JP2000121716A - Radio wave propagation estimating equipment - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable estimating propagating wave number and the propagating direction of each wave, with high precision. SOLUTION: This estimating equipment is provided with a relative matrix operating means 9 for calculating a relative matrix from a receiving signal, a minimum characteristic value calculating means 10 for calculating a minimum characteristic value from the relative matrix, a characteristic vector operating means 11 for calculating a characteristic vector corresponding to a calculated minimum characteristic value, an MUSIC spectrum calculating means 12 for calculating an MUSIC spectrum from a calculated characteristic vector, an propagating direction estimating means 13 for estimating the propagating direction of waves containing an artificial wave corresponding to a peak waveform, from the position of each peak waveform contained in the MUSIC spectrum, a power operating means 14 for calculating each power value of the propagating wave or the artificial wave which correspond to each estimated peak waveform, and an propagating wave number estimating means 15 for estimating the propagating wave number, by regarding each peak waveform of a power value of at least a specified value out of power values as the peak waveform of the propagating wave.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a radio wave arrival direction estimating apparatus for estimating a radio wave arrival direction, and more particularly to a radio wave arrival direction estimating apparatus for receiving a plurality of radio waves arriving from each direction with a plurality of antennas, The present invention relates to a radio wave direction-of-arrival estimation device that estimates the direction of arrival of an incoming wave.

[0002]

2. Description of the Related Art For example, when controlling illegal radio waves,
It is necessary to specify the source of the illegal radio wave. In such a case, measurement vehicles equipped with the radio wave arrival direction estimating device are arranged at a plurality of positions, and the arrival direction of the illegal radio wave is measured by the radio wave arrival direction estimating device mounted on each measurement vehicle. Then, draw an extension of the direction of arrival of the same illegal radio wave from the position of each measurement vehicle,
The position of the intersection on each of these extensions is estimated as the transmission position of the illegal radio wave.

In a radio wave arrival direction estimating apparatus having such a function, when a plurality of radio wave arrival directions are simultaneously estimated (measured) using a plurality of antennas, generally, M
This is performed using the USIC (Multiple Signal Classification) method.

[0004] The configuration and calculation procedure of the radio wave direction-of-arrival estimation apparatus using the MUSIC method will be described below with reference to FIG.

[0005] In FIG.
For example, M antennas 1 are disposed on, for example, the measurement vehicle described above. Then, from the distant K transmission positions 2, K radio waves are incident on the M antennas 1 as arriving waves 3. Since the transmitting position 2 is far away, it is assumed that the angle of incidence (the direction of arrival) of each one incoming wave 3 with respect to each antenna 1 does not change. It is assumed that the phases at the time of incidence are different due to the mutual positional relationship between the antennas 1.

[0006] Each receiving section 4 sends out a received signal received by each antenna 1 to the MUSIC method operation processing section 5. MUS
The IC method calculation processing unit 5 estimates and calculates the number K of the incoming waves 3 and the arrival direction θ of each of the incoming waves 3 and displays them on the display 6.

Next, a detailed estimation method of the arrival direction (incident angle) of each arriving wave 3 in the MUSIC method operation processing section 5 will be described.

Each signal of K incoming waves 3 (arriving wave signal)
And u _i (t), and the arrival direction (incident angle) of the incident angle is θ _i (i = 1,
, K), and each signal (received signal) received by each receiving unit 4 via each of the M antennas 1 is assumed to be x _i (t) (i = 1, 2, 3, …, M). Further, let the noise included in each received signal x _i (t) be n _i (t) (i = 1, 2, 3,..., M).

[0009] Each received signal of each antenna 1 has K
Since each of the K arriving signal signals and the one irrelevant noise between the antennas are included, each received signal x _i (t)
Can be represented by a received signal vector x (t) expressed as a vector from i = 1 to i = M. (1) to (3)
It can be expressed by an expression.

Here, [] T indicates a transposed matrix. In this specification, vectors are represented by lowercase bold characters (x (t), u (t), n (t), a (θ), β).
, And the matrix is indicated by capital letters (A, E, P, R).

