JP2020165811A - Position estimating device and position estimating method - Google Patents

Abstract

To provide a position estimating device which heightens the separability of a plurality of targets by an FFT process and heightens the accuracy of distance estimation by a 2-frequency comparison process.SOLUTION: A position estimating device 100 comprises: transmit units 110, 111 for transmitting a signal wave changed to two frequencies with a prescribed frequency difference; a receive unit 121 for receiving a reflected wave of the signal wave having been reflected by a plurality of targets and generating a beat signal derived by frequency-converting the reflected wave with the frequency of the signal wave; a first signal processing unit 125 for generating, for each of beat signals corresponding to two frequencies, separate beat signals with which phase information is retained and the azimuth angles or distances of the plurality of targets are separated from each other; and a second signal processing unit 126 for estimating the distance to a target as the position of the target from a prescribed frequency difference and a phase difference of separate beat signals corresponding to two frequencies.SELECTED DRAWING: Figure 1

Description

The present invention relates to a position estimation device and a position estimation method for estimating the position of a target (for example, the azimuth angle of the target and the distance to the target).

Patent Document 1 describes a two-frequency comparison type radar device. In this radar device, the distance to the target is calculated from the phase difference of the power spectrum of the beat signal. Patent Document 2 describes a radar device that combines a two-frequency comparison method and an FMCW method. In this radar device, the distance to the target is calculated from the phase difference of the beat signals corresponding to the two frequencies.

JP-A-2009-042061 Japanese Unexamined Patent Publication No. 2011-232115

In the two-frequency comparison method, the distance can be accurately obtained, but it is necessary to have one target. On the other hand, in the distance estimation by FFT, the distances of a plurality of targets can be obtained, but the distance estimation accuracy is low. Also, there must be a sufficient distance difference.

Regarding these points, the radar device described in Patent Document 1 can deal with a plurality of targets by FFT processing, and can obtain an accurate distance by processing by bifrequency comparison for the targets separated by FFT processing. That is, in the radar device described in Patent Document 1, the distances of a plurality of targets can be accurately obtained. However, it may not be possible to separate a plurality of targets even with the FFT process, and in this case, the two-frequency comparison process cannot be applied to the targets that cannot be separated by the FFT process.

An object of the present invention is to provide a position detection device and a position estimation device that enhance the separation ability of a plurality of targets by FFT processing and enhance the distance estimation accuracy by two-frequency comparison processing.

The position estimation device of the present invention is a position estimation device that estimates the positions of a plurality of targets, and includes a transmission unit that transmits signal waves changed to two frequencies by a predetermined frequency difference, and the plurality of signal waves. For each of the receiving unit that receives the reflected wave reflected by the target of the above and generates a beat signal in which the reflected wave is frequency-converted at the frequency of the signal wave, and the beat signal corresponding to the two frequencies. A first signal processing unit that generates a separated beat signal in which the azimuth angles or distances of the plurality of targets are separated from each other while maintaining phase information, the predetermined frequency difference, and the separation corresponding to the two frequencies. It includes a second signal processing unit that estimates the distance from the target as the position of the target from the phase difference of the beat signal.

The position estimation method of the present invention is a position estimation method for estimating the positions of a plurality of targets, in which a signal wave changed into two frequencies by a predetermined frequency difference is transmitted, and the signal wave is generated by the plurality of targets. The reflected reflected wave was received, the reflected wave was frequency-converted at the frequency of the signal wave to generate a beat signal, and phase information was maintained for each of the beat signals corresponding to the two frequencies. , The separated beat signals in which the azimuth angles or distances of the plurality of targets are separated from each other are generated, and the position of the target is determined from the predetermined frequency difference and the phase difference of the separated beat signals corresponding to the two frequencies. Estimate the distance to the target.

According to the present invention, the separability of a plurality of targets by the FFT processing can be enhanced, and the distance estimation accuracy by the two-frequency comparison processing can be improved.

It is a figure which shows the structure of the radar apparatus (position estimation apparatus) which concerns on this embodiment. An example of the target position measurement result (radar chart: XY Cartesian coordinate system) based on the beat signals corresponding to the three frequencies, and the target position measurement result (radar chart: XY Cartesian coordinate system) based on the beat signals corresponding to the two frequencies. It is a figure which shows an example of a coordinate system). It is a figure which shows the vector of the beat signal and the difference vector corresponding to three frequencies in FIG. It is a figure which shows the intensity (upper side) of the difference vector when the target is located at the distance 2m and 6m, and the phase φ and the phase difference Δφ (lower side) of the difference vector. It is a figure which shows the intensity (upper side) of the difference vector when the target is located at the distance 2m and 4m, and the phase φ and the phase difference Δφ (lower side) of the difference vector. It is a figure which shows an example of the measurement result (radar chart: XY Cartesian coordinate system) of the position of a target based on a beat signal corresponding to two frequencies. It is a figure for demonstrating the analysis (procedure 3-1) of the measured 4 antenna data by the 1st processing part in the radar apparatus (position estimation apparatus) shown in FIG. It is a figure for demonstrating the analysis (procedure 3-1) of the measured 4 antenna data by the 1st processing part in the radar apparatus (position estimation apparatus) shown in FIG. It is a figure for demonstrating the analysis (procedure 3-1) of the measured 4 antenna data by the 1st processing part in the radar apparatus (position estimation apparatus) shown in FIG. It is a figure for demonstrating the analysis (procedure 3-1) of the measured 4 antenna data by the 1st processing part in the radar apparatus (position estimation apparatus) shown in FIG. It is a figure for demonstrating the analysis (procedure 3-1) of the measured 4 antenna data by the 1st processing part in the radar apparatus (position estimation apparatus) shown in FIG. It is a figure which shows an example of a synthetic signal x _m '. It is a figure which shows an example of the estimation function Y'. It is a figure which shows an example of the estimation result (radar chart: XY orthogonal coordinate system) of a distance. It is a flowchart which shows the position estimation process by the radar apparatus (position estimation apparatus) shown in FIG. It is a flowchart which shows the azimuth angle high separation ability process (DOA1) in the position estimation process shown in FIG. It is a figure which shows an example of the estimation result (radar chart: XY Cartesian coordinate system) of the distance of the comparative example. It is a flowchart of the azimuth angle high separation ability processing (DOA2) in the position estimation processing by the radar apparatus (position estimation apparatus) which concerns on modification 1. It is a figure which shows the calculation result of the estimation function Y of the modification 1. It is a figure which shows the calculation result of the estimation function Y of the modification 1. It is a flowchart of the azimuth angle high separation ability processing (DOA3) in the position estimation processing by the radar apparatus (position estimation apparatus) which concerns on modification 2. FIG. It is a figure for demonstrating the analysis (procedure 1) of the measured 4 antenna data by the 1st processing part in the radar apparatus (position estimation apparatus) which concerns on modification 2. FIG. It is a figure which shows the calculation result of the composite signal _xn (k)', and the estimation function Y'of the modification 2. It is a figure for demonstrating an example of distance estimation which concerns on modification 3 (FMCW method). It is a figure for demonstrating another example of distance estimation which concerns on modification 3 (pulse method). It is a figure for demonstrating an example of the high resolution processing of the azimuth angle (or the distance) in the position estimation processing by the radar apparatus (position estimation apparatus) 100 which concerns on modification 4. FIG. It is a figure for demonstrating an example of the high resolution processing of the azimuth angle (or the distance) in the position estimation processing by the radar apparatus (position estimation apparatus) 100 which concerns on modification 4. FIG.

Hereinafter, an example of the embodiment of the present invention will be described with reference to the accompanying drawings. In addition, the same reference numerals are given to the same or corresponding parts in each drawing.

(Radar device)
FIG. 1 is a diagram showing a configuration of a radar device (position estimation device) according to the present embodiment. The radar device 100 shown in FIG. 1 is a device used for various in-vehicle radars such as an in-vehicle millimeter-wave radar in a 76 GHz band or a 79 GHz band, or an in-vehicle radar in a 24 GHz band.

The radar device 100 transmits a signal wave from the transmitting antenna Tx, and the reflected wave reflected by the target is received by the plurality of receiving antennas Rx. The radar device 100 detects (estimates) the position of the target (distance to the target, the azimuth angle of the target) and the speed (relative speed) of the target based on the received signals received by the plurality of receiving antennas Rx. The radar device 100 includes a control unit 101, a transmission control unit 110, a transmission unit 111, a reception signal processing unit 120, a reception unit 121, one transmission antenna Tx, and a plurality of reception antennas Rx.

The control unit 101, the transmission control unit 110, and the reception signal processing unit 120 are composed of, for example, arithmetic processors such as a DSP (Digital Signal Processor) and an FPGA (Field-Programmable Gate Array). Various functions of the control unit 101, the transmission control unit 110, and the reception signal processing unit 120 are realized, for example, by executing predetermined software (program, application) stored in the storage unit. Various functions of the control unit 101, the transmission control unit 110, and the reception signal processing unit 120 may be realized by the cooperation of hardware and software, or may be realized only by hardware (electronic circuit).

The control unit 101 controls the operation of the entire radar device 100.
The transmission control unit 110 controls the transmission unit 111.
The transmission unit 111 includes a VCO (Voltage-controlled oscillator), a modulator, an amplifier, and the like. The VCO generates a predetermined high frequency signal based on the frequency control signal from the transmission control unit 110. The modulator modulates the high frequency signal from the VCO with the modulated signal from the transmission control unit 110 to generate a transmission signal. The generated transmission signal is transmitted via the amplifier and the transmission antenna Tx. The transmission unit 111 and the transmission control unit 110 constitute the transmission unit described in the claims.

The receiving unit 121 includes an amplifier, an IQ mixer, a filter, an ADC and the like. The receiving unit 121 receives the received signal via the receiving antenna Rx and the amplifier.
The IQ mixer generates a beat signal (IF signal) obtained by frequency-converting a received signal (reflected wave) with the frequency of a transmission signal (signal wave). More specifically, the IQ mixer detects the received signal by orthogonal phase detection with a high frequency signal from the VCO, and outputs an I signal having the same phase and a Q signal having a phase shift of 90 degrees (π / 2).
The ADC performs analog-to-digital conversion by sampling the I signal and Q signal (beat signal) input through the filter at a predetermined sampling frequency. The ADC outputs the I signal and Q signal (beat signal) converted into digital signals to the reception signal processing unit 120.

The reception signal processing unit 120 detects (estimates) the position of the target (distance to the target, the orientation angle of the target), and the speed (relative speed) of the target based on the I signal and the Q signal (beat signal: digital signal). To do.
Hereinafter, the estimation of the target position by the received signal processing unit 120 will be described. The reception signal processing unit 120 includes a first signal processing unit 125 and a second signal processing unit 126 for estimating the position of the target.

