JP2020143918A - Incoming direction estimation device and incoming direction estimation method - Google Patents

Abstract

To provide an incoming direction estimation technology having a high accuracy of error correction.SOLUTION: An incoming direction estimation device includes an acquisition unit, a calculation unit, a correction unit, and an estimation unit. The acquisition unit acquires reception signals of virtual array antennas including a first and second virtual antennas that are generated by a combination of a plurality of transmission antennas and a plurality of reception antennas and have the same reciprocal path distance of the signal to an identical target. The calculation unit calculates a correlation matrix of the reception signals based on the reception signal for each of the plurality of virtual antennas including the first and second virtual antennas, and calculates an eigenvector based on the eigenvalue decomposition of the correlation matrix of the reception signals. The correction unit corrects the eigenvector based on elements of the eigenvector corresponding to the first and second virtual antennas. The estimation unit estimates an incoming direction of a radio wave based on the eigenvector corrected by the correction unit.SELECTED DRAWING: Figure 1

Description

The present invention relates to a technique for estimating the arrival direction of radio waves.

The radar device irradiates radio waves and receives radio waves (reflected waves) reflected from the target to estimate the direction of arrival of the reflected waves. The method of estimating the arrival direction is a method of calculating the arrival direction (angle) from information on the phase difference and amplitude difference of the received signals obtained by a plurality of receiving antennas that receive the reflected wave.

Japanese Unexamined Patent Publication No. 2011-33498

The number of antennas is very important in calculating the arrival direction (angle), and greatly affects the angle accuracy performance and the angle separation performance. However, an increase in the number of antennas leads to an increase in the cost of the radar device. Therefore, in recent years, attention has been paid to a technique for generating a virtual array antenna having a number of virtual antennas exceeding the number of receiving antennas by combining a plurality of transmitting antennas and a plurality of receiving antennas.

In the virtual array antenna, the error existing in one transmitting antenna is reflected in the error of a plurality of virtual antennas, and the range affected by the error becomes wide. For example, in the virtual array antenna VAx1 generated by the combination of the two transmitting antennas Tx1 and Tx2 having an antenna spacing of 4d and the four receiving antennas Rx1 to Rx4 having an antenna spacing of d as shown in FIG. 7, the transmitting antenna Tx1 The existing error affects the four virtual antennas VRx5 to VRx8.

Similarly, in the virtual array antenna, the error existing in one receiving antenna is reflected in the error of the plurality of virtual antennas, and the range affected by the error becomes wide. For example, in the virtual array antenna VAx1 generated by the combination of the four transmitting antennas Tx1 to Tx4 having an antenna spacing of d and the two receiving antennas Rx1 and Rx2 having an antenna spacing of 4d shown in FIG. 8, the receiving antenna Rx1 The existing error affects the four virtual antennas VRx1, VRx3, VRx5, and VRx7.

Therefore, a plurality of transmitting antennas and a plurality of receiving antennas are arranged so that a plurality of virtual antennas having the same signal round-trip distance for the same target can be obtained, and the signal round-trip distance for the same target is the same. A correction technique for correcting the received signal of the virtual antenna so as to cancel the difference between the received signals of the two virtual antennas has been proposed (see, for example, Patent Document 1).

However, random noise such as thermal noise is generated at each of the plurality of transmitting antennas and the plurality of receiving antennas. Therefore, the noise contained in each received signal of each virtual antenna is different. Therefore, in the above correction technique, the difference between the received signals of the two virtual antennas having the same round-trip distance of the signals for the same target is canceled while the received signals of the two virtual antennas contain different noises. However, the correct error correction is not performed. In particular, when the signal-to-noise ratio (SNR) is low, the influence of noise becomes large, so that the accuracy of error correction is significantly reduced.

In view of the above problems, an object of the present invention is to provide an arrival direction estimation technique with high accuracy of error correction.

The arrival direction estimation device according to the present invention is a first virtual antenna and a second virtual antenna that are generated by a combination of a plurality of transmitting antennas and a plurality of receiving antennas and have the same round-trip distance of signals for the same target. A correlation matrix of the received signal is obtained based on the received signal for each of a plurality of virtual antennas including the acquisition unit for acquiring the received signal of the virtual array antenna including the first virtual antenna and the second virtual antenna. A calculation unit that calculates and calculates an eigenvector based on the eigenvalue decomposition of the correlation matrix of the received signal, an element of the eigenvector corresponding to the first virtual antenna, and an element of the eigenvector corresponding to the second virtual antenna. Based on the above, the configuration includes a correction unit that corrects the eigenvector and an estimation unit that estimates the arrival direction of the radio wave based on the eigenvector after the correction by the correction unit (first configuration).

In the arrival direction estimation device of the first configuration, the estimation unit corresponds to the element corresponding to the first virtual antenna and the second virtual antenna in the eigenvector after being corrected by the correction unit. Even if it is a configuration (second configuration) in which a combined eigenvector is generated, a correlation matrix of the combined eigenvectors is generated, and the arrival direction of radio waves is estimated based on the correlation matrix of the combined eigenvectors. Good.