X (t) = [x ₁ (t),..., X _M (t)] ^T = M ^1/2 [a (θ ₁ ), a (θ ₂ ),..., A (θ _K )] u (t) + n (t ) ... (1) = Au (t) + n (t) ... (2) = A [u 1 (t), ..., u K (t)] T + [n 1 (t ),..., N _M (t)] ^T (3) where a (θ _i ) (M-dimension) is called a direction vector, and each antenna is referred to as a first antenna 1. This is a vector representing the phase difference, and is expressed by equation (4).

A (θ _i ) = M ^−1/2 [1, exp (−jkd cos θ _i ),..., Exp {(−jk (M−1) dc os θ _i ｝]] ^T (4) : Propagation constant (= 2π / λ) λ; wavelength In other words, the received signal vector is defined as each direction vector a (θ _i ) indicating the phase difference in each antenna 1 and each incoming signal u (t) indicating each incoming wave signal. The noise can be represented by n (t).

First, a correlation matrix R of the received signal vector x (t) shown by the equations (1) to (3) is obtained. In determining the correlation matrix R, it is assumed that there is no correlation between the incoming signal and the noise. Then, the correlation matrix R is obtained by the equations (5) to (7). Here, [] ^H indicates a conjugate transposed matrix. Also, E [
[] Indicates an expected value.

[0014] R = E [x (t) x H (t)] ... (5) = AE [u (t) u H (t)] A H + E [n (t) n H (t)] ... ( 6) = APA ^H + σ ² I (7) where σ ² is noise power.

Next, with respect to the correlation matrix R derived in this manner, M eigenvalues (λ ₁ ,
_{λ 2, ..., λ i,} ..., determine the λ _M).

Further, each eigenvalue (λ₁, Λ_Two,…,
λ_i, ..., λ_M) For the eigenvector (β₁, Β
_Two, ..., β_i, ..., β_M) Is calculated. That is,
Eigenvalue λ _iAnd the eigenvector β_iIs defined by equation (8).
It is.

Rβ _i = λ _i β _i (i = 1, 2, 3,..., M) (8) However, when the degree of the matrix becomes fifth or higher, a finite number of calculation operations using a computer The above-mentioned M eigenvalues (λ ₁ , λ ₂ ,..., Λ _i ,..., Λ _M ) cannot be obtained. Therefore, it is necessary to obtain these eigenvalues by an iterative method such as the Jacobi method or the QR method.

In this iterative method, the eigenvalues (λ ₁ , λ ₂ ,
, Λ _i ,..., Λ _M ) and then calculating the eigenvectors (β ₁ , β ₂ ,..., Β _i ,..., Β _M ) and the eigenvalues (λ ₁ , λ ₂ ,. λ _i , ..., λ _M )
And eigenvectors (β ₁ , β ₂ ,…, β _i ,…,
β _M ) at the same time.

The calculated eigenvalues (λ ₁ , λ ₂ ,..., Λ
_i ,..., λ _M ) are, for example, λ ₁ , λ ₂ ,.
Then, λ ₁ ≧ λ ₂ ≧,... ≧ λ _K ≧ λ _{K + 1} ≧ =,..., Λ _M = σ ² (9) In this case, all the eigenvalues corresponding to the noise can be regarded as equal, so that the number of eigenvalues λ larger than the noise power σ ² can be estimated as the number K of the arriving waves 3.

That is, (λ₁, Λ_Two, ..., λ_K )
It can be specified as the eigenvalue of the arriving wave 3, and (λ _{K + 1} , ..., λ_M=
σ^Two) Can be specified as the eigenvalue of noise.

Then, from the equations (7) and (8), the eigenvectors (β _{K + 1} ,..., Β _M ) corresponding to the eigenvalues of the noise (λ _{K + 1} ,..., Λ _M = σ ² ) are obtained. Equation (10) holds.

APA ^H β _i = 0 K + 1 ≦ i ≦ M (10) Equation (10) means that the direction vector a (θ _k ) of the signal and the eigenvector β _i for noise are orthogonal to each other. , (11).

