the present invention.

FIG. 3 is a diagram for explaining the operation rearrangement of elements of a steering vector matrix according to the present invention.

FIG. 4 is a diagram illustrating rearrangement of elements when M = 4 and i = 16 in the first embodiment.

FIG. 5 is a diagram showing a method of calculating an azimuth evaluation function according to Embodiment 2 of the present invention.

FIG. 6 is a diagram showing a method of calculating an azimuth evaluation function according to Embodiment 3 of the present invention.

FIG. 7 is a diagram illustrating a matrix operation in a method of calculating an azimuth evaluation function according to a third embodiment.

FIG. 8 is a diagram illustrating a matrix operation of a method for calculating an azimuth evaluation function according to the third embodiment.

FIG. 9 is a diagram illustrating a direction estimation method of a conventional radio wave direction finding apparatus.

[Explanation of symbols]

1 an array antenna, 2 a step of obtaining a correlation matrix,
3 a step of obtaining an eigenvalue and an eigenvector of a correlation matrix, 4 a step of obtaining an eigenvector corresponding to a minimum eigenvalue, 5 a step of obtaining an azimuth evaluation function, 6 a step of obtaining an arrival direction of an incident wave from a peak of the azimuth evaluation function, 11 other steering Extracting a steering vector element having a different element from the vector as an operation target, rearranging the order of the row vector of the steering vector matrix, 12 dividing the steering vector matrix into sub-matrices, 13
Performing FFT processing, 14 performing multiplication, 15 performing IFFT processing, 16 rearranging step, 17 calculating the reciprocal of the absolute value,
18 complex conjugate calculation means, 19 rearrangement step, 4
1 sub-matrix division step, 42 step of dividing into a real part and an imaginary part, 43 addition / subtraction means, 44 multiplication means,
45 Converting to a complex number.

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-8-201498 (JP, A) JP-A-8-29513 (JP, A) JP-A-7-140221 (JP, A) JP-A-6-2014 347529 (JP, A) JP-A-5-196716 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01S 3/00-3/74 G01S ⁷ /00-7/42

Claims (3)

(57) [Claims] 1. Correlation processing is performed on received signals from a plurality of element antennas arranged at substantially equal intervals on a circle, an eigenvector is calculated, an evaluation function is calculated, and the arrival direction of the received signal is obtained from the result. In the configuration, an operation object extraction step of extracting, as an operation object, a steering vector having elements different from other steering vector elements in a steering vector having phase characteristics and amplitude characteristics for each direction of the element antenna as an element, An FFT processing step of performing a fast Fourier transform on the extracted steering vector elements and the eigenvector elements corresponding to the calculated receiver noise, a multiplication step of multiplying the result obtained by the FFT processing, Inverse FFT processing for performing an inverse fast Fourier transform on the element of the multiplication result And performing the same operation on the values of the elements obtained by the inverse FFT processing and other steering vector elements to which the same operation as the steering vector extracted as the operation target can be applied. A radio wave direction detecting method in which an evaluation function is calculated with the obtained value. 2. A complex conjugate number for calculating a complex conjugate number of an FFT processing result of the element of the operation target steering vector with respect to another steering vector having an arrangement reverse to the sequence of the element of the operation target steering vector. A calculating step is provided, and the calculated complex conjugate number is F
2. The radio wave direction detecting method according to claim 1, wherein a subsequent multiplication step is input instead of the FT processing step. 3. Correlation processing is performed on received signals from a plurality of element antennas arranged at substantially equal intervals on a circle, an eigenvector is calculated, an evaluation function is calculated, and the arrival direction of the received signal is obtained from the result. In the configuration, an operation target extraction step of extracting, as an operation target, a steering vector element having an element different from other steering vector elements in a steering vector element having a phase characteristic and an amplitude characteristic for each direction of the element antenna, A sub-matrix extraction step of dividing a matrix composed of the elements of the steering vector extracted as the operation target into sub-matrices and extracting a necessary sub-matrix; and a required eigenvector corresponding to the calculated receiver noise. An addition / subtraction step for obtaining a sum and a difference for the elements; And a multiplication step for multiplying the vector with the output of the sub-matrix and the subtraction step as an element, Telecommunications azimuth detection method to calculate the evaluation function using the values of elements obtained by the multiplication step.