In the arrival direction estimation device of the second configuration, the calculation unit multiplies the feature matrix obtained by arranging the plurality of coupling eigenvectors and the complex conjugate transposed matrix of the feature matrix to obtain the correlation matrix of the coupling eigenvectors. It may be the configuration to be generated (third configuration).

In the arrival direction estimation device of the first configuration, the calculation unit is a single unit based on the received signal for each of the plurality of virtual antennas including both the first virtual antenna and the second virtual antenna. The correlation matrix of the received signal is calculated, and the correction unit corrects the element of the eigenvector corresponding to the virtual antenna generated by combining either the transmitting antenna or the receiving antenna which is the same as the second virtual antenna. It may be the configuration (fourth configuration).

In the arrival direction estimation device having the first to third configurations, the calculation unit includes the first correlation matrix based on the received signal for each of the plurality of virtual antennas including the first virtual antenna, and the second. The second correlation matrix based on the received signal for each of the plurality of virtual antennas including the virtual antenna is calculated as a separate correlation matrix of the received signals, and the first correlation matrix based on the eigenvalue decomposition of the first correlation matrix is calculated. The eigenvector and the second eigenvector based on the eigenvalue decomposition of the second correlation matrix are calculated, and the correction unit uses the element of the first eigenvector corresponding to the first virtual antenna and the second eigenvector. The second eigenvector is corrected based on the element of the second eigenvector corresponding to the virtual antenna of the above, and the estimation unit is the first eigenvector and the second eigenvector after being corrected by the correction unit. The configuration may be such that the arrival direction of the radio wave is estimated based on the above (fifth configuration).

In the arrival direction estimation device of the fifth configuration, the estimation unit combines the first eigenvector and the second eigenvector after being corrected by the correction unit to generate the combined eigenvector (). It may be the sixth configuration).

The arrival direction estimation method according to the present invention is generated by a combination of a plurality of transmitting antennas and a plurality of receiving antennas, and the first virtual antenna and the second virtual antenna having the same round-trip distance of signals for the same target. A correlation matrix of the received signal is obtained based on the acquisition step of acquiring the received signal of the virtual array antenna including the first virtual antenna and the received signal for each of a plurality of virtual antennas including the first virtual antenna and the second virtual antenna. The calculation step of calculating and calculating the eigenvector based on the eigenvalue decomposition of the correlation matrix of the received signal, the element of the eigenvector corresponding to the first virtual antenna, and the element of the eigenvector corresponding to the second virtual antenna. Based on the above, the configuration includes a correction step of correcting the eigenvector and an estimation step of estimating the arrival direction of the radio wave based on the eigenvector after being corrected by the correction step (seventh configuration).

According to the arrival direction estimation technique according to the present invention, the accuracy of error correction can be improved.

The figure which shows the structure of the radar apparatus which concerns on 1st and 2nd Embodiment The figure which shows the antenna arrangement in 1st and 2nd Embodiment A flowchart showing the schematic operation of the radar device according to the first embodiment. A flowchart showing the schematic operation of the radar device according to the second embodiment. Diagram showing a modified example of antenna arrangement Diagram showing other variants of antenna placement The figure which shows the further modification example of the antenna arrangement The figure which shows an example of the general MIMO antenna arrangement Diagram showing other examples of common MIMO antenna arrangements

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the drawings.

<1. First Embodiment>
<1-1. Radar device configuration>
FIG. 1 is a diagram showing a configuration of a radar device 1 according to the present embodiment. The radar device 1 is mounted on a vehicle such as an automobile. When the radar device 1 is mounted on the front end of the own vehicle, the radar device 1 acquires the target data related to the target existing in front of the own vehicle by using the transmitted wave. The target data includes the distance to the target, the relative speed of the target with respect to the radar device 1, and the like. However, in order to explain the radar device 1 according to the present embodiment as an example of the arrival direction estimation device, only the part related to the arrival direction estimation will be described in the following description.

As shown in FIG. 1, the radar device 1 mainly includes two transmitting units 2, a receiving unit 3, and a signal processing device 4. The radar device 1 is a so-called MIMO (Multi Input Multi Output) radar device.

The transmission unit 2 includes a signal generation unit 21 and a transmitter 22. The transmitter 22 modulates the signal generated by the signal generation unit 21 to generate a transmission signal. The two transmitting antennas 23 receive transmission signals from separate transmission units 2, convert the transmission signals into transmission wave TW, and output the transmission signals. The two transmission signals output from each of the two transmission units 2 are signals orthogonal to each other (orthogonal signals).

The receiving unit 3 includes a plurality of receiving antennas 31 and a plurality of individual receiving units 32 connected to the plurality of receiving antennas 31. In the present embodiment, the receiving unit 3 includes, for example, four receiving antennas 31 and four individual receiving units 32. The four individual receiving units 32 correspond to the four receiving antennas 31, respectively. Each of the four receiving antennas 31 corresponds to receiving channels ch1 to ch4. Each receiving antenna 31 receives the reflected wave RW from the object and acquires the received signal, and each individual receiving unit 32 processes the received signal obtained by the corresponding receiving antenna 31.