Β _i ^Ha (θ _k ) = 0 K + 1 ≦ i ≦ M 1 ≧ k ≧ K (11) Therefore, the evaluation function P (θ) of the equation (12) is calculated.

[0024]

(Equation 1) That is, the evaluation function P (θ) shown in the equation (12) is obtained by adding the inner product value of the direction vector a (θ) of the signal shown in the equation (11) and the eigenvector β _i to the noise from the (K + 1) th to the Mth. And the reciprocal.

Therefore, as shown in FIG. 4, the evaluation function P (θ) calculated by the equation (12) is represented by the arrival direction (angle) on the horizontal axis.
In the MUSIC spectrum indicated as
By searching for the positions (peak positions) of the peak waveforms, the arrival directions (incident angles) θ ₁ , θ ₂ of each of the K arriving waves 3 are obtained.
…, Θ _K is obtained.

[0026]

However, using the MUSIC method of the above algorithm, the number of arriving waves K
In the radio wave arrival direction estimating apparatus for estimating each arrival direction (incident angle) θ, the following problems still remain.

That is, in the conventional MUSIC method, the eigenvalue λ of the correlation matrix R of the received signal vector x (t)
, The number K of arriving waves 3 is estimated, and as shown in Expression (12),
The arrival direction (incident angle) θ of each arriving wave 3 is estimated from the peak position of the MUSIC spectrum obtained by performing the inner product operation of the signal direction vector a (θ) and the plurality of noise eigenvectors β _i in each direction. are doing.

In this case, the number of arriving waves K is estimated by assuming that all the eigenvalues for each of the calculated eigenvalues corresponding to each of the incoming waves 3 and each of the noises are equal. Of course, it is desirable that all eigenvalues corresponding to each noise are ideally the same. But in reality,
There is considerable variation in the simulation results, and it is difficult to estimate the number of incoming waves without error when the number of incoming waves K is automatically estimated.

For this reason, if the number of arriving waves is estimated to be smaller than the actual number of arriving waves, the eigenvector corresponding to the eigenvalue of the arriving wave 3 is used for the evaluation function P (θ) in equation (12). There is a concern that the peak waveform of the arriving signal disappears and the measurement error in the arriving direction increases.

When the level difference among the plurality of arriving waves 3 is large, the peak value of the arriving direction θ of the arriving wave 3 having a small level on the MUSIC spectrum becomes small, and the arriving direction θ is accurately estimated. There are concerns that cannot be made.

The present invention has been made in view of such circumstances, and by using only one minimum eigenvalue, even if each eigenvalue corresponding to each noise has a variation, It is an object of the present invention to provide a radio wave arrival direction estimating device capable of estimating the number of arrival waves and the arrival direction of each arrival wave with high accuracy and high speed even when the level difference between the arrival waves is large.

[0032]

According to the present invention, a plurality of arriving waves arriving from each direction are received via a plurality of antennas arranged at different positions from each other, and the arriving signals are received from each antenna. The present invention is applied to a radio wave arrival direction estimating apparatus that estimates the number of incoming waves and the direction of arrival of each incoming wave.

In order to solve the above problem, in the radio wave direction-of-arrival estimation apparatus of the present invention, a correlation matrix calculating means for calculating a correlation matrix from each received signal, and one minimum eigenvalue from the calculated correlation matrix Calculating the minimum eigenvalue, calculating the eigenvector corresponding to the calculated minimum eigenvalue, calculating the eigenvector, and calculating the calculated eigenvector and the phase difference between the antennas depending on the direction of arrival. MUSIC spectrum calculating means for calculating a MUSIC spectrum by performing an inner product operation with a direction vector, and arrival of an incoming wave including a false wave corresponding to the corresponding peak waveform from the position of each peak waveform included in the calculated MUSIC spectrum Direction-of-arrival estimating means for estimating the direction; Power calculation means for calculating each power value of an incoming wave or a false wave corresponding to a peak waveform, and regarding each of the calculated power values, a peak waveform having a power value not less than a specified value is regarded as a peak waveform of the incoming wave. And an incoming wave number estimating means for estimating the number of incoming waves.