<Procedure 1>
First, the transmission control unit 110 and the transmission unit 111 sequentially transmit transmission signals (signal waves) changed into three frequencies with a predetermined frequency difference Δf. For example, the transmission control unit 110 and the transmission unit 111 shift the center frequency of the transmission signal (signal wave) at predetermined frequency intervals (for example, at intervals of several tens of MHz). The receiving unit 121 sequentially receives the reflected wave whose transmission signal is reflected by the target as a reception signal. The receiving unit 121 frequency-converts the received signal with the frequency of the transmission signal to generate an analog-digitally converted beat signal (digital signal).

(A) to (c) of FIG. 2 show an example of the measurement result (radar chart: XY orthogonal coordinate system) of the position of the target based on the beat signals corresponding to the three frequencies. In the example of FIG. 2, a predetermined frequency difference is Δf, the center frequency of the first transmission signal is f1, the center frequency of the second transmission signal is f2 = f1 + Δf, and the third transmission signal. The center frequency of is f3 = f2 + Δf.

<Procedure 2>
Next, the first signal processing unit 125 generates a beat signal corresponding to two frequencies having a predetermined frequency difference Δf by the difference processing of the beat signals corresponding to the three frequencies (frequency shift method Ver.1). ..

(D) and (e) of FIG. 2 show an example of the measurement result (radar chart: XY Cartesian coordinate system) of the position of the target based on the beat signals corresponding to the two frequencies. The example of FIG. 2D is, for example, the measurement result of the position of the target based on the beat signal of the difference between the beat signal corresponding to the center frequency f1 and the beat signal corresponding to the center frequency f2. The example of FIG. 2 (e) is, for example, the measurement result of the position of the target based on the beat signal of the difference between the beat signal corresponding to the center frequency f2 and the beat signal corresponding to the center frequency f3.

The procedure 2 will be described with reference to FIG. 3A to 3C are diagrams showing the vectors of beat signals corresponding to the three frequencies in FIGS. 2A to 2C. The beat signal is a composite vector of a target component composed of a reflected wave reflected by the target and a transmission-transmission coupling component composed of a signal wave directly incident on the receiving unit 121 from the transmitting unit 111.
Since the transmission / reception coupling component has a constant phase and intensity regardless of the frequency shift, the difference processing between the beat signal corresponding to the center frequency f1 and the beat signal corresponding to the center frequency f2 is performed to obtain (d) of FIG. As shown in (e) and (f), the difference vector (1) of the beat signal from which the inter-transmission coupling component has been removed is obtained (corresponding to (d) in FIG. 2). Similarly, as shown in FIGS. 3 (d) and 3 (e), the beat from which the transmission / reception coupling component has been removed by the difference processing between the beat signal corresponding to the center frequency f2 and the beat signal corresponding to the center frequency f3. The signal difference vector (2) is obtained (corresponding to (e) in FIG. 2). In addition to the transmission / reception coupling component, if the target does not change its relative position with respect to the radar device 100 (for example, reflection from the bumper of the vehicle on which the radar device 100 is mounted), the frequency shift does not matter. Since the phase and intensity are constant, it is possible to obtain a difference vector of the beat signal from which such a component is removed.

The phase φ of the target component is determined by the reciprocating distance 2r between the radar device and the target with respect to the wavelength λ of the radar wave as follows.
fn = f1 + Δf · (n-1): Center frequency of radar wave. n = 1,2,3
c: Speed of light Further, the amount of phase change Δφ when the frequency is shifted by Δf is expressed by the following equation.
This Δφ cannot be obtained directly from the vector of the beat signal, but it can be obtained by using the difference between the vectors in the combination of f2 and f1 or the combination of f3 and f2.

In the present embodiment, the phase of the target component is constant, but the phase of the target component may be slightly rotated depending on how the transmission / reception coupling component is removed.

Note that step 2 does not necessarily have to be performed. In this case, in step 1, a transmission signal (transmission wave) changed into two frequencies with a predetermined frequency difference Δf is transmitted in order, and a beat signal corresponding to the two frequencies with a predetermined frequency difference Δf is received. The received beat signal itself may be used as a beat signal corresponding to two frequencies having a predetermined frequency difference Δf.

<Procedure 3>
Here, the inventors of the present application need to satisfy the following items (i) to (iii) in order to improve the separability of a plurality of targets by the FFT process and the distance estimation accuracy by the two-frequency comparison process. We have obtained the knowledge.
(I) Multiple targets can be detected separately for the azimuth (ii) Have phase information (iii) Estimate the same azimuth under two frequency conditions Each item is outlined below.

(I) Multiple targets can be detected separately for the azimuth:
It is necessary to be able to separate and detect multiple targets for the azimuth. For example, if it is detected that the azimuth angles of the detected targets are different from each other, this item (i) is satisfied. In addition, the peaks in the azimuth angle directions obtained from a plurality of targets are large. This item (i) is not satisfied when the peaks overlap and the peaks cannot be separated from each other.

(Ii) Having phase information:
It is necessary that the phase information is maintained for each of the target positions detected at the two frequencies. For example, when the radar device 100 measures the target position by IQ orthogonal demodulation of the signals received at the respective frequencies, the I signal and the Q signal each transmit each other's signals in the pre-stage of the procedure 4 described later. If it can be separated, this item (ii) is satisfied. On the other hand, if the I signal and the Q signal cannot be separated, this item (ii) is not satisfied.

(Iii) Estimating the same azimuth under two frequency conditions:
In the previous step of step 4 described later, it is necessary that the azimuth angle positions of a plurality of targets obtained from each of the two frequencies (including the beat signal) used are separated from each other and estimated to have substantially the same azimuth angle. is there. For example, when the azimuth angle positions of a plurality of targets obtained from each of the two frequencies are substantially the same (for example, the difference between the azimuth angle positions is within 5 deg), this matter (Iii) is satisfied. On the other hand, when the azimuth angle position of one of the plurality of targets obtained from each of the two frequencies is significantly different (for example, the difference in the azimuth angle position is 10 deg or more), this matter. (Iii) is not satisfied.

With reference to FIG. 4, an example in which the target is located at distances 2 m and 6 m in front of the radar device will be examined. FIG. 4 is a diagram showing the intensity of the difference vector (upper side) when the target is located at the distances of 2 m and 6 m, and the phase φ and the phase difference Δφ (lower side) of the difference vector. On the upper side of FIG. 4, the above-mentioned difference vector (1) is shown by a solid line, and the above-mentioned difference vector (2) is shown by a broken line. In the lower part of FIG. 4, the phase φ (1) of the difference vector (1) is shown by a solid line, the phase φ (2) of the difference vector (2) is shown by a broken line, and the phase difference Δφ is shown by a chain line.

For example, in a pulse radar device in the 24 GHz band, the frequency bandwidth may be set to 200 MHz or less in consideration of the influence on other devices using the same frequency band and the reduction of the influence from other devices. In such a case, the pulse width has a spread of about 3 m in terms of distance. On the upper side of FIG. 4, the same degree of uncertainty in distance estimation occurs. Since the difference vector has a property of becoming larger as the distance is longer, the peak at a distance of 6 m is higher than the peak at a distance of 2 m.
On the other hand, in Δφ on the lower side of FIG. 4, the value changes stepwise. The constant value of Δφ means that the estimated distance is a constant value, and by utilizing this property, the uncertainty due to the pulse width as shown in the intensity graph on the upper side of FIG. 4 is eliminated. be able to.
However, the above distance estimation by Δφ is established when the target component vector consists of a single target. In the vicinity of 4 m where the peak tails overlap on the upper side of FIG. 4, the vectors of the two target components interfere with each other, so that the Δφ on the lower side of FIG. 4 does not become a constant value, causing uncertainty in estimation. ..
This makes it necessary to estimate the distance regardless of the pulse width.

Next, with reference to FIG. 5, an example in which the target is located at distances 2 m and 4 m in front of the radar device will be examined. FIG. 5 is a diagram showing the intensity of the difference vector (upper side) when the target is located at the distances of 2 m and 4 m, and the phase φ and the phase difference Δφ (lower side) of the difference vector. On the upper side of FIG. 5, the above-mentioned difference vector (1) is shown by a solid line, and the above-mentioned difference vector (2) is shown by a broken line. In the lower part of FIG. 5, the phase φ (1) of the difference vector (1) is shown by a solid line, the phase φ (2) of the difference vector (2) is shown by a broken line, and the phase difference Δφ is shown by a chain line.

On the upper side of FIG. 5, the peaks overlap and separation detection cannot be performed. At this time, a region where the distance estimation in which the value of Δφ is not constant is uncertain spreads. In particular, this uncertain region overlaps with a region having high signal strength as shown in the upper part of FIG. In such a case, it is difficult to estimate the correct distance.
Assuming an application for detecting multiple targets near the radar device, it is not possible to assume a situation where the target distances are different by 4 m or more, so it is necessary to be able to separate and detect multiple targets at the azimuth angle. ((I) above).

In order to satisfy the above-mentioned requirements (i) to (iii), in step 3, the first signal processing unit 125 has a plurality of beat signals in which phase information is maintained for each of the beat signals corresponding to the two frequencies. Generates a separate beat signal in which the azimuths or distances of the targets are separated from each other.

FIG. 6 shows the measurement result (radar chart: XY Cartesian coordinate system) of the target position based on the beat signals corresponding to the two frequencies obtained by executing the high separation capability processing of the target azimuth angle described later. An example is shown. In this example, the measurement result of the target position based on any beat signal has peaks (blackened areas) along the two angular directions, and the target position based on any beat signal. In the measurement result of, there are peaks in the vicinity of the 0 deg direction and in the vicinity of the direction inclined by 30 degrees counterclockwise from the 0 deg direction.
(I) It can be seen that a plurality of targets can be separated and detected with respect to the azimuth (iii), and the same azimuth can be estimated under two frequency conditions.

Hereinafter, the high resolution of the azimuth angle of the target by the first signal processing unit 125 will be described in detail. The first signal processing unit 125 includes a first processing unit 11, a second processing unit 12, and a third processing unit 13. Hereinafter, the transmitting unit transmits a signal wave having coherence, and the receiving unit receives the reflected wave by four receiving antennas.

<< Procedure 3-1 >>
The first processing unit 11 linearly synthesizes K virtual single wave signals x _n (k)'with the received signals x _n generated by receiving signal waves from N existing receiving antennas. Approximate as a signal (N is 4, n is the number of N receiving antennas 0 to 3, K is 15, and k is the number of K virtual single targets. ).