Each individual receiving unit 32 includes a mixer 33 and an A / D converter 34. The received signal obtained by the receiving antenna 31 is amplified by a low noise amplifier (not shown) and then sent to the mixer 33. A transmission signal from each transmitter 22 of each transmission unit 2 is input to the mixer 33, and each transmission signal and a reception signal are mixed in the mixer 33. As a result, a beat signal having a beat frequency that is the difference between the frequency of each transmission signal and the frequency of the reception signal is generated. The beat signal generated by the mixer 33 is converted into a digital signal by the A / D converter 34 and then output to the signal processing device 4.

The signal processing device 4 includes a microcomputer including a CPU (Central Processing Unit), a memory 41, and the like. The signal processing device 4 stores various data to be calculated in the memory 41 which is a storage device. The memory 41 is, for example, a RAM (Random Access Memory) or the like. The signal processing device 4 includes a transmission control unit 42, a Fourier transform unit 43, and a data processing unit 44 as functions realized by software in a microcomputer. The transmission control unit 42 controls each signal generation unit 21 of each transmission unit 2. The data processing unit 44 includes a peak extraction unit 45, an acquisition unit 46, a calculation unit 47, a correction unit 48, and an estimation unit 49.

Since the Fourier transform unit 43 receives the reflected waves from a plurality of targets in a superposed state on the receiving antenna 31, the frequency component based on the reflected waves of each target is obtained from the beat signal generated based on the received signal. (For example, FFT (Fast Fourier Transfer) processing) is performed. In the FFT process, the reception level and phase information are calculated for each frequency point (sometimes called a frequency bin) set at a predetermined frequency interval.

The peak extraction unit 45 detects a peak from the result of FFT processing or the like by the Fourier transform unit 43.

The acquisition unit 46 acquires the reception signal of the virtual array antenna based on the peak value extracted by the peak extraction unit 45. Details of the virtual array antenna will be described later.

The calculation unit 47 calculates the first eigenvector and the second eigenvector based on the received signal of the virtual array antenna. Details of the first eigenvector and the second eigenvector will be described later.

The correction unit 48 corrects the second eigenvector. A specific example of the correction will be described later.

The estimation unit 49 estimates the arrival direction of the radio wave based on the first eigenvector and the second eigenvector after being corrected by the correction unit 48. A specific example of estimation will be described later.

<1-2. Antenna placement>
FIG. 2 is a diagram showing an antenna arrangement in this embodiment. As shown in FIG. 2A, the two transmitting antennas 23 are arranged along the horizontal direction with an antenna spacing of 3d, and as shown in FIG. 2B, the four receiving antennas 31 are the same along the horizontal direction. It is arranged at the antenna spacing d of. The distance between the three antennas in the receiving antenna 31 does not have to be exactly the same, and it is sufficient that the distance between the three antennas can be regarded as the same in consideration of design errors and variations. Further, the antenna spacing in the transmitting antenna 23 does not have to be strictly three times the antenna spacing in the receiving antenna 31, and is regarded as three times the antenna spacing in the receiving antenna 31 in consideration of design errors and variations. I wish I could.

The virtual array antenna shown in FIG. 2C is generated by the combination of the two transmitting antennas 23 shown in FIG. 2A and the four receiving antennas 31 shown in FIG. 2B. The virtual array antenna shown in FIG. 2C is composed of eight virtual antennas VRx1 to VRx8.

By applying MIMO technology, it is possible to obtain a number of virtual antennas that exceeds the number of receiving antennas. The virtual antennas VRx1 to VRx4 are arranged along the horizontal direction at an antenna spacing d, and the virtual antennas VRx5 to VRx8 are arranged along the horizontal direction at an antenna spacing d. Then, the position of the virtual antenna VRx4 and the position of the virtual antenna VRx5 overlap. The virtual antenna VRx4 and the virtual antenna VRx5 having overlapping positions have the same round-trip distance of signals with respect to the same target.

The reception signal of the reception antenna 31 of 1ch includes the reception signal of the virtual antenna VRx1 and the reception signal of the virtual antenna VRx5 that are orthogonal to each other. Similarly, the received signal of the 2ch receiving antenna 31 includes the received signal of the virtual antenna VRx2 and the received signal of the virtual antenna VRx6 that are orthogonal to each other. Similarly, the reception signal of the reception antenna 31 of 3ch includes the reception signal of the virtual antenna VRx3 and the reception signal of the virtual antenna VRx7 that are orthogonal to each other. Similarly, the reception signal of the reception antenna 31 of 4ch includes the reception signal of the virtual antenna VRx4 and the reception signal of the virtual antenna VRx8 that are orthogonal to each other.

<1-3. Outline operation of radar device>
FIG. 3A is a flowchart showing a schematic operation of the radar device 1. The radar device 1 periodically repeats the process shown in FIG. 3A at regular time intervals.

First, the transmitting antenna 23 outputs the transmitted wave TW (step S10). Next, the receiving antenna 31 receives the received wave RW and acquires the received signal (step S20). Next, the signal processing device 4 acquires a predetermined number of beat signals (step S30). Next, the Fourier transform unit 43 executes an FFT operation on the beat signal (step S40).

Then, the peak extraction unit 45 extracts the peak from the result of the FFT calculation (step S50).