In the radio wave direction of arrival estimating apparatus thus configured, only one minimum eigenvalue of a plurality of eigenvalues respectively corresponding to a plurality of incoming wave signals and a plurality of noises which can be calculated from the correlation matrix is calculated. I do. This eigenvalue is always an eigenvalue for noise.

Then, one eigenvector corresponding to the one minimum eigenvalue is obtained, and the MUSIC spectrum is obtained using only the obtained one eigenvector.

However, since the inner product of the direction vector a (θ) of the signal and the eigenvector β _i for noise is not added from K + 1 to M, as shown by the conventional evaluation function of equation (12), In the present application, since the inner product of Expression (11) is not calculated over the eigenvectors corresponding to all the noises, the MUSIC spectrum has a plurality of false waveforms in addition to the respective peak waveforms corresponding to the original arriving waves. A peak waveform is generated.

Therefore, the power of the false wave is equal to the noise power, and each noise power is much smaller than the power of each arriving wave. Therefore, the power value of the arriving wave or the false wave corresponding to each peak waveform is calculated. Then, each peak waveform having a power value equal to or larger than a predetermined value is set as the peak waveform of the incoming wave, and the direction of arrival is obtained.

Therefore, even if the eigenvalue corresponding to the noise received together with each arriving signal greatly varies in each reception signal of each antenna, each arriving wave is reliably distinguished from each noise, and the number of arriving waves and each The direction of arrival of an incoming wave can be estimated.

[0039]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing a schematic configuration of a radio wave direction-of-arrival estimation apparatus according to an embodiment of the present invention. FIG.
The same parts as those of the conventional radio wave arrival direction estimating device shown in FIG. Therefore, the detailed description of the overlapping part will be omitted.

M antennas 1 are arranged on a measuring vehicle at equal intervals d from each other on a straight line. Then, from the distant K transmission positions 2, K radio waves are incident on the M antennas 1 as arriving waves 3. Since the transmitting position 2 is far away, it is assumed that the direction of arrival (incident angle) of one incoming wave 3 to each antenna 1 does not change. Then, it is assumed that the phases at the time of incidence are different due to the positional relationship between the antennas 1.

Each receiving section 4 sends out a received signal received by each antenna 1 to the MUSIC method operation processing section 7. MUS
The IC method calculation processing section 7 estimates and calculates the number K of the incoming waves 3 and the arrival direction θ of each of the incoming waves 3 and displays them on the display 6.

The MUSIC method operation processing unit 7 in the radio wave direction-of-arrival estimation apparatus of this embodiment is different from the MUSIC method operation processing unit 5 in the conventional apparatus shown in FIG. Minimum eigenvalue calculator 1
0, an eigenvector operation unit 11, a MUSIC spectrum calculation unit 12, an arrival direction estimation unit 13, a power operation unit 14, and an arrival wave number estimation unit 15.

Next, the operations of the respective units 8 to 15 constituting the MUSIC method operation processing unit 7 will be described in order. For the sake of simplicity, as in FIG. 3, each signal (arriving wave signal) of each of the K arriving waves 3 is represented by u _i (t) and the incident angle (direction of arrival).
Is _defined as θ _i (i = 1, 2, 3,..., K), and each signal (received signal) received by each receiver 4 via each of the M antennas 1 is represented by x
_i (t) (i = 1, 2, 3,..., M). Furthermore, let the noise included in each received signal x _i (t) be n _i (t) (i = 1,2,3,
…, M).

Each receiving unit 4 converts each received signal x _i (t) into A /
The data is D-converted and sent to the memory 8 of the MUSIC method operation processing unit 7. The memory 8 temporarily stores A / D converted digital received signals x _i (t).

The next correlation matrix calculator 9 calculates each direction vector a indicating the phase difference in each antenna 1 described above.
(Θ _i ), a correlation matrix R of the received signal vector x (t) of the equations (1) to (4) shown by an incoming signal u (t) indicating a signal of each incoming wave and each noise n (t). Is calculated using the above-described equations (5) to (7).