Specifically, the first processing unit 11 assumes that the received signals x _n of the N receiving antennas are the combined signals x _n (0) of the K single wave signals x _n (k)'. Perform the following processing. That is, the first processing unit 11
(1) By Fourier transforming the combined signals x _n (k) of N receiving antennas, the azimuth angle distribution (azimuth angle spectrum) of the signal wave is generated as an estimation function Y (k).
(2) For the single-wave signal forming the peak of the generated estimation function Y (k), the single-wave signal x _n (k)'of N receiving antennas is generated.
(3) every N number of receiving antennas, the synthesized signal _x n (k), subtracts the single wave signal _x n (k) 'to multiplied by the predetermined subtraction ratio _{M M · x n (k)} ' Then, the remaining combined signals x _n (k') of the N receiving antennas are generated.
x _n (k') = x _n (k) -M · x _n (k)'
k'is a value obtained by incrementing k by 1.
The first processing unit 11 repeats the above processes (1) to (3) for the remaining combined signal x _n (k') until k'= K.
In this way, by incrementing k by 1 and repeating the processes (1) to (3) above for the combined signal x _n (k), the received signals x _n of the N receiving antennas are K. It can be approximated as a linear composite signal of a single wave signal x _n (k)'.

In the above and later sections, x _n is the actually measured receiving antenna signal, and x _n (k) is the receiving antenna signal from which k single wave signal components have been subtracted (before subtraction, that is, k = 0). When x _n (0) = x _n ), Y (k) is an estimation function calculated by Fourier transforming x _n (k) (before subtraction, that is, when k = 0, Y (0) = Y0). , X _n (k)'is the receiving antenna signal of the single wave signal forming the peak of Y (k), and Y (k)'is the estimation function calculated by Fourier transforming x _n (k)'. Yes, k'= k + 1, and x _m'is a received signal obtained by synthesizing K single wave signals (m: antenna number including real and non-real).

7 to 11 are diagrams for explaining the analysis (procedure 3-1) of the actually measured 4 antenna data by the first processing unit 11.
FIG. 7A is a diagram showing an example of the received signals x _n of the four receiving antennas, that is, the combined signals x _n (0). In (A) of FIG. 7, the real part I _n and an imaginary part Q _n of the received signal x _n received in real receiving antenna (number n = 0 to 3) are shown by black circles and black triangles. Incidentally, in FIG. 7 (A) is a real part _{I n} and an imaginary part _{Q n} are open circles and open triangles of the received signal of the receiving antenna (number n = -4~-1,4~11 nonexistent for reference Although shown, these signals are not used in this procedure 3-1. The horizontal axis is the receiving antenna number n, and the vertical axis is the receiving intensity.
FIG. 7A illustrates a case where two signal waves, a signal wave from an azimuth angle of 0 degrees and a signal wave from an azimuth angle of 20 degrees, are received. The amplitude A and phase φ of the signal wave from the azimuth angle 0 degrees are A = 1 and φ = 0 degrees, and the amplitude A and phase φ of the signal wave from the azimuth angle 20 degrees are A = 1 and φ = 180 degrees. is there.

First, the first processing unit 11 generates the azimuth angle distribution (azimuth angle spectrum) of the signal wave as the estimation function Y (0) by Fourier transforming the combined signals x _n (0) of the four receiving antennas. ..
FIG. 7B is a diagram showing an example of the estimation function Y (0). In FIG. 7B, the real part Re_Y (0) and the imaginary part Im_Y (0) of the estimation function Y (0), and the size | Y (0) | which is the square root of the sum of squares thereof are medium. It is shown by the thick solid line, the solid line, and the thickest solid line. The horizontal axis is the estimated angle θ, and the vertical axis is the reception intensity.
The azimuth θ0 of the peak p0 at | Y (0) | is not the azimuth of the true signal wave, but a false azimuth caused by the interference of the two signal waves, but it does not matter at this stage. ..

Next, the first processing unit 11 is a single receiving antenna (number n = 0 to 3) existing for the single wave signal forming the peak p0 (estimated angle θ0) of the generated estimation function | Y (0) |. Generate a one-wave signal x _n (0)'.
FIG. 7C is a diagram showing an example of a single wave signal x _n (0)'. The (C) of FIG. 7, 'the real part _I n (0) of the' single wave signal _x n (0) of the actual receiving antenna (number n = 0 to 3) and the imaginary part _Q n (0) ' Is indicated by black circles and black triangles. The horizontal axis is the receiving antenna number n, and the vertical axis is the receiving intensity.
For example, the first processing unit 11 calculates the real part I ₀ (0)'and the imaginary part Q ₀ (0)' of the single wave signal x ₀ (0)'with the receiving antenna number n = 0 by the following equation. ..
I ₀ (0)'= Re [Y (0) @ Azimuth = θ0] / 4
Q ₀ (0)'= Im [Y (0) @ Azimuth = θ0] / 4
Further, the first processing unit 11 is a real part of a single wave signal x _n (0)'of receiving antenna number n = 0 to 3 based on the phase difference Δφ0 of the signal waves between the receiving antennas arranged at equal intervals d. calculating the I _n (0) 'and the imaginary part _Q n (0)'.
In (0)'= Re [x _n (0)' · exp (j · Δφ0 · n)]
Qn (0)'= Im [x _n (0)' · exp (j · Δφ0 · n)]
Δφ0 = -2π (d / λ) sin (θ0)

Next, first processing unit 11, the azimuth angle distribution (azimuth angle 'by Fourier transform, a single wave signal x _{n (0)'} of the four receiving antennas single wave signal x _{n (0)} Spectrum) is generated as the estimation function Y (0)'.
FIG. 7D is a diagram showing an example of the estimation function Y (0)'. In FIG. 7D, the real part Re_Y (0)'and the imaginary part Im_Y (0)' of the estimation function Y (0)', and the size | Y (0)' which is the square root of the sum of squares thereof. | Is indicated by a medium-thick solid line, a solid line, and the thickest solid line. The horizontal axis is the estimated angle θ, and the vertical axis is the reception intensity.

Next, the first processing unit 11 multiplies the single wave signal x _n (0)'by a predetermined subtraction rate M from the received signal x _n (0) for each of the four existing receiving antennas. Subtract x _n (0)'to calculate the remaining combined signal x _n (1) of the four receiving antennas.
x _n (1) = x _n (0) -M · x _n (0)'
FIG. 8 (E) is a diagram showing an example of the remaining synthetic signal x _n (1). The (E) in FIG. 8, the real part I _{n (1)} and an imaginary part Q _n of the remainder of the synthesis signal x _n of the four receiving antennas real _{(1) (1)} is shown by black circles and black triangles ing. The horizontal axis is the receiving antenna number n, and the vertical axis is the receiving intensity. Further, the predetermined subtraction rate M is 0.4.

The first processing unit 11 repeats the above processing for the remaining composite signal x _n (k') until k'= K.
Specifically, the first processing unit 11 Fourier transforms the remaining composite signals x _n (1) of the four receiving antennas to perform an azimuth angle distribution (azimuth angle spectrum) of the composite signals x _n (1). Is generated as the estimation function Y (1).
FIG. 8F is a diagram showing an example of the estimation function Y (1). In FIG. 8 (F), the real part Re_Y (1) and the imaginary part Im_Y (1) of the estimation function Y (1), and the size | Y (1) | which is the square root of the sum of squares thereof are medium. It is shown by the thick solid line, the solid line, and the thickest solid line. The horizontal axis is the estimated angle θ, and the vertical axis is the reception intensity.
The first processing unit 11 has the single wave signal x _n (0) shown in FIG. 7 (D) from the estimation function Y (0) of the combined signal x _n (0) shown in FIG. 7 (B). The estimation function Y (0) of the remaining composite signal x _n (1) of the four receiving antennas is subtracted by subtracting MY (0)', which is obtained by multiplying the estimation function Y (0) of'by a predetermined subtraction rate M. 1) may be calculated.
Y (1) = Y (0) -MY (0)'

In FIG. 8 (F), since the component of the single wave signal x _n (0)'forming the false apex is reduced, the peak p1 is obtained at an azimuth angle closer to the true one. However, since the single wave signal x _n (0)'contains a true received signal component in addition to forming a false peak, it is inappropriate to set the subtraction rate M = 1, and M <0. About 5.5 is considered appropriate.
Further, when p0 is a true azimuth, a peak is formed at Y (1) or the same azimuth thereafter, and the truth of p0 may be determined using the result. ..

Next, the first processing unit 11 of the four receiving antennas (numbers n = 0 to 3) for the single wave signal forming the peak p1 (estimated angle θ1) of the generated estimation function | Y (1) |. Generate a single wave signal x _n (1)'. The method of generating the single wave signal x _n (1)'is the same as the method of generating the single wave signal x _n (0)' described above.
FIG. 8 (G) is a diagram showing an example of a single wave signal x _n (1)'. The (G) in FIG. 8, a single wave signal _x n (1) 'real part _I n (1) of' and an imaginary part _Q n of the real four receive antennas (number n = 0 to 3) ( 1)'is indicated by a black circle and a black triangle. The horizontal axis is the receiving antenna number n, and the vertical axis is the receiving intensity.

Next, the first processing unit 11 Fourier transforms the single wave x _n (1)'of the four receiving antennas to obtain the azimuth angle distribution (azimuth angle spectrum) of the single wave signal x _n (1)'. ) Is calculated as the estimation function Y (1)'.
FIG. 8H is a diagram showing an example of the estimation function Y (1)'. In FIG. 8 (H), the real part Re_Y (1)'and the imaginary part Im_Y (1)' of the estimation function Y (1)', and the size | Y (1)' which is the square root of the sum of squares thereof. | Is indicated by a medium-thick solid line, a solid line, and the thickest solid line. The horizontal axis is the estimated angle θ, and the vertical axis is the reception intensity.

In this way, the first processing unit 11 increments k by 1 and repeats the above processing with respect to the combined signal x _n (k), so that the first processing unit 11 has four pieces as shown in FIGS. 9 and 10. The received signal x _n (0) of the receiving antenna is approximated as a linear composite signal of 15 single wave signals x _n (k)'.

When the above process is sufficiently repeated (15 times in this embodiment), the received signal _xn and the estimation function Y approach 0 as shown in FIGS. 10 (I) and 10 (J). Looking at the process along the way, the vertices of | Y (k) | alternately show the true azimuths of around 0 degrees and around 20 degrees. In this way, by gradually removing the tentative azimuth component corresponding to the peak obtained each time from the actually measured signal, a peak close to the true value can be obtained.
At this time, the actually measurable received signal x _n (0) and the estimation function Y (0) are 15 single wave signals x _n (0)'to x _n (14)' and their estimation function Y (0). It is probable that it could be decomposed into'~ Y (14)'. However, the azimuth angle of this individual single wave signal is not a true value.

(L) of FIG. 11 shows the 15 single wave signal _x n (0) of the _'~x n (14)' results of linear synthesized, (M) in FIG. 11, 15 single wave The result of linearly synthesizing the signal estimation signals Y (0)'to Y (14)' is shown. According to (L) and (M) of FIG. 11, (A) and (B) of FIG. 7 can be reproduced well, and it can be confirmed that this decomposition treatment is appropriate.