In step S60 following step S50, the acquisition unit 46 acquires the reception signal of the virtual array antenna shown in FIG. 2C based on the peak value extracted by the peak extraction unit 45. That is, the acquisition unit 46 acquires a received signal _x 1 ~x ₈ virtual antennas VRx1～VRx8.

In step S70 subsequent to step S60, calculation unit 47, based on the received signal _x 1 ~x ₈ virtual antennas VRx1～VRx8, calculates a first eigenvector and the second eigenvector.

Calculator 47, the first correlation matrix R1 based on the received signal _x 1 ~x ₄ virtual antennas VRx1～VRx4, a second correlation matrix R2 based on the received signal _x 5 ~x ₈ virtual antennas VRx5~VRx8 , Is calculated. The first correlation matrix R1 can be calculated by the following equation (1), and the second correlation matrix R2 can be calculated by the following equation (2). Here, an _{X1 = [x 1, x 2} , x 3, x 4] T, a _{X2 = [x 5, x 6} , x 7, x 8] T, E [·] the time average processing [・] ^H represents the complex conjugate translocation.
R1 = E [X1X1 ^H ] ... (1)
R2 = E [X2X2 ^H ] ・ ・ ・ (2)

The calculation unit 47 decomposes the first correlation matrix R1 and the second correlation matrix R2 into eigenvalues, respectively. Signal and noise can be separated by eigenvalue decomposition. Since the first correlation matrix R1 and the second correlation matrix R2 are matrices of 4 rows and 4 columns, four eigenvalues of the first correlation matrix R1 can be obtained, and four eigenvalues of the second correlation matrix R2 can also be obtained. ..

In the present embodiment, the calculation unit 47 calculates a first eigenvector corresponding to each specific eigenvalue of the first correlation matrix R1, and a second eigenvector corresponding to each specific eigenvalue of the second correlation matrix R2. Is calculated. The specific eigenvalue means an eigenvalue whose absolute value is equal to or greater than a predetermined value.

Hereinafter, among the eigenvalues of the first correlation matrix R1, only the eigenvalue Φ ₁₁ having the largest absolute value and the eigenvalue Φ ₁₂ having the second largest absolute value correspond to the specific eigenvalues, and the eigenvalues of the second correlation matrix R2 Of these, the case where only the eigenvalue Φ ₂₁ having the largest absolute value and the eigenvalue Φ ₂₂ having the second largest absolute value correspond to a specific eigenvalue will be described as an example.

Calculator 47, corresponding to the first eigenvector _{e 11,} first eigenvector _{e 12,} second eigenvector _{e 21} corresponding to the eigenvalues [Phi ₂₁ corresponding to the eigenvalues [Phi ₁₂ and eigenvalues [Phi _22, corresponding to the eigenvalue [Phi ₁₁ The second eigenvector e ₂₂ is calculated. The first eigenvectors e ₁₁ and e ₁₂ and the second eigenvectors e ₂₁ and e ₂₂ calculated by the calculation unit 47 are not affected by noise. The first eigenvectors e ₁₁ and e ₁₂ and the second eigenvectors e ₂₁ and e ₂₂ calculated by the calculation unit 47 are represented by the following equations (3) to (6).
e ₁₁ = [e ₁₁₁ , e ₁₁₂ , e ₁₁₃ , e ₁₁₄ ] ^T ... (3)
e ₁₂ = [e ₁₂₁ , e ₁₂₂ , e ₁₂₃ , e ₁₂₄ ] ^T ... (4)
e ₂₁ = [e ₂₁₁ , e ₂₁₂ , e ₂₁₃ , e ₂₁₄ ] ^T ... (5)
e ₂₂ = [e ₂₂₁ and e ₂₂₂ , e ₂₂₃ , e ₂₂₄ ] ^T ... (6)

In step S80 following step S70, the correction unit 48 corrects the second eigenvectors e ₂₁ and e ₂₂ .

If there is no error between the two transmitting antennas 23, the element e _{114 of} the first eigenvector e ₁₁ corresponding to the virtual antenna VRx4 and the element e _{211 of} the second eigenvector e ₂₁ corresponding to the virtual antenna VRx5 match. Should be. Accordingly, the correction unit 48, as the elements of the first eigenvector _{e 11} elements _{e 114} and second corrected corresponding to the virtual antenna VRx5 the eigenvectors e _'21 corresponding to the virtual antenna VRx4 match, the The eigenvector e ₂₁ of 2 is corrected.

Similarly, if there is no error between the two transmitting antennas 23, the element e _{124 of} the first eigenvector e ₁₂ corresponding to the virtual antenna VRx4 and the element e _{221 of} the second eigenvector e ₂₂ corresponding to the virtual antenna VRx5. Should match. Accordingly, the correction unit 48, as the elements of the first eigenvector elements _{e 12 e 124} and second corrected corresponding to the virtual antenna VRx5 eigenvectors e _'22 corresponding to the virtual antenna VRx4 match, the The eigenvector e ₂₂ of 2 is corrected.