The minimum eigenvalue calculation unit 10 obtains one minimum eigenvalue λ _M from the _M eigenvalues, which is the number of antennas 1, in the correlation matrix R derived as described above. Therefore, the eigenvalue λ _M corresponding to the minimum noise is obtained.

When the order of the matrix is equal to or higher than five, the above-mentioned one eigenvalue λ _M cannot be obtained by a finite number of calculation operations using a computer. So, Jacobi method and Q
The eigenvalue is obtained by an iterative method such as the R method.

Next, the eigenvector operation unit 11 calculates one eigenvector β _M for one calculated minimum eigenvalue λ _M by using the correlation matrix R by the above-mentioned equation (8). In this way, one eigenvector β _M for the minimum noise is obtained.

The MUSIC spectrum calculator 12 calculates the one eigenvector β _M and the direction vector a
The evaluation function P expressed by the equation (13) in which the inner product value with (θ) is the reciprocal is
Calculate _M (θ).

[0051] _{P M (θ) = 1 /} | β M H a (θ) | in ² ... (13) the evaluation function P _M (theta), the evaluation function P of a conventional (12) (theta) As shown, the direction vector a of the signal
Since the inner product value of (θ) and the eigenvector β _i for noise is not added from K + 1 to M, in the present embodiment, the inner product of equation (11) is calculated over the eigenvectors corresponding to all noises. Since the calculation is not performed, as shown in FIG. 2, a plurality of peak waveforms 16 are generated as false waves in the MUSIC spectrum, in addition to the respective peak waveforms 16 corresponding to the original incoming waves 3.

The direction-of-arrival estimating unit 13 uses the MUS shown in FIG.
From the peak position of each peak waveform 16 in the IC spectrum, the arrival direction (incident angle) θ of the incoming wave 3 or the false wave corresponding to the relevant peak waveform 16 is obtained.

Next, the power calculator 14 calculates the received power value p _i of the arriving wave 3 or the false wave corresponding to each peak waveform 16 for which the arriving direction θ has been determined.

Specifically, the arriving wave 3 or the
Direction of arrival of false wave (incident angle) θ₁, Θ _Two…, Θ_KThan,
Calculate the matrix A (M × K) of equation (2) to obtain equation (14)
You.

A = M ^1/2 [a (θ ₁ ), a (θ ₂ ),..., A (θ _K )] (14) From the equation (7), the power matrix P (K order × K When the following is calculated, equations (15), (16) and (17) are obtained.

APA ^H = R−σ ² I (15) (A ^H A) P (A ^H A) = A ^H (R−σ ² I) A (16) P = (A ^H A) ⁻¹ A ^H (R−σ ² I) A (A ^H A) ⁻¹ (17) In the obtained power matrix P (K order × K order), the ith diagonal component is the arrival direction (incident angle). Arrival wave 3 of θ _i
Or it becomes the power value p _{i of the} false wave.

[0057] the arrival wave number estimating unit 15 is regarded as the peak waveform 16 of the incoming waves 3 each peak waveform 16 of the prescribed value p _s or more power values among the power values p _i of each peak waveform 16 which is calculated, The number of incoming waves K is estimated.

Then, the number of arriving waves M of the radio wave estimated by the arriving wave number estimating section 15 and the arriving direction θ of each arriving wave 3 are displayed on the display 6.

In the radio wave direction-of-arrival estimating apparatus thus formed, a plurality of eigenvalues λ respectively corresponding to a plurality of incoming waves and a plurality of noises which can be calculated from the correlation matrix R.
_{1, λ 2, ..., λ} i, ..., only one of the smallest eigenvalue lambda _M of lambda _M is calculated. This eigenvalue λ _M is always an eigenvalue for noise.

Then, one eigenvector β _M of one minimum eigenvalue λ _M corresponding to the noise is obtained,
The MUSIC spectrum shown in FIG. 2 is obtained by using only one obtained eigenvector β _M.