<< Procedure 3-2 >>
The second processing unit 12 has (16-4) non-existent (16-4) non-existent receiving antennas (numbers -4 to 4) for every 15 single-wave signals x _n (k)', according to a predetermined phase difference Δφ. Generates a single wave signal of -1,4 to 11). in this way,
x _{n + 1} (k)'= x _n (k)' · e ^jΔφ
Phase difference Δφ = -2π (d / λ) sin (θ)
In consideration of, it is possible to estimate the received signal of the receiving antenna that does not actually exist. This is equivalent to enlarging the receiving antenna opening.

The second processing unit 12 has 15 for each of the four existing receiving antennas (numbers n = 0 to 3) and the non-existing 12 non-existing receiving antennas (numbers -4 to -1, 4 to 11). By synthesizing the single-wave signal x _n (k)'and the single-wave signal of 15 non-existent receiving antennas, 4 receiving antennas (numbers n = 0 to 3) and 12 non-existing receiving antennas. generating a synthesized signal _{x m} 'real reception antennas (number -4~-1,4~11) (M> a N a M = 16, m is four receive antennas and (M-N) Number of non-existent receiving antennas).

FIG. 12 is a diagram showing an example of the composite signal x _m '. Comparing FIG. 12 and FIG. 7 (A), it is a value that cannot be confirmed by an actual radar device, but when compared with the true value signal _xn (0), a actually measurable receiving antenna (number 0). For ~ 3), about 5 signals x _{m'before and} after showed values close to the true value.
This corresponds to the receiving antenna opening expanding by (4 + 5) / 4≈2.3 times.

<< Procedure 3-3 >>
The third processing unit 13 generates an estimation function Y'by Fourier transforming the generated composite signal x _m ', and sets the azimuth angle of the Y'peak to the azimuth angle of the signal wave, that is, the azimuth angle of the wave source (wave source position). ).
By this process, the 15 single-wave signals x _n (k)'obtained in step 1 are combined (in the process of synthesis, the components close to the true azimuth strengthen each other, and the components far away are canceled and attenuated. ), Forming a peak close to the true azimuth.
FIG. 13 is a diagram showing an example of the estimation function Y'. In FIG. 13, peaks appear at azimuth angles close to 0 degrees and 20 degrees, which are true azimuth angles.
However, an erroneous peak occurs between the two peaks. This is the effect of the first peak p0 in the analytical process of procedure 1. This point will be described in Modification 1 described later.

As described above, according to the high resolution of the azimuth angle of the target in step 3, the received signal x _n is a single wave signal x _n (k)'generated from K virtual single targets. A single wave signal of a non-existent non-existent receiving element is generated for each single wave signal x _n (k)'according to a predetermined phase difference (reception antenna opening expansion), and a single signal is generated. A combined signal x _m'is generated by combining a wave signal x _n (k)'and a single wave signal of a non-existent receiving element. As a result, even if the received signal _xn includes signals from two or more targets, it includes a non-existent receiving antenna according to the phase difference between the receiving antennas of the single wave signal, that is, the receiving antenna opening is expanded. A composite signal x _m'can be generated.
The receiving antenna aperture enlarged synthesized signal x _{m 'is} Fourier transformed, by generating an azimuth angle distribution (azimuth angle spectrum) (estimation function), it is possible to increase the resolution of the two or more peaks. Therefore, when estimating these peaks as the azimuth angles of two or more targets, the separability of the estimation of the azimuth angles of the targets can be enhanced.

Further, according to the high resolution of the azimuth of the target in step 3, the number of single wave signals generated by the MN non-existent receiving elements in the second processing unit 12 is the number of the N receiving elements. It is more than twice the number of single wave signals x _n (k)'. As a result, the separation limit of the azimuth angles of the plurality of targets can be made smaller than the Rayleigh limit, and the azimuth angle estimation using the Fourier transform by the third processing unit 13 can be made highly separable.

As described above, in step 3-1 in the estimation function Y, the shape of the peak including the phase is analyzed up to the base, and the peak shape including the phase is reproduced by linear synthesis of a plurality of waves, and the procedure 3- In 2, the receiving antenna opening is enlarged according to a predetermined phase difference. As a result, in the high separation ability of the azimuth angle of the target in step 3,
(Ii) It can be seen that the phase information is not impaired, that is, the phase information is possessed.
Further, as described above (FIG. 6), by increasing the separation ability of the azimuth angle of the target in step 3,
(I) It can be seen that a plurality of targets can be separated and detected with respect to the azimuth (iii), and the same azimuth can be estimated under two frequency conditions.

(Procedure 4)
Next, the second signal processing unit 156 estimates the distance r from the target as the position of the target from the predetermined frequency difference Δf and the phase difference Δφ of the separated beat signals corresponding to the two frequencies. Specifically, the second signal processing unit 156 uses the following equation (1) based on a predetermined frequency difference Δf (Hz) and the speed of light c (m / s) to determine the phase difference of the separated beat signals corresponding to the two frequencies. The distance r (m) from the target is estimated from Δφ (rad) (frequency shift method Ver.2).

According to the example of FIG. 6, even if the target azimuth angle is enhanced to be separated in the procedure 3, the distance estimation of about 3 m due to the pulse width is ambiguous in each of the targets TA and TB. However, when the phase difference Δφ of the same xy coordinates is examined for the two-frequency beat signals (1') and (2'), the values are the same. From this feature, the distance can be re-estimated using Δφ and remapped on the xy coordinates.

For example, if the distance is m, the azimuth is n, the amplitude at an arbitrary mn is Amn, and the phase difference at an arbitrary mn is Δφmn, the position in a predetermined range n (n may be all or only one). The phase difference Δφn can be expressed as the following equation.
(M is the upper limit of m. M may be an arbitrary value or all m.)
In step 4, Δφn may be obtained from the above equation for the azimuth angle n in the predetermined range mapped in FIG. 6, and the distance r corresponding to Δφn may be estimated using the above equation (1).

Alternatively, the concept of histogram may be introduced.
Let the amount indicating the coordinates be the distance m and the azimuth angle n. (M is the distance at the time of procedure 3, n is the angle at the time of procedure 3. However, n is common to steps 3 and 4.)
For a certain n = k
A phase difference Δφ (m, k) satisfying f (s) ≤ Δφ (m, k) <f (s + 1) is specified from within a predetermined range of m (the predetermined range may be all). It is assumed that f_ (s) is associated with the distance r (s) in step 4. (For example, f (s) is proportional to the distance r (s))
-The amplitude A (m, k) associated with Δφ (m, k) satisfying the above is summed A (k) with respect to m in the predetermined range.
-Let the histogram value of r (s) be sumA (k).
This is repeated for s within a predetermined range (all s may be used).
These processes are performed for a predetermined k (all ks may be used).
Then, in step 4, the above-mentioned Δφ (m, k) is obtained for the azimuth angle n in the predetermined range mapped in FIG. 6, and the distance r corresponding to Δφ (m, k) is calculated using the above equation (1). You may estimate.

FIG. 14 shows an example of the distance estimation result (radar chart: XY Cartesian coordinate system). According to the example of FIG. 6, even if the target azimuth angle is enhanced to be separated in the procedure 3, the distance estimation of about 3 m due to the pulse width is ambiguous. On the other hand, the two-frequency comparison process of step 4 is performed, and in each of the targets TA and TB, the phase difference between the two-frequency beat signals (1') and (2') in FIG. 6 is used to obtain the target. By estimating the distance and remapping it on the radar chart, it can be seen that the distance estimation accuracy is improved in the example of FIG.

Next, the position estimation operation by the radar device (position estimation device) 100 will be described with reference to FIGS. 15 and 16. FIG. 15 is a flowchart showing the position estimation process by the radar device (position estimation device) shown in FIG. 1, and FIG. 16 shows the azimuth angle high separation ability process (DOA1) in the position estimation process shown in FIG. It is a flowchart.

<Procedure 1>
First, as shown in FIG. 15, the transmission control unit 110 and the transmission unit 111 sequentially transmit transmission signals (signal waves) changed to three frequencies with a predetermined frequency difference Δf. At this time, the receiving unit 121 sequentially receives the reflected wave whose transmission signal is reflected by the target as a reception signal, and generates a beat signal (digital signal) whose frequency is converted from the reception signal by the frequency of the transmission signal (S100). The order of the frequencies of the transmission signals to be transmitted may be arbitrarily determined.

<Procedure 2>
Next, the first signal processing unit 125 generates a beat signal corresponding to two frequencies having a predetermined frequency difference Δf by the difference processing of the beat signals corresponding to the three frequencies (frequency shift method Ver.1). (S200).

<Procedure 3>
Next, the first signal processing unit 125 generates a separated beat signal in which the azimuth angles or distances of a plurality of targets are separated from each other with the phase information maintained for each of the beat signals corresponding to the two frequencies. (S300).
<< Procedure 3-1 >>
Specifically, as shown in FIG. 16, the first processing unit 11 sets k = 0 (S1), and the received signals x _n of the N receiving antennas are K single wave signals x _n (S1). Assuming that the composite signal x _n (0) of k)'is performed, the following processing is performed. That is, the first processing unit 11
(1) By Fourier transforming the combined signals x _n (k) of N receiving antennas, the azimuth angle distribution (azimuth angle spectrum) of the signal wave is generated as an estimation function Y (k) (S2).
(2) For the single-wave signal forming the peak of the generated estimation function Y (k), the single-wave signal x _n (k)'of N receiving antennas is generated (S3).
(3) every N number of receiving antennas, the synthesized signal _x n (k), subtracts the single wave signal _x n (k) 'to multiplied by the predetermined subtraction ratio _{M M · x n (k)} ' Then, the remaining combined signals x _n (k) of the N receiving antennas are generated (S4).
x _n (k') = x _n (k) -M · x _n (k)'
k'is a value obtained by incrementing k by 1.
The first processing unit 11 increments k by 1 (S5), and until k = K (S6), the remaining composite signals x _n (k) are processed in the above (1) to (3). repeat.

<< Procedure 3-2 >>
Next, the second processing unit 12 has (16-4) non-existent (16-4) non-existent receiving antennas (numbers) for each of the 15 single-wave signals x _n (k)', according to a predetermined phase difference Δφ. -4 to -1, 4 to 11) single wave signals are generated (reception antenna opening expansion) (S7).
The second processing unit 12 has 15 receiving antennas (numbers n = 0 to 3) and 12 non-existing receiving antennas (numbers -4 to -1, 4 to 11) that do not exist. By synthesizing the single-wave signal x _n (k)'and the single-wave signal of 15 non-existent receiving antennas, 4 receiving antennas (numbers n = 0 to 3) and 12 non-existing receiving antennas. generating a synthesized signal _{x m} 'real reception antennas (number -4~-1,4~11) (S8).

<< Procedure 3-3 >>
Next, the third processing unit 13 generates an estimation function Y'by Fourier transforming the generated composite signal x _m '(S9), and sets the azimuth angle of the peak of Y'to the azimuth angle of the signal wave, that is, Estimated as the azimuth of the target (S10).