The second eigenvector e _'21 and e' ₂₂ after correction is expressed by (7) - (8) below.
_{e '21 = (e 114 /} e 211) × [e 211, e 212, e 213, e 214] T ··· (7)
_{e '22 = (e 124 /} e 221) × [e 221, e 222, e 223, e 224] T ··· (8)

In step S90 following step S80, the estimation unit 49 generates a first eigenvector _{e 11} and the corrected second eigenvector e _'21 and coupling eigenvectors _{e C1} that combines the corrected first eigenvector _{e 12} generating a binding eigenvectors e _C2 coupled to the second eigenvector e _'22 after. In the above combination, matching elements are combined into one element.

The combined eigenvectors e _C1 and e _C2 are represented by the following equations (9) to (10).
e _C1 = [e ₁₁₁ , e ₁₁₂ , e ₁₁₃ , e ₁₁₄ , (e ₁₁₄ / e ₂₁₁ ) x e ₂₁₂ , (e ₁₁₄ / e ₂₁₁ ) x e ₂₁₃ , (e ₁₁₄ / e ₂₁₁ ) x e ₂₁₄ ] ^T ... (9)
e _C2 = [e ₁₂₁ , e ₁₂₂ , e ₁₂₃ , e ₁₂₄ , (e ₁₂₄ / e ₂₂₁ ) x e ₂₂₂ , (e ₁₂₄ / e ₂₂₁ ) x e ₂₂₃ , (e ₁₂₄ / e ₂₂₁ ) x e ₂₂₄ ] ^T ... (10)

Further, the estimation unit 49 multiplies the feature matrix A obtained by arranging the join eigenvector e _C1 and the join eigenvector e _C2 in order and the complex conjugate transpose matrix of the feature matrix A to generate the correlation matrix R _ev of the join eigenvector. .. The correlation matrix R _ev is represented by the following equation (11).
R _ev = AA ^H
_{R ev = [e C1 e C2} ] × [e C1 H e C2 H] ··· (11)

Finally, the estimation unit 49 calculates the angle spectrum based on the correlation matrix R _ev by using a known arrival direction estimation method such as MUSIC (Multiple Signal Classification) and ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques). , Estimate the direction of arrival of radio waves (angle of the target) based on the calculated angle spectrum.

By executing the operation shown in the flowchart of FIG. 3A described above, the signal and the noise can be separated, and the error correction can be performed without being affected by the noise. Therefore, the accuracy of error correction can be improved.

<2. Second embodiment>
Next, a second embodiment of the present invention will be described. The description of the portion that overlaps with the first embodiment will be omitted. In the second embodiment, the processing of the calculation unit 47, the correction unit 48, and the estimation unit 49 is different.

The calculation unit 47 calculates one correlation matrix based on the received signal of the virtual array antenna, and calculates the eigenvector from the correlation matrix.

The correction unit 48 corrects the eigenvector.

The estimation unit 49 estimates the arrival direction of the radio wave based on the eigenvector after the correction by the correction unit 48.

Hereinafter, the schematic operation of the radar device 1 according to the second embodiment will be described with reference to FIG. 3B. The second embodiment is different from the first embodiment in that the processes of steps S70'to S80'are executed instead of steps S70 to S80.

In step S70 'subsequent to step S60, calculation unit 47, based on the received signal _x 1 ~x ₈ virtual antennas VRx1～VRx8, calculates the eigenvectors.

Calculator 47 calculates a single correlation matrix R based on the received signal _x 1 ~x ₈ virtual antennas VRx1～VRx8. The correlation matrix R can be calculated by the following equation (12). Here, X = [x ₁ , x ₂ , x ₃ , x ₄ , x ₅ , x ₆ , x ₇ , x ₈ ] ^T.
R = E [XX ^H ] ・ ・ ・ (12)

The calculation unit 47 decomposes the correlation matrix R into eigenvalues. Signal and noise can be separated by eigenvalue decomposition. Since the correlation matrix R is a matrix with 8 rows and 8 columns, eight eigenvalues of the correlation matrix R can be obtained.

In the present embodiment, the calculation unit 47 calculates the eigenvectors corresponding to each specific eigenvalue of the correlation matrix R. The specific eigenvalue means an eigenvalue whose absolute value is equal to or greater than a predetermined value.

Hereinafter, among the eigenvalues of the correlation matrix R, a description will be an example in which only the most absolute value is larger eigenvalues [Phi ₁ and the absolute value to the second large eigenvalue [Phi ₂ corresponds to a specific eigenvalue.

The calculation unit 47 calculates the eigenvector e ₁ corresponding to the eigenvalue Φ ₁ and the eigenvector e ₂ corresponding to the eigenvalue Φ ₂ . The eigenvectors e ₁ and e ₂ calculated by the calculation unit 47 are not affected by noise. The eigenvectors e ₁ and e ₂ calculated by the calculation unit 47 are represented by the following equations (13) to (14).
e ₁ = [e ₁₁ , e ₁₂ , e ₁₃ , e ₁₄ , e ₁₅ , e ₁₆ , e ₁₇ , e ₁₈ ] ^T ... (13)
e ₂ = [e ₂₁ , e ₂₂ , e ₂₃ , e ₂₄ , e ₂₅ , e ₂₆ , e ₂₇ , e ₂₈ ] ^T ... (14)

In step S80'following step S70', the correction unit 48 corrects the eigenvectors e ₁ and e ₂ .