However, in the MUSIC spectrum, a plurality of peak waveforms 16 are generated as spurious waves, in addition to the peak waveforms 16 corresponding to the original incoming waves 3. Therefore, to calculate the power value p _i of the arrival wave 3 or sham wave corresponding to each peak waveform 15, each peak waveform 16 of the prescribed value p _s or more power values to the peak waveform of the incoming waves, obtains an arrival direction .

Therefore, since only the smallest eigenvalue of each eigenvalue corresponding to a number of noises having different values is used, for example, the received signal of each antenna 1 includes Even if the corresponding eigenvalue λ greatly varies, it is possible to reliably distinguish each arriving wave 3 from each false wave and estimate the number K of arriving waves and the arriving direction θ of each arriving wave.

Further, since only one minimum eigenvalue λ _M and the corresponding eigenvector β _M need be calculated, the amount of calculation is greatly reduced as compared with the conventional apparatus which calculates all the eigenvalues and the corresponding eigenvectors. it can.

Similarly, the evaluation function P _M (θ) shown in the expression (13) of the embodiment device is the evaluation function P _M (θ) shown in the expression (12) of the conventional device.
Compared with (θ), the processing speed is greatly simplified, so that the processing speed can be greatly improved.

The calculation of the power value p _i of each arriving wave 3 or false wave corresponding to each peak waveform 16 is performed by calculating each eigenvalue λ
The processing load is much smaller than that of the calculation calculation, so that even if the power calculation processing is added, the processing load of the entire apparatus does not increase so much, and the processing is performed at a higher speed than the conventional apparatus. be able to.

[0066]

As described above, in the radio wave direction of arrival estimating apparatus of the present invention, the MUSIC spectrum is calculated by using only one of the plurality of eigenvalues obtained from the correlation matrix of the received signal. I have.

Therefore, even if the eigenvalues corresponding to the respective noises vary, and even if the level difference between the incoming waves is large, the number of incoming waves and the direction of arrival of each incoming wave are changed. It can be estimated with high accuracy and at high speed.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of a radio wave direction of arrival estimation apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing a MUSIC spectrum calculated by the radio wave direction-of-arrival estimation apparatus according to the embodiment;

FIG. 3 is a block diagram showing a schematic configuration of a conventional radio wave direction-of-arrival estimation device;

FIG. 4 shows an M calculated by the conventional radio wave arrival direction estimating apparatus.
Diagram showing USIC spectrum

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Antenna 2 ... Transmission position 3 ... Arrival wave 4 ... Receiving part 5, 7 ... MUSIC method operation processing part 6 ... Display 8 ... Memory 9 ... Correlation matrix operation part 10 ... Minimum eigenvalue operation part 11 ... Eigenvector operation part 12 ... MUSIC spectrum calculation unit 13 ... arrival direction estimation unit 14 ... power calculation unit 15 ... arrival wave number estimation unit

Claims (1)

[Claims] 1. A plurality of arriving waves arriving from respective directions are received via a plurality of antennas arranged at mutually different positions, and the number of the arriving waves and the number of each arriving wave are determined based on a reception signal of each antenna. A radio wave arrival direction estimating apparatus for estimating a wave arrival direction, comprising: a correlation matrix calculating means (9) for calculating a correlation matrix from each of the received signals; and a minimum eigenvalue calculation for calculating one minimum eigenvalue from the calculated correlation matrix. Means (10), an eigenvector calculating means (11) for calculating one eigenvector corresponding to the calculated one eigenvalue, and a phase difference between the calculated one eigenvector and each of the antennas which differs depending on the direction of arrival. A MUSIC spectrum calculating means (12) for calculating an MUSIC spectrum by performing an inner product operation with the indicated directional vector; DOA estimation means for estimating the direction of arrival of an incoming wave including Nisenami corresponding to the relevant peak waveform from the position of the peak waveforms included in MUSIC spectrum was (1
3) and power calculation means (14) for calculating each power value of an incoming wave or a false wave corresponding to each peak waveform whose arrival direction is estimated.
And radio wave arrival estimation means (15) for estimating the number of arrival waves by regarding each peak waveform having a power value not less than a specified value among the calculated power values as the peak waveform of the arrival wave. Direction estimation device.