<Procedure 4>
Next, as shown in FIG. 15, the second signal processing unit 156 determines the distance r from the target as the position of the target from the predetermined frequency difference Δf and the phase difference Δφ of the separated beat signals corresponding to the two frequencies. Estimate (S400). Specifically, the second signal processing unit 156 uses the above equation (1) based on the predetermined frequency difference Δf and the speed of light c to obtain the distance r from the phase difference Δφ of the separated beat signals corresponding to the two frequencies. (Frequency shift method Ver.2) (S400).

Here, in the radar device, for example, the distance is estimated by measuring the time required for the transmitted pulse signal to reciprocate with the target. However, for example, in the 24 GHz band, the frequency bandwidth that can be used for pulse formation needs to be, for example, 200 MHz or less in consideration of the influence of other devices that use the same frequency band. .. Therefore, the pulse width has a spread of about 3 m in terms of distance, and there is a problem that the estimated distance becomes ambiguous by that amount.

In this regard, according to the radar device (position estimation device) 100 of the present embodiment, by combining the frequency shift method (procedure 4) and the high separation ability of the azimuth (procedure 3), in other words, By combining the target information detected separately in the azimuthal direction and the phase difference information generated by the frequency shift, it is possible to improve the separation ability of multiple targets by FFT processing and improve the distance estimation accuracy by dual frequency comparison processing. it can.

For example, in the examples of FIGS. 6 and 14, the target TA at a distance of 1.5 m from the radar device and the target TB at a distance of 1 m from the radar device can be separately detected and before and after with a distance difference of 0.5 m. The relationship can be determined.

FIG. 17 shows an example of the distance estimation result (radar chart: XY Cartesian coordinate system) of the comparative example. In the comparative example, the azimuth angle of the azimuth of the procedure 3 in the procedures 1 to 4 of the present embodiment was not enhanced. As described above, if the two targets cannot be separated and detected by the azimuth angle, a plurality of target components are mixed in the difference vector, and the distance estimation by the two-frequency comparison process cannot be performed correctly.

(Modification example 1)
In the above-described embodiment, as shown in FIG. 13, an erroneous peak may occur between the two peaks. This is the effect of the first peak p0 in the analytical process of procedure 1.
In the first modification, the peak of the first i times in the procedure 3-1 is ignored.

FIG. 18 is a flowchart of the azimuth angle high separation capability processing (DOA2) in the position estimation processing by the radar device (position estimation device) 100 according to the modification 1.

<< Procedure 3-2 >>
The second processing unit 12 has K single-wave signals for each of N receiving antennas (numbers 0 to 3) and MN non-existing receiving antennas (numbers -4 to -1, 4 to 11). 'by combining the synthetic signal x _m of the n receive antennas and M-n pieces of imaginary receive _antenna' x n _(k) when generating a K-number of single wave signal x _{n (k)} 'And the corresponding i of the single-wave signals of the K non-existent receiving antennas are not used.
Δφ0 = -2π (d / λ) sin (θ _k )
* M is the receiving antenna number (-4 to 11)

FIG. 19A is a diagram showing a calculation result of the estimation function Y of the modification 1. As shown in FIG. 19A, the erroneous peak that occurred at the azimuth angle of 7 degrees can be reduced. As described above, in step 1, the first i batches (i = 2 is considered appropriate) of the single wave signal in which there is a high possibility that a peak is generated at an erroneous angle due to the interference of a plurality of signal waves is obtained in step 2. By not using it, the occurrence of erroneous peaks can be reduced.

However, instead, the peak intensity corresponding to the true value of 0 degrees, which is close to the erroneous peak, is lowered, and the balance with the peak intensity of 20 degrees, which is originally the same intensity, is lost.

This peak intensity imbalance can be alleviated by reducing the value of the subtraction rate M used in step 1 (for example, M = 0.4 → 0.3) (see FIG. 19B).
Since the peak intensity is a value that greatly changes depending on the state of the target (shape, material, orientation), etc., it is considered that this imbalance is not a serious problem if it is within the permissible range.

In the present embodiment, the predetermined i times that are not used are set to the first i times, but may be set to i times other than the first.

Next, with reference to FIG. 18, the azimuth angle high separation capability processing (DOA2) in the position estimation processing by the radar device (position estimation device) 100 of the modification 1 will be described. The azimuth high separation capability process shown in FIG. 18 includes step S11 instead of step S8 (procedure 3-2) in FIG. 16, and the azimuth angle high separation capability process (DOA1) of the above-described embodiment is included. ) Is different.

In step S11, the second processing unit 12 has K numbers for every N receiving antennas (numbers 0 to 3) and MN non-existing receiving antennas (numbers -4 to -1, 4 to 11). 'by combining the synthetic signal of n receive antennas and M-n pieces of imaginary receive antennas x _m' single wave signal x _{n (k)} when generating a K-number of single wave signal x Of the single wave signals of _n (k)'and the corresponding K non-existent receiving antennas, i are not used.

(Modification 2)
In the second modification, the procedures 3-1 to 3-3 of the modification 1 are performed, and then the procedures 3-1 to 3-3 are performed again. In step 3-1 of the second lap, a single wave signal x _n (k)'is generated at the azimuth angle obtained in the first lap.

FIG. 20 is a flowchart of the azimuth angle high separation capability processing (DOA3) in the position estimation processing by the radar device (position estimation device) 100 of the modification 2.

When steps 3-1 to 3-3 of the first modification are performed, an azimuth angle of 24 degrees is obtained as shown in FIG. 19B.
<< Procedure 3-1 on the second lap >>
The first processing unit 11
(2) 'when generating the first, already synthesized signal x _m by the third processing unit _13' single wave signal x _n of N receive antennas _(k) based on the azimuth angle of the target which is estimated using the hand,
(2) With respect to the single-wave signal forming the already estimated target azimuth estimation function Y (k) in the generated estimation function Y (k), the single-wave signals x _n of N receiving antennas ( k) Generate'.

For example, as shown in FIG. 21, the first processing unit 11 does not use the peak of the estimation function | Y (0) | of the combined signal x _n (0), but a signal having an azimuth angle of 24 degrees obtained in the first round. A single-wave signal to be formed is generated, and the combined signal _xn (0) is first subtracted from this single-wave signal.
At this time, the first processing unit 11 sets the subtraction rate M to a large value only at the beginning, and greatly reduces the factors affected by the erroneous peak at the azimuth angle of 7 degrees (for example, M = 0.8).

In the second lap of step 3-1 as a result of eliminating the influence of the erroneous peak with an azimuth angle of 7 degrees, the component for the first i times lost in the first modification can be recovered (see FIG. 22).
In addition, the error of the peak corresponding to the true value of 0 degrees can be reduced (see FIG. 22).

In the second modification, further using this result, if the first subtraction of the procedure 3-1 on the third lap is performed at -1 degree, the error of the peak with a true value of 24 degrees can also be reduced.

Next, with reference to FIG. 20, the azimuth angle high separation ability processing (DOA3) in the position estimation processing by the radar device (position estimation device) 100 of the modification 2 will be described. The high separation performance treatment shown in FIG. 20 is different from the above-described embodiment in that the azimuth angle high separation performance treatment shown in FIG. 16 further includes steps S12 and S13.

In step S12, the azimuth angle high separation ability processing (DOA2) of the modification 1 shown in FIG. 18 is performed. After that, the processing of the second lap is performed.
In the processing of the second lap, after setting k = 0 (S1), the first processing unit 11 moves.
(1) By Fourier transforming the combined signals x _n (0) of N receiving antennas, the azimuth angle distribution (azimuth angle spectrum) of the signal wave is generated as an estimation function Y (0).
(2) Instead of the single wave signal forming the peak of the generated estimation function Y (0), based on the azimuth angle of the wave source already estimated by the third processing unit 13 using the combined signal x _m ', For the single wave signal forming the estimated function Y (0) of the azimuth angle of the wave source already estimated in the generated estimation function Y (0), the single wave signals x _n (0)'of N receiving antennas. To generate
(3) every N number of receiving antennas, the synthesized signal _{x n (0), 'M} · x n (0) multiplied by the predetermined subtraction ratio M in' single wave signal _x n (0) by subtracting the Then, the remaining combined signals x _n (1) of the N receiving antennas are generated (S13).

In other words, in the first subtraction of the second lap, the first processing unit 11 does not have the peak of the estimation function | Y (0) | of the composite signal x _n (0), but the azimuth angle 24 obtained in the first lap. A single-wave signal that forms a degree signal is generated, and the combined signal _xn (0) is first subtracted from this single-wave signal.
At this time, the first processing unit 11 sets the subtraction rate M to a large value only at the beginning, and greatly reduces the factors affected by the erroneous peak at the azimuth angle of 7 degrees (for example, M = 0.8).
After that, the processes of steps S2 to S10 are performed as described above.

(Modification 3)
The procedure for estimating the azimuth angle of the above-described embodiments and modifications can be applied to the distance estimation as it is. First, the outline of the distance estimation according to the modification 3 will be described.
FIG. 23 is a diagram for explaining an example of distance estimation according to the modification 3 (FMCW method), and FIG. 24 is a diagram for explaining another example of the distance estimation according to the modification 3 (FMCW method). Pulse method).
FIG. 23 (A) shows the received signals Rx0 to Rx3 for each of the four receiving antennas that receive the FMCW system signal wave, and the vertical axis thereof is the square root of the sum of squares of the real number component and the imaginary number component. ) | I + jQ |, the real number component Rx_I and the imaginary number component Rx_Q of the received signal, and the horizontal axis is time. FIG. 23B shows an estimation function Y obtained by Fourier transforming the received signals Rx0 to Rx3 shown in FIG. 23A, and the vertical axis thereof is the square root of the sum of squares of the real number component and the imaginary number component. ) | I + jQ |, the real number component Rx_I and the imaginary number component Rx_Q of the estimation function, and the horizontal axis is the frequency. FIG. 23 (C) shows an estimation function Y obtained by converting the horizontal axis of the estimation function Y shown in FIG. 23 (B) into a distance.
FIG. 24 (B) is an estimation function Y obtained by Fourier transforming the received signals Rx0 to Rx3 for each of the four receiving antennas that receive the pulse type signal wave or the received signals Rx0 to Rx3 shown in FIG. 24 (A). The vertical axis is the amplitude of the received signal or the estimation function Y (square root of the sum of squares of the real component and the imaginary component) | I + jQ |, the real component Rx_I and the imaginary component Rx_Q of the received signal or the estimation function Y, and the horizontal axis. Is the time. FIG. 24A shows the received signals Rx0 to Rx3 obtained by inverse Fourier transforming the received signals Rx0 to Rx3 shown in FIG. 24B, and the vertical axis thereof is the amplitude of the received signal (the square of the real number component and the imaginary number component). The square root of the sum) | I + jQ |, the real component Rx_I and the imaginary component Rx_Q of the received signal, and the horizontal axis is the frequency. FIG. 24 (C) shows an estimation function Y in which the horizontal axis of the estimation function Y shown in FIG. 24 (B) is converted into a distance.