If if no error between the two transmitting antennas 23, should match the element _{e 15} of eigenvectors _{e 1} corresponding to eigenvectors _{e 1} element _{e 14} and virtual antennas VRx5 corresponding to the virtual antenna VRx4. In other words, each element _e 15 _{to e 18} of eigenvectors _{e 1} corresponding to the virtual antenna VRx5~VRx8 a virtual antennas are generated using the same transmitting antenna as the virtual antenna VRx5 _is the difference between _{e 14} and _{e 15} Has a corresponding error. Accordingly, the correction unit 48, as the eigenvectors e _'1 element after correction corresponding to element _{e 14} and virtual antennas VRx5 eigenvectors _{e 1} corresponding to the virtual antenna VRx4 matches, it corrects the eigenvectors _{e 1.} In particular, each element e _{15 to} e ₁₈ of the eigenvector e ₁ corresponding to the virtual antennas VRx5 to VRx8 is corrected.

Similarly, if there is no error between the two transmitting antennas 23, it should match the eigenvectors _{e 2} elements _{e 25} corresponding to the virtual antenna VRx5 an element _{e 24} eigenvector _{e 2} corresponding to the virtual antenna VRx4 .. Accordingly, the correction unit 48, as the eigenvector e _'2 elements after correction corresponding to the element _{e 24} and virtual antennas VRx5 eigenvector _{e 2} corresponding to the virtual antenna VRx4 matches, corrects the eigenvectors _{e 2.}

Eigenvectors e _'1 and e' ₂ after correction is expressed by (14) - (17) below.
k ₁ = (e ₁₄ / e ₁₅ ) ・ ・ ・ (14)
k ₂ = (e ₂₄ / e ₂₅ ) ・ ・ ・ (15)
^{_{e '1 = e 1 T ×}} [1, 1, 1, 1, k 1, k 1, k 1, k 1] ··· (16)
^{_{e '2 = e 2 T ×}} [1, 1, 1, 1, k 2, k 2, k 2, k 2] ··· (17)

In step S90 following step S80', the estimation unit 49 generates the coupling eigenvector e _C1 . Binding eigenvectors _{e C1,} in the eigenvector e _'1 corrected a vector summarizing the elements corresponding to the virtual antenna VRx4 and virtual antennas VRx5 one. This is because the correction unit 48 made corrections so that both elements match. Similarly, the estimation unit 49 also generates the coupling eigenvector e _C2 . Binding eigenvectors _{e C2,} in the eigenvector e _'2 after correction, which is a vector summarizing the elements corresponding to the virtual antenna VRx4 and virtual antennas VRx5 one.

The combined eigenvectors e _C1 and e _C2 are represented by the following equations (18) to (19).
e _C1 = [e ₁₁ , e ₁₂ , e ₁₃ , e ₁₄ , k ₁ x e ₁₆ , k ₁ x e ₁₇ , k ₁ x e ₁₈ ] ^T ... (18)
e _C2 = [e ₂₁ , e ₂₂ , e ₂₃ , e ₂₄ , k ₂ x e ₂₆ , k ₂ x e ₂₇ , k ₂ x e ₂₈ ] ^T ... (19)

Further, the estimation unit 49 multiplies the feature matrix A obtained by arranging the join eigenvector e _C1 and the join eigenvector e _C2 in order and the complex conjugate transpose matrix of the feature matrix A to generate the correlation matrix R _ev of the join eigenvector. .. The correlation matrix R _ev is represented by the following equation (20).
R _ev = AA ^H
_{R ev = [e C1 e C2} ] × [e C1 H e C2 H] ··· (20)

Finally, the estimation unit 49 calculates the angle spectrum based on the correlation matrix R _ev by using a known arrival direction estimation method such as MUSIC (Multiple Signal Classification) and ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques). , Estimate the direction of arrival of radio waves (angle of the target) based on the calculated angle spectrum.

The radar device 1 according to the present embodiment exhibits a unique effect that cannot be achieved by the radar device 1 according to the first embodiment.

In the radar device according to the first embodiment, there is a possibility that uncertainty of combination may occur when the eigenvalues have the same magnitude. For example, consider the situation of Φ ₁₁ ≈ Φ ₁₂ and Φ ₂₁ ≈ Φ ₂₂ . At this time, the eigenvectors corresponding to Φ ₁₁ and Φ ₂₁ and the eigenvectors corresponding to Φ ₁₂ and Φ ₂₂ are the correct combinations. However, due to the influence of noise or the like, there is a possibility that the eigenvectors corresponding to Φ ₁₁ and Φ ₂₂ and the eigenvectors corresponding to Φ ₁₂ and Φ ₂₁ are mistakenly combined. If the combination of the eigenvectors is incorrect in this way, the estimation unit 49 cannot calculate the correct angle.

Such a condition can occur, for example, when there are similar targets at the same distance and velocity but at different angles. For example, when vehicles of the same model run in parallel in the same position in the lanes on both sides.

On the other hand, in the radar device 1 according to the present embodiment, since a single correlation matrix is generated using the received signals of all the virtual antennas, there is no possibility that the uncertainty of the combination will occur in principle. Therefore, the radar device 1 according to the present embodiment can stably estimate the correct angle even in the above-mentioned state.