For example, when applying the above-mentioned azimuth angle estimation procedure to the distance estimation of an FMCW (Frequency Modulated Continuous Wave) radar,
Instead of the received signals _xn of the _n receiving antennas of FIG. 7A, the received signals for the time of (A) of FIG. 23 may be used for each receiving antenna. That is, in FIG. 7A, the horizontal axis is the time number of the time-series received data instead of the receiving element number.
Further, instead of the estimation function Y of the angle distribution (angle spectrum) of FIG. 7B after the Fourier transform, the estimation function Y of the distance distribution of FIG. 23 (C) may be used for each receiving antenna. That is, in FIG. 7B, the horizontal axis is the distance instead of the azimuth.

On the other hand, when the above-mentioned azimuth angle estimation procedure is applied to the distance estimation of the pulse radar, the received signal strength data corresponding to the time can be obtained as shown in FIG. 24 (B), but the time axis is converted into the distance. The distance estimation function Y can be obtained simply by doing so. Therefore, there is no data before Fourier transform like the FMCW radar, but it is possible to generate data before Fourier transform by performing inverse Fourier transform on the time axis data as shown in FIG. 24 (A). Corresponds to the received signals _xn of the _n receiving antennas of FIG. 7A in the azimuth angle estimation. The same processing can be performed by expanding the frequency band of the pulse signal in step 3-3 instead of increasing the time series data in step 3-2.

The distance estimation method of the present embodiment can be applied to various devices using the Fourier transform.
For example, in the FMCW type radar device, the distance is estimated by Fourier transforming the time series data of the received signal. Therefore, as described above, the distance estimation method of the present embodiment is applied to the FMCW type radar device. It is possible.

On the other hand, the pulse type radar device does not have the above Fourier transform part (when the pulse method is expressed based on the FMCW method, the pulse method is a radar device that performs an operation equivalent to the Fourier transform on an analog circuit”. is there). Therefore, although it cannot be diverted as it is, the above-mentioned received signal x _n equivalent can be generated by inverse Fourier transforming the measurement data.

Further, the distance estimation method of the present embodiment is not limited to the FMCW method and the pulse method, and can be applied to the sophistication of the relative velocity analysis (Doppler spectrum analysis) and the detailed analysis of the micro Doppler.

Next, the details of the distance estimation according to the modified example 3 will be described.
(In the case of FMCW)
For example, as shown in (A) of FIG. 23 (corresponding to (A) of FIG. 7 described above), the signal waves (Rx0 to Rx3) received by the receiving antenna are frequency-modulated waves whose frequency changes on the time axis. _Let x _n be the received signal at N times when the time axis is divided (N is an antenna of 2 or more, and n is the number of N times).
The first signal processing unit 125 in the radar device (position estimation device) 100 performs the following processing for each receiving antenna, that is, for each received signal x _n (Rx0 to Rx3) received by the receiving antenna.

<< Procedure 3-1 >>
Similar to the above-described embodiment, the first processing unit 11 converts the received signal _xn generated by receiving the signal wave by the receiving antenna into a single wave generated from K virtual single targets. Approximate as a composite wave signal of the signal x _n (k)'(K is an integer of 2 or more, and k is the number of K virtual single targets).

Specifically, the first processing unit 11 assumes that the received signals x _{n at} N times are the combined signals x _n (0) of K single wave signals x _n (k)', and the following Is processed. That is, the first processing unit 11
(1) By Fourier transforming the combined signal x _n (k) of N times, the distance distribution of the signal wave is generated as the estimation function Y (k) ((C) of FIG. 23: FIG. 7 described above. (B) equivalent),
(2) For the single-wave signal forming the peak of the generated estimation function Y (k), a single-wave signal x _n (k)'at N times is generated.
(3) Every N times, subtract M · x _n (k)', which is obtained by multiplying the single wave signal x _n (k)' by a predetermined subtraction rate M, from the composite signal x _n (k). , Generates the remaining composite signals x _n (k') of N times.
x _n (k') = x _n (k) -M · x _n (k)'
k'is a value obtained by incrementing k by 1.
The first processing unit 11 repeats the above processes (1) to (3) for the remaining combined signal x _n (k') until k'= K.
In this way, by incrementing k by 1 and repeating the processes (1) to (3) above for the combined signal x _n (k), K received signals x _n at N times can be obtained. Can be approximated as a linear composite signal of the single wave signal x _n (k)'.

<< Procedure 3-2 >>
Similar to the above-described embodiment, the second processing unit 12 has MN non-existent non-existent times for each K single-wave signals x _n (k)'according to a predetermined phase difference. Generate a single wave signal. The number of MN non-existent time single wave signals generated is more than twice the number of N time single wave signals x _n (k)'. As a result, the separation limit of the distances of the plurality of wave sources can be reduced, and the distance estimation using the Fourier transform by the third processing unit 13 can be made highly separable.

The second processing unit 12 synthesizes K single-wave signals x _n (k)'for each of N existing times and MN non-existing times that do not exist, as in the above-described embodiment. by, generates a synthesized signal of the N times and M-N pieces of imaginary time x _{m 'is} an (M> N, m is the N times and (M-N) pieces of imaginary time It is the number of).

<< Procedure 3-3 >>
Similar to the above-described embodiment, the third processing unit 13 generates an estimation function Y'by Fourier transforming the generated composite signal x _m ', and estimates the distance of the Y'peak as the distance to the target. ..
The first signal processing unit 125 may perform the above-described processing for each receiving antenna and always adopt the estimation result of any one of the four receiving antennas, or the first signal processing unit 125 may always adopt the estimation result of the four receiving antennas. The estimation result of the receiving antenna having the maximum value may be adopted.

As described above, according to the high resolution of the target distance in step 3, the received signal x _n is the single wave signal x _n (k)'generated from K virtual single targets. Approximate by synthesis, for each single wave signal x _n (k)', a single wave signal of a non-existent non-existent receiving element is generated according to a predetermined phase difference, and the single wave signal x _n (k) 'And a composite signal x _m'combined with the single wave signal of the non-existent receiving element is generated. As a result, even if the received signal x _n includes signals from two or more targets, it is possible to generate a widened composite signal x _m '.
By Fourier transforming this widebanded composite signal x _m'to generate a distance distribution (estimation function), the ability to separate two or more peaks can be enhanced. Therefore, when estimating these peaks as the distances to two or more targets, the separability of the estimation of the distances from the targets can be enhanced.

In addition, in the high separation ability of the target distance in the procedure 3, the second processing unit 12 has K pieces for each of N times and MN non-existing times, as in the above-described modification 1. 'by combining the synthetic signal x _m of the n times and M-n pieces of imaginary _time' single wave signal x _{n (k)} when generating a K-number of single wave signal x _n ( Of the k)'and the corresponding K non-existent time single wave signals, i need not be used (i is an integer less than K).
According to this, as in the above-described modification 1, in step 3-1 it is appropriate that the first i times (i = 2 or so) in which a peak is likely to occur at an erroneous distance due to the interference of a plurality of signal waves. By not using the single-wave signal (which seems to be) in step 3-2, the occurrence of false peaks can be reduced.

Further, in order to increase the separation ability of the target distance in step 3, the steps 3-1 to 3-3 are performed again and then the steps 3-1 to 3-3 are performed again in the same manner as in the modification 2 described above. In step 3-1 of the first lap, a single wave signal x _n (k)'may be generated at the distance obtained in the first lap. Specifically, the first processing unit 11
(2) When the single wave signal x _n (k)'of N times is first generated, the distance to the target estimated using the combined signal x _m'already generated by the third processing unit 13 is set. On the basis of,
(2) For the single wave signal forming the estimated function Y (k) of the distance to the target already estimated in the generated estimation function Y (k), the single wave signal x _n (k) at N times. )'May be generated.
At this time, the first processing unit 11 may set the subtraction rate M to be large only at the beginning, and may greatly reduce the factors affected by the erroneous peak.
According to this, as in the above-mentioned modification 2, as a result of eliminating the influence of the erroneous peak in the second lap of the procedure 3-1 the component of the first i times lost by not using the first i times. Can be regained.

As explained above, even with the high separation ability of the target distance in step 3.
(I) It is possible to separate and detect a plurality of targets for a distance (ii) It has phase information (iii) It satisfies that it can be estimated at the same azimuth angle under two frequency conditions. Therefore, even in the radar device (position estimation device) 100 of the modified example 3, by combining the frequency shift method (procedure 4) and the high separation ability of the distance (procedure 3), in other words, the separation is detected by the distance. By combining the target information and the information on the phase difference generated by the frequency shift, the separability of a plurality of targets by the FFT processing can be improved, and the distance estimation accuracy by the two-frequency comparison processing can be improved.

(In case of pulse modulation)
For example, as shown in (A) of FIG. 24 (corresponding to (A) of FIG. 7 described above), the signal wave (Rx0 to Rx3) received by the receiving antenna is a pulse-modulated wave whose intensity changes in a pulse shape on the time axis. _Let x _n be the received signal at N frequencies obtained by dividing the frequency axis obtained by converting the time axis (N is an integer of 2 or more, and n is the number of N frequencies).
The first signal processing unit 125 in the radar device (position estimation device) 100 performs the following processing for each receiving antenna, that is, for each received signal x _n (Rx0 to Rx3) received by the receiving antenna.

<< Procedure 3-1 >>
Similar to the above-described embodiment, the first processing unit 11 converts the received signal _xn generated by receiving the signal wave by the receiving antenna into a single wave generated from K virtual single targets. Approximate as a composite wave signal of the signal x _n (k)'(K is an integer of 2 or more, and k is the number of K virtual single targets).

Specifically, the first processing unit 11 assumes that the received signals x _n of N frequencies are the combined signals x _n (0) of K single wave signals x _n (k)', and the following Is processed. That is, the first processing unit 11
(1) By Fourier transforming the combined signal x _n (k) of N frequencies, the distance distribution of the signal wave is generated as the estimation function Y (k) ((C) of FIG. 24: FIG. 7 described above. (B) equivalent)
(2) For the single-wave signal forming the peak of the generated estimation function Y (k), a single-wave signal x _n (k)'of N frequencies is generated.
(3) For each N frequencies, subtract M · x _n (k)'from the composite signal x _n (k) by multiplying the single wave signal x _n (k)' by a predetermined subtraction rate M. Then, the remaining composite signals x _n (k') of N frequencies are generated.
x _n (k') = x _n (k) -M · x _n (k)'
k'is a value obtained by incrementing k by 1.
The first processing unit 11 repeats the above processes (1) to (3) for the remaining combined signal x _n (k') until k'= K.
In this way, by incrementing k by 1 and repeating the processes (1) to (3) above for the combined signal x _n (k), K received signals x _n of N frequencies can be obtained. Can be approximated as a linear composite signal of the single wave signal x _n (k)'.

<< Procedure 3-2 >>
Similar to the above-described embodiment, the second processing unit 12 has MN non-existent frequencies for each K single-wave signals x _n (k)'according to a predetermined phase difference. Generate a single wave signal. The number of MN non-existent frequency single wave signals generated is more than twice the number of N frequency single wave signals x _n (k)'. As a result, the separation limit of the distances of the plurality of targets can be made smaller than the Rayleigh limit, and the distance estimation using the Fourier transform by the third processing unit 13 can be made highly separable.

The second processing unit 12 synthesizes K single-wave signals x _n (k)'for each of the existing N frequencies and the non-existing MN non-existing frequencies, as in the above-described embodiment. by, for generating a synthesized signal x _{m 'of} the N frequency and M-N pieces of imaginary frequencies are (M> N, m is N frequency and (M-N) pieces of imaginary frequencies It is the number of.).

<< Procedure 3-3 >>
Similar to the above-described embodiment, the third processing unit 13 generates an estimation function Y'by Fourier transforming the generated composite signal x _m ', and estimates the distance of the Y'peak as the distance to the target. ..
The first signal processing unit 125 may perform the above-described processing for each receiving antenna and always adopt the estimation result of any one of the four receiving antennas, or the first signal processing unit 125 may always adopt the estimation result of the four receiving antennas. The estimation result of the receiving antenna having the maximum value may be adopted.

As described above, according to the high resolution of the target distance in step 3, the received signal x _n is the single wave signal x _n (k)'generated from K virtual single targets. Approximate by synthesis, for each single wave signal x _n (k)', a single wave signal of a non-existent non-existent receiving element is generated according to a predetermined phase difference, and the single wave signal x _n (k) 'And a composite signal x _m'combined with the single wave signal of the non-existent receiving element is generated. As a result, even if the received signal x _n includes signals from two or more targets, it is possible to generate a widened composite signal x _m '.
By Fourier transforming this widebanded composite signal x _m'to generate a distance distribution (estimation function), the ability to separate two or more peaks can be enhanced. Therefore, when estimating these peaks as the distances to two or more targets, the separability of the estimation of the distances from the targets can be enhanced.

In addition, in the high resolution of the target distance in step 3, in the second processing unit, K singles are used for each of N frequencies and MN non-existent frequencies, as in the above-described modification 1. 'by combining the synthetic signal x _m of the n frequencies and M-n pieces of imaginary _frequencies' first wave signal x _{n (k)} when generating a K-number of single wave signal x _{n (k} )'And the corresponding K non-existent frequency single wave signals need not be used (i is an integer less than K).
According to this, as in the above-described modification 1, in step 3-1 it is appropriate that the first i times (i = 2 or so) in which a peak is likely to occur at an erroneous distance due to the interference of a plurality of signal waves. By not using the single-wave signal (which seems to be) in step 2, the occurrence of erroneous peaks can be reduced.

Further, in order to increase the separation ability of the target distance in step 3, the steps 3-1 to 3-3 are performed again and then the steps 3-1 to 3-3 are performed again in the same manner as in the modification 2 described above. In step 3-1 of the first lap, a single wave signal x _n (k)'may be generated at the distance obtained in the first lap. Specifically, the first processing unit 11
(2) Based on the distance to the target estimated using the combined signal x _m'already generated by the third processing unit when the single wave signal x _n (k)'of N frequencies is first generated. hand,
(2) N-frequency single-wave signals x _n (k)'for single-wave signals forming the already-estimated distance estimation function Y (k) in the generated estimation function Y (k). May be generated.
At this time, the first processing unit 11 may set the subtraction rate M to be large only at the beginning, and may greatly reduce the factors affected by the erroneous peak.
According to this, as in the above-mentioned modification 2, as a result of eliminating the influence of the erroneous peak in the second lap of the procedure 3-1 the component of the first i times lost by not using the first i times. Can be regained.

As explained above, even with the high separation ability of the target distance in step 3.
(I) It is possible to separate and detect a plurality of targets for a distance (ii) It has phase information (iii) It satisfies that it can be estimated at the same azimuth angle under two frequency conditions. Therefore, even in the radar device (position estimation device) 100 of the modified example 3, by combining the frequency shift method (procedure 4) and the high separation ability of the distance (procedure 3), in other words, the separation is detected by the distance. By combining the target information and the information on the phase difference generated by the frequency shift, the separability of a plurality of targets by the FFT processing can be improved, and the distance estimation accuracy by the two-frequency comparison processing can be improved.

In step 3, even if both the above-described embodiment, the high separation ability of the target azimuth of the modification 1 or 2 and the high separation ability of the target distance of the modification 2 are performed in order. Good.

(Modification example 4)
Compressed sensing (Orthogonal Matching Pursuit) technology may be used as the processing for increasing the resolution of the azimuth angle (or distance) in step 3 of the above-described embodiment and modification.
In the following, an example in which the compressed sensing technique is applied to the high-resolution processing of the azimuth angle in the procedure 3 of the above-described embodiment will be described. 25 and 26 are diagrams for explaining an example of high resolution processing of the azimuth angle (or distance) in the position estimation processing by the radar device (position estimation device) 100 according to the modification 4.

In the model shown in FIG. 25, estimates of N K sparse _{x 0} from M measurement signal y (M <N, K < M). For example, considering the example of the above-described embodiment, the beat signal for each target is estimated with y as the reception signal of the four receiving antennas and x _i0 and x _{iK at} 256 x _0s . In this case, M is the number of receiving antennas 4, N is 256, K is the number of targets 2, and the matrix A is 4 × 256.

For example, in this estimation, the following processing is performed.
0. Set x to a 0 vector. At this time, clears the list T to store the subscript i0~iK indicating the position value of x ₀ is not zero (list T is empty).
1. 1. Find the vector in the observation matrix that is closest (highly correlated) with yA (T) x (T). Add the column number of the vector to list T.
2. Find x (T) that minimizes the norm of yA (T) x (T). For example, it is obtained by the generalized inverse matrix.
3. 3. Repeat steps 1 and 2 above several times.
These processes approximate the process of step 3-1 described above, that is, the received signal received by the N receiving elements as a composite wave signal of the single wave signal reflected from the K virtual single targets. Corresponds to the processing to be performed.

After performing the above processing for frequency 1, for frequency 2, x (T) that minimizes the norm of yA (T) x (T) by using the above list T and the received signal y of frequency 2. Ask for. For example, use the generalized inverse matrix. When Ax is calculated using x obtained in this way and A expanded to M = 16, the process of step 3-2 described above, that is, for each K single-wave signal of MN non-existent receiving elements In addition, a single wave signal of MN non-existent receiving elements is generated according to a predetermined phase difference, and K single waves are generated for each of N receiving elements and MN non-existing receiving elements. This corresponds to a process of generating a composite signal of N receiving elements and MN non-existing receiving elements by synthesizing a one-wave signal. This makes it possible to estimate the same azimuth at both frequencies 1 and 2. (Since it is expected to be estimated at the same azimuth, frequencies 1 and 2 may be performed separately.)

In addition, instead of the complex model of FIG. 25, the real part and imaginary part separation model shown in FIG. 26 may be used.

As described above, even with the high resolution of the target azimuth (or distance) used by the compressed sensing technology in step 3.
(I) A plurality of targets can be separated and detected with respect to the azimuth angle (distance) (ii) The phase information is possessed (iii) The same azimuth angle can be estimated under two frequency conditions. Therefore, even in the radar device (position estimation device) 100 of the modified example 4, by combining the frequency shift method (procedure 4) and the high resolution of the azimuth (or distance) (procedure 3), in other words, By combining the target information separated and detected by the azimuth (or distance) and the phase difference information generated by the frequency shift, the separation ability of multiple targets by FFT processing is improved, and the distance estimation accuracy by two-frequency comparison processing is improved. Can be enhanced.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and various modifications and modifications can be made. For example, in the above-described embodiment, the position estimation device assuming the form of one transmitting element and four receiving elements has been illustrated, but the present invention has a plurality of target positions from the received signals from two or more receiving elements. It can be applied to a position estimation device that estimates.

The present invention can also be applied to a position estimation device in a device using two or more transmitting elements. For example, a form having two transmitting elements and four receiving elements is equivalent to a form having one transmitting element and eight receiving elements. As described above, the position estimation device described above uses not only the received signal x _n received by the physically existing receiving element but also a plurality of transmitting elements as the received signal x _n received by the N receiving elements. The position of the wave source may be estimated using the received signal of the receiving element that virtually exists.

The present invention can also be applied to a form in which a plurality of transmitting elements and one receiving element are used. For example, a form having four transmitting elements and one receiving element is equivalent to a form having one transmitting element and four receiving elements.
For example, a LiDAR device in the case where an optical array using a laser beam as a transmission wave is a transmission element corresponds to such a form. In such a LiDAR device, the increase in the number of transmitting elements is limited due to the attenuation of the laser beam due to the branching of the optical waveguide, so that the resolution is low. The present invention makes it possible to improve the separation ability of the LiDAR device by applying it to such a LiDAR device.

Further, in the above-described embodiment, a radar device has been exemplified as a position estimation device. However, the position estimation device of the present invention is not limited to this, and various types of signal waves having coherence (for example, electromagnetic waves (radio waves, terahertz waves, laser light, X-rays) and sound waves (ultrasonic waves)) are received. It can be applied to an apparatus (for example, LiDAR, sound source localization using a microphone array, millimeter wave body scanner, MRI, X-ray CT, optical CT, ultrasonic inspection machine, etc.).
In addition, in the millimeter wave body scanner, MRI, X-ray CT, optical CT, and ultrasonic inspection machine, it is easy to make a multi-array element and a synthetic aperture is possible. Therefore, a plurality of techniques different from the present invention may be used. It is possible to improve the separation ability of the incoming wave. Therefore, it is expected that the present invention is suitably used for sound source localization using a radar device, LiDAR, and a microphone array.