<3. Others>
The various technical features disclosed herein can be modified in addition to the above embodiments without departing from the gist of the technical creation. In addition, a plurality of embodiments and modifications shown in the present specification may be combined and implemented to the extent possible.

In the above-described embodiment, the error between the transmitting antennas is corrected, but when the radar device has a hardware configuration in which an error is unlikely to occur between the transmitting antennas and an error is likely to occur between the receiving antennas, the error between the receiving antennas is likely to occur. You may try to correct the error of.

When correcting the error between the receiving antennas, for example, as shown in FIG. 4 (a), four transmitting antennas 23 are arranged along the horizontal direction with the same antenna spacing d, and in FIG. 4 (b). As shown, the two receiving antennas 31 may be arranged along the horizontal direction with an antenna spacing of 3d. The calculation unit 47, a virtual antenna VRx1, VRx3, VRx5, and a reception signal _{x 1,} x 3, _{x 5,} and the first correlation matrix R1 based on _{x 7} of VRx7, virtual antennas VRx2, VRx4, VRx6, and a reception signal _{x 2, x 4, x 6} , and the second correlation matrix R2 based on _{x 8} of VRx8, may be calculated.

In this case, the reception signal of the reception antenna 31 of 1ch includes the reception signal of the virtual antenna VRx1 orthogonal to each other, the reception signal of the virtual antenna VRx3, the virtual antenna VRx5, and the virtual antenna VRx7. Similarly, the received signal of the 2ch receiving antenna 31 includes the received signal of the virtual antenna VRx2 orthogonal to each other, the virtual antenna VRx4, the virtual antenna VRx6, and the received signal of the virtual antenna VRx8.

Further, in the above-described embodiment, the position of the virtual antenna VRx4 and the position of the virtual antenna VRx5 only overlap, but a plurality of sets of virtual antennas having overlapping positions may be used. For example, in the case of the antenna arrangement shown in FIG. 5, the position of the virtual antenna VRx3 and the position of the virtual antenna VRx5 overlap, and the position of the virtual antenna VRx4 and the position of the virtual antenna VRx6 overlap. In this case, for example, the first eigenvector _{e 11} second eigenvector e _'21 element and a first correction to match the post elements _{e 113} and the correction corresponding to the virtual antenna VRx5 that correspond to virtual antennas VRx3 determine the amount, the second correction amount as an element _{e 114} and elements of the second eigenvector e _'21 after correction corresponding to the virtual antenna VRx6 matches the first eigenvector _{e 11} corresponding to the virtual antenna VRx4 Is obtained, and the final correction amount of the second eigenvector e ₂₁ can be determined based on the first correction amount and the second correction amount. For example, the simple average of the first correction amount and the second correction amount may be used as the final correction amount, and the weighted average of the first correction amount and the second correction amount in consideration of the characteristics of each receiving antenna may be used. It may be the final correction amount. The same applies to the final correction amount of the second eigenvector e ₂₂ .

Further, by providing three or more transmitting antennas, a plurality of sets of virtual antennas having overlapping positions may be provided. For example, in the case of the antenna arrangement shown in FIG. 6, the position of the virtual antenna VRx4 and the position of the virtual antenna VRx5 overlap, and the position of the virtual antenna VRx8 and the position of the virtual antenna VRx9 overlap. In this case, for example, calculation unit 47 also calculates a third correlation matrix of R3 based on the received signal _x 9 _{~x 12} virtual antennas VRx9～VRx12, third eigenvectors third correlation matrix of R3 and eigenvalue decomposition e ₃₁ and e ₃₂ are also calculated. Then, a first correction amount, such as the elements of the first eigenvector _{e 11} elements _{e 114} and second corrected corresponding to the virtual antenna VRx5 the eigenvectors e _'21 corresponding to the virtual antenna VRx4 matches obtains a second correction amount as the third eigenvector e _'31 elements after correction corresponding to the element _{e 118} and virtual antenna VRx9 the first eigenvector _{e 11} corresponding to the virtual antenna VRx8 match, The third eigenvector e ₃₁ can be corrected by the correction amount obtained by multiplying the first correction amount and the second correction amount. The same applies to the correction of the third eigenvector e _32.

Further, although the in-vehicle radar device has been described in the above-described embodiment, the present invention can be applied to an infrastructure radar device, a ship surveillance radar device, an aircraft surveillance radar device, and the like installed on a road or the like.

Further, in the above-described embodiment, the estimation unit 49 multiplies the feature matrix A obtained by arranging the coupling eigenvector e _C1 and the coupling eigenvector e _C2 in order by the complex conjugate transposed matrix of the feature matrix A, and correlates the coupling eigenvectors. The matrix R _ev was generated. This makes it possible to improve the accuracy of the angle estimation of the targets even when the two targets are present at very similar angles.