100 Radar device (position estimation device)
101 Control unit 110 Transmission control unit (transmission unit)
111 Transmitter (transmitter)
120 Received signal processing unit 121 Reception unit 125 1st signal processing unit 126 2nd signal processing unit 11 1st processing unit 12 2nd processing unit 13 3rd processing unit Tx transmitting antenna Rx receiving antenna

Claims (19)

A position estimator that estimates the positions of multiple targets.
A transmitter that transmits a signal wave changed to two frequencies with a predetermined frequency difference,
A receiving unit that receives a reflected wave whose signal wave is reflected by the plurality of targets and generates a beat signal obtained by frequency-converting the reflected wave at the frequency of the signal wave.
A first signal processing unit that generates a separated beat signal in which phase information is maintained and the azimuths or distances of the plurality of targets are separated from each other for each of the beat signals corresponding to the two frequencies.
A second signal processing unit that estimates the distance to the target as the position of the target from the predetermined frequency difference and the phase difference of the separated beat signals corresponding to the two frequencies.
A position estimation device. The predetermined frequency difference is Δf (Hz), the speed of light is c (m / s), the phase difference Δφ (rad) of the separated beat signals corresponding to the two frequencies, and the distance r (m) from the target. The second signal processing unit estimates the distance r (m) from the target by the following equation (1).
The position estimation device according to claim 1. The transmitter transmits a signal wave changed to at least three frequencies with the predetermined frequency difference.
The first signal processing unit generates the beat signal corresponding to the two frequencies of the predetermined frequency difference by the difference processing of the beat signal corresponding to the at least three frequencies.
The position estimation device according to claim 1 or 2. The transmission unit transmits the signal wave having coherence, and
The receiving unit receives the reflected wave by N receiving elements, and receives the reflected wave.
The first signal processing unit receives each of the beat signals corresponding to the two frequencies.
The reflected wave by the N receiving elements, where N is an integer of 2 or more, n is the number of the N receiving elements, K is an integer of 2 or more, and k is the number of K virtual single targets. With the first processing unit that approximates the received signal x _n generated by receiving the above as a composite wave signal of the single wave signal x _n (k)'reflected from the K virtual single targets. ,
M is an integer larger than N, m is the number of the N receiving elements and MN non-existing receiving elements, and a predetermined position is set for each of the K single wave signals x _n (k)'. A single wave signal of the MN non-existent receiving elements is generated according to the phase difference, and the K singles are generated for each of the N receiving elements and the MN non-existing receiving elements. 'by combining the synthetic signal x _m of the n number of receiving elements and the M-n pieces of imaginary receiving _{element-frequency} signal x _{n (k)} and a second processing unit for generating as said separating beat signal ,
Using the generated synthetic signal x _m ', a third processing unit that estimates the azimuth angle of the target as the position of the target, and
The position estimation device according to any one of claims 1 to 3. In the second processing unit, the number of single-wave signals generated by the MN non-existent receiving elements is twice or more the number of single-wave signals x _n (k)'of the N receiving elements. The position estimation device according to claim 4. The first processing unit assumes that the received signals x _n of the N receiving elements are the combined signals x _n (0) of the K single wave signals x _n (k)'.
(1) The azimuth-angle distribution of the reflected wave is generated as an estimation function Y (k) by Fourier transforming the combined signals x _n (k) of the N receiving elements.
(2) For the single-wave signal forming the peak of the generated estimation function Y (k), the single-wave signal x _n (k)'of the N receiving elements is generated.
(3) for each of said N number of receiving elements, wherein the composite signal _x n (k), the single wave signal _x n (k) 'to multiplied by the predetermined subtraction ratio _{M M · x n (k)} ' Is subtracted to generate the remaining composite signal _xn (k') of the N receiving elements.
Here, k'is a value obtained by incrementing k by 1.
By repeating the processes (1) to (3) above for the remaining combined signal x _n (k') until k'= K, the received signals x _n of the N receiving elements can be obtained. Approximate as a composite wave signal of the K single wave signals x _n (k)',
The position estimation device according to claim 4 or 5. The second processing unit sets i as an integer less than K, and outputs K single-wave signals x _n (k)'for each of the N receiving elements and the MN non-existing receiving elements. by combining, the _'when generating, the K single-wave signal x _{n (k)'} synthesized signal x _m of the n number of receiving elements and the M-n pieces of imaginary receiving elements and corresponding The position estimation device according to any one of claims 4 to 6, wherein a predetermined i of the single wave signals of the K non-existent receiving elements is not used. When (2) the single wave signal x _n (k)'of the N receiving elements is first generated, the first processing unit produces the combined signal x _m'already generated by the third processing unit. Based on the azimuth of the target estimated using, (2) a single wave forming the estimation function Y (k) of the already estimated azimuth of the target in the generated estimation function Y (k). For the signal, the single wave signal x _n (k)'of the N receiving elements is generated.
The position estimation device according to claim 6. The transmission unit transmits the signal wave having coherence, and
The receiving unit receives the reflected wave by N receiving elements, and receives the reflected wave.
The signal wave and the reflected wave are frequency-modulated waves whose frequency changes on the time axis, the received signal at N times obtained by dividing the time axis is _xn , N is an integer of 2 or more, and n is the N times. Number, K as an integer of 2 or more, k as the number of K virtual single targets
The received signal x _n generated by receiving the reflected wave by the receiving element is a composite wave signal of the single wave signal x _n (k)'generated from the K virtual single targets. The first processing unit that approximates as
M is an integer larger than N, m is the number of the N time and MN non-existent time, and each of the K single wave signals x _n (k)'has a predetermined phase difference. Correspondingly, the MN non-existent time single wave signals are generated, and the K single wave signals x _n (for each of the N time and the MN non-existent time). 'by combining the synthetic signal x _m of the N times and the M-N pieces of imaginary _time' k) and the second processing unit for generating as said separation beat signal,
Using the generated synthetic signal x _m ', a third processing unit that estimates the distance to the target as the position of the target, and
The position estimation device according to any one of claims 1 to 3. In the second processing unit, the number of generations of the MN non-existent time single wave signals is more than twice the number of the number of the N time single wave signals x _n (k)'. , The position estimation device according to claim 9. The first processing unit assumes that the received signals x _{n at the} N times are the combined signals x _n (0) of the K single wave signals x _n (k)'.
(1) By Fourier transforming the composite signal x _n (k) of the N time times, the distance distribution with the target is generated as an estimation function Y (k).
(2) For the single wave signal forming the peak of the generated estimation function Y (k), the single wave signal x _n (k)'at the above N times is generated.
(3) M · x _n (k)' obtained by multiplying the single wave signal x _n (k)' by a predetermined subtraction rate M from the combined signal x _n (k) at each of the N time times. Subtraction is performed to generate the remaining composite signal _xn (k') of the N times.
Here, k'is a value obtained by incrementing k by 1.
By repeating the processes (1) to (3) above for the remaining composite signal x _n (k') until k'= K, the received signals x _n at the N times are obtained. Approximate as a composite wave signal of K single wave signals x _n (k)',
The position estimation device according to claim 9 or 10. The second processing unit synthesizes the K single-wave signals x _n (k)'for each of the N times and the MN non-existent times, where i is an integer less than K. said K that by the _'when generating, the said K single wave signal x _{n (k)'} n number of times and the synthesized signal x _m of the M-n pieces of imaginary time and the corresponding The position estimation device according to any one of claims 9 to 11, which does not use a predetermined i of the non-existent time single wave signals. The first processing unit uses (2) the composite signal x _m'already generated by the third processing unit when first generating the single wave signal x _n (k)'at the N time. Based on the estimated distance to the target, (2) a single-wave signal forming the estimated function Y (k) of the already estimated distance to the target in the generated estimation function Y (k). Generates the single-wave signal x _n (k)'at the above N times.
The position estimation device according to claim 11. The transmission unit transmits the signal wave having coherence, and
The receiving unit receives the reflected wave by N receiving elements, and receives the reflected wave.
A pulse-modulated wave in which the signal wave and the reflected wave change in intensity in a pulse shape on the time axis, a received signal at N frequencies obtained by dividing the frequency axis obtained by converting the time axis is _xn , and N is an integer of 2 or more. Let n be the number of the N frequencies, K be an integer of 2 or more, and k be the number of K virtual single targets.
The received signal x _n generated by receiving the reflected wave by the receiving element is a composite wave signal of the single wave signal x _n (k)'generated from the K virtual single targets. The first processing unit that approximates as
M is an integer larger than N, m is the number of the N frequencies and MN non-existent frequencies, and each of the K single-wave signals x _n (k)'has a predetermined phase difference. Correspondingly, the MN non-existing frequency single wave signals are generated, and for each of the N frequencies and the MN non-existing frequencies, the K single wave signals x _n ( 'by combining the synthetic signal x _m of the N frequencies and the M-N pieces of imaginary _frequencies' k) and the second processing unit for generating as said separation beat signal,
Using the generated synthetic signal x _m ', a third processing unit that estimates the distance to the target as the position of the target, and
The position estimation device according to any one of claims 1 to 3. In the second processing unit, the number of MN non-existent frequency single wave signals generated is more than twice the number of the N frequency single wave signals x _n (k)'. The position estimation device according to claim 14. The first processing unit assumes that the received signals x _n of the N frequencies are the combined signals x _n (0) of the K single wave signals x _n (k)'.
(1) By Fourier transforming the composite signal _xn (k) of the N frequencies, the distance distribution with the target is generated as an estimation function Y (k).
(2) For the single wave signal forming the peak of the generated estimation function Y (k), the single wave signal x _n (k)'of the N frequencies is generated.
(3) For each of the N frequencies, M · x _n (k) ′ obtained by multiplying the single wave signal x _n (k) ′ by a predetermined subtraction factor M from the combined signal x _n (k) is obtained. Subtract to generate the remaining composite signal _xn (k') of the N frequencies.
Here, k'is a value obtained by incrementing k by 1.
By repeating the processes (1) to (3) above for the remaining combined signal x _n (k') until k'= K, the received signals x _n of the N frequencies are obtained. Approximate as a composite wave signal of K single wave signals x _n (k)',
The position estimation device according to claim 14 or 15. The second processing unit synthesizes the K single-wave signals x _n (k)'for each of the N frequencies and the MN non-existent frequencies, where i is an integer less than K. said K that by the _'when generating, the said K single wave signal x _{n (k)'} n number of frequency and the M-n pieces of the composite signal x _m of the non-existent frequencies and the corresponding The position estimation device according to any one of claims 14 to 16, which does not use a predetermined i of the non-existent frequency single wave signals of the above. The first processing unit uses (2) the composite signal x _m'already generated by the third processing unit when first generating the single wave signal x _n (k)'of the N frequencies. Based on the estimated distance to the target, (2) a single-wave signal forming the estimated function Y (k) of the already estimated distance to the target in the generated estimation function Y (k). Generates the single wave signal x _n (k)'of the N frequencies.
The position estimation device according to claim 13. A position estimation method that estimates the positions of multiple targets.
A signal wave changed into two frequencies with a predetermined frequency difference is transmitted,
The signal wave receives the reflected wave reflected by the plurality of targets, and generates a beat signal obtained by frequency-converting the reflected wave at the frequency of the signal wave.
For each of the beat signals corresponding to the two frequencies, a separated beat signal in which the phase information is maintained and the azimuth angles or distances of the plurality of targets are separated from each other is generated.
The distance to the target is estimated as the position of the target from the predetermined frequency difference and the phase difference of the separated beat signals corresponding to the two frequencies.
Position estimation method.