For example, in the case of a radar device such as a ship monitoring radar device in which a scene in which two targets exist at very similar angles is not assumed, the estimation unit 49 of the coupling eigenvector e _C1 and the coupling eigenvector e _C1 . multiplied by the complex conjugate transposed vector to generate the correlation matrix _{R ev1} binding eigenvectors _{e C1,} generates a correlation matrix _{R ev2} binding eigenvectors _{e C2} by multiplying the complex conjugate transposed vector of binding eigenvectors _{e C2} and binding eigenvectors _{e C2} Then, the angle spectrum may be calculated based on the correlation matrices R _ev1 and R _ev2, respectively, and the arrival direction of the radio wave (the angle of the target) may be estimated based on the calculated angle spectrum.

1 Radar device 2 Transmitter 23 Transmitter antenna 3 Receiver 31 Receiver antenna 4 Signal processing device 46 Acquisition unit 47 Calculation unit 48 Correction unit 49 Estimator unit VRx1 to VRx12 Virtual antenna

Claims (8)

Acquires the received signal of a virtual array antenna including a first virtual antenna and a second virtual antenna, which are generated by a combination of a plurality of transmitting antennas and a plurality of receiving antennas and have the same round-trip distance of signals for the same target. Acquisition department and
The correlation matrix of the received signal is calculated based on the received signal for each of the plurality of virtual antennas including the first virtual antenna and the second virtual antenna, and the eigenvector is calculated based on the eigenvalue decomposition of the correlation matrix of the received signal. And the calculation unit that calculates
A correction unit that corrects the eigenvector based on the element of the eigenvector corresponding to the first virtual antenna and the element of the eigenvector corresponding to the second virtual antenna.
An estimation unit that estimates the arrival direction of radio waves based on the eigenvector after correction by the correction unit, and an estimation unit.
A device for estimating the direction of arrival. The estimation unit
In the eigenvector after being corrected by the correction unit,
A coupling eigenvector that summarizes the element corresponding to the first virtual antenna and the element corresponding to the second virtual antenna is generated.
A correlation matrix of the combined eigenvectors is generated.
The direction of arrival of radio waves is estimated based on the correlation matrix of the combined eigenvectors.
The arrival direction estimation device according to claim 1. The calculation unit
A correlation matrix of the combined eigenvectors is generated by multiplying a feature matrix obtained by arranging a plurality of the combined eigenvectors with a complex conjugate transposed matrix of the feature matrix.
The arrival direction estimation device according to claim 2. The calculation unit
A single correlation matrix of the received signals is calculated based on the received signals for each of the plurality of virtual antennas including both the first virtual antenna and the second virtual antenna.
The correction unit corrects the element of the eigenvector corresponding to the virtual antenna generated by combining either the transmitting antenna or the receiving antenna, which is the same as the second virtual antenna.
The arrival direction estimation device according to claim 1. The calculation unit
A first correlation matrix based on the received signal for each of the plurality of virtual antennas including the first virtual antenna, and a second correlation matrix based on the received signal for each of the plurality of virtual antennas including the second virtual antenna. And are calculated as separate correlation matrices for the received signals.
The first eigenvector based on the eigenvalue decomposition of the first correlation matrix,
A second eigenvector based on the eigenvalue decomposition of the second correlation matrix is calculated.
The correction unit
The second eigenvector is corrected based on the element of the first eigenvector corresponding to the first virtual antenna and the element of the second eigenvector corresponding to the second virtual antenna.
The estimation unit
The arrival direction of the radio wave is estimated based on the first eigenvector and the second eigenvector after being corrected by the correction unit.
The arrival direction estimation device according to claim 1. The calculation unit
A first correlation matrix based on the received signal for each of the plurality of virtual antennas including the first virtual antenna, and a second correlation matrix based on the received signal for each of the plurality of virtual antennas including the second virtual antenna. And are calculated as separate correlation matrices for the received signals.
The first eigenvector based on the eigenvalue decomposition of the first correlation matrix,
A second eigenvector based on the eigenvalue decomposition of the second correlation matrix is calculated.
The correction unit
The second eigenvector is corrected based on the element of the first eigenvector corresponding to the first virtual antenna and the element of the second eigenvector corresponding to the second virtual antenna.
The estimation unit
The arrival direction of the radio wave is estimated based on the first eigenvector and the second eigenvector after being corrected by the correction unit.
The arrival direction estimation device according to claim 2 or 3. The estimation unit
The first eigenvector and the second eigenvector after being corrected by the correction unit are combined to generate the combined eigenvector.
The arrival direction estimation device according to claim 6. Acquires the reception signal of a virtual array antenna including a first virtual antenna and a second virtual antenna, which are generated by a combination of a plurality of transmitting antennas and a plurality of receiving antennas and have the same round-trip distance of signals for the same target. Acquisition process and
The correlation matrix of the received signal is calculated based on the received signal for each of the plurality of virtual antennas including the first virtual antenna and the second virtual antenna, and the eigenvector is calculated based on the eigenvalue decomposition of the correlation matrix of the received signal. And the calculation process to calculate
A correction step of correcting the eigenvectors based on the elements of the eigenvectors corresponding to the first virtual antenna and the elements of the eigenvectors corresponding to the second virtual antenna.
An estimation step of estimating the arrival direction of radio waves based on the eigenvector after correction by the correction step, and an estimation step.
A method of estimating the direction of arrival.

Images