JP5114187B2 - Electronic scanning radar apparatus, received wave direction estimation method, and received wave direction estimation program - Google Patents

Description

  The present invention relates to an electronic scanning radar apparatus suitable for in-vehicle use, which uses a reflected wave from a target with respect to a radiated transmission wave and detects the target, and an arrival wave direction estimation program used therefor.

Conventionally, as an in-vehicle radar, an electronic scanning type radar using a method such as FMCW (Frequency Modulated Continuous Wave) radar, multi-frequency CW (Continuous Wave) radar, or pulse radar is used.
In each of the above radars, an array antenna arrival wave direction estimation method is used as a technique for detecting the direction of an incoming wave (or a received wave) from a target.
This incoming wave direction estimation method includes beam scanning method such as beamformer method and capon method, linear prediction method such as maximum entropy (MEM) method, minimum norm method, MUSIC (MUltiple SIgnal Classification) method, ESPRIT There is a null operation method called a super-resolution (high accuracy) algorithm such as (Estimation of Signal Parameters via Rotational Invariance Techniques) method (for example, see Non-Patent Documents 1 and 2).

In addition, the direction of arrival wave used for in-vehicle radar is estimated only by digital beam forming (DBF) of the beamformer method (for example, see Patent Document 1), or the detection accuracy of the direction of arrival of the received wave (or In recent years, a method combining DBF and the maximum entropy method (for example, see Patent Documents 2 and 3) has been performed in order to improve the target resolution.
In addition, super-resolution algorithms such as MUSIC are applied to in-vehicle radars. Compared to ordinary personal computers, microprocessors used for in-vehicle use have lower arithmetic processing functions, so estimation algorithms are simplified. Have been developed (for example, Patent Documents 4, 5, 6, and 7).

The super resolution algorithm such as MUSIC creates a correlation matrix from the received wave data of each array antenna, and detects the arrival direction of the received wave from the correlation matrix by eigenvalue calculation. Here, the eigenvalue calculation (calculation) is to obtain eigenvalues and eigenvectors.
At this time, as the white noise component of the correlation matrix is removed, the accuracy of detection of the arrival direction is improved. Therefore, the correlation matrix is created by ensemble averaging of received data.
For example, FMCW radar obtains as many samples as possible from the received beat signal data set (time-series data in a certain time interval that can be converted to frequency domain data) and uses an averaged correlation matrix. Become. The number of samples is called the number of snapshots (see, for example, Non-Patent Documents 1 and 2).

However, in the in-vehicle radar, the distance and relative speed with respect to the target constantly fluctuate. Therefore, even if the number of snapshots is increased vaguely, the detection accuracy is not always improved.
Conversely, in order to earn the number of snapshots during the control cycle (or detection cycle, eg, 100 ms) for detecting the target, the number of times the array antenna is multiplied by the number of snapshots at the same time as the frequency resolution processing for the received signal. There is a limit to the number of
In order to improve the detection accuracy without increasing the snapshot, in Patent Document 4, for example, a correlation matrix for each beat frequency at the previous (or last) control cycle is stored, and the current control cycle is stored. A method of weighted addition (weighted average) of the correlation matrix of the beat frequency where the target is present and the previous (or the previous) correlation matrix of the same beat frequency, or storing this weighted addition correlation matrix for each beat frequency In addition, there is described a method for further weighted addition of the correlation matrix of the beat frequency where the target exists in the current control cycle and the correlation matrix obtained by weighting and adding the same beat frequency.

Moreover, in the said patent document 5, it produces with the frequency which shows a peak value, and its vicinity (for example, +/- 2 frequency resolution part) from the same peak waveform in which the target exists in beat frequency. A method is described in which each correlation matrix is averaged to increase the number of snapshots.
In this Patent Document 5, it is described that a correlation matrix between frequency domains is averaged by using a past correlation matrix.
Patent Document 6 describes estimation of the arrival direction of a received wave by combining a roadside shape estimation method and a method of averaging current and past correlation matrices.

Patent Document 6 describes a method for determining the weighting coefficient (or forgetting coefficient: a constant indicating the degree of forgetting) in real time in order to average the current and past correlation matrices.
Kikuma Nobuyoshi, Adaptive Signal Processing with Array Antenna, Science and Technology Publishers, 1998 Nobuyoshi Kikuma, Adaptive Antenna Technology, Ohmsha, 2003 JP 2000-284044 A JP 2006-275840 A JP 2006-308542 A JP 2007-040806 A JP 2006-145251 A JP 2006-242695 A JP 2006-284182 A

  However, in the above-mentioned Patent Document 4, in the conventional correlation matrix averaging means, when the past correlation matrix is frequency-resolved into all beat frequencies, for example, Fourier transform is performed in 256 discrete times, 128 discrete frequencies are used. There is a problem that a large-capacity memory corresponding to “number of data of correlation matrix × total beat frequency” is required. In addition, since the beat frequency of the past correlation matrix is the same as the beat frequency of the current target, when the distance to the target is tracking at a constant distance, the detection data is averaged. On the other hand, if the distance to the target varies, there is a concern that the target may not exist at the previous frequency, and the data for detection will deteriorate.

Further, in Patent Document 5, as a conventional correlation matrix averaging method, the correlation matrix is always averaged between neighboring beat frequencies, and then averaged with the past correlation matrix.
For this reason, it is premised on that the signal intensity as the presence level of the target is shown in the frequency band including the nearby beat frequency, and can be applied only when the discrete frequency resolution of the beat frequency is very fine. In addition, since it is always performed in combination with the averaging of correlation matrices between neighboring beat frequencies, the arrival direction of the received wave is not detected using the past correlation matrix alone.
Furthermore, Patent Document 6 is based on the premise that the roadside shape estimation method is performed in the conventional correlation matrix averaging method as well as Patent Document 5, and the received wave using the past correlation matrix alone is used. The direction of arrival is not detected, and the processing is complicated.

  The present invention has been made in view of such circumstances, and an electronic scanning radar apparatus and a reception device that detect the arrival direction of a received wave with high accuracy by simple calculation using a past correlation matrix alone. An object is to provide a wave direction estimation program.

An electronic scanning radar apparatus according to the present invention is an electronic scanning radar apparatus mounted on a moving body, and includes a transmission unit that transmits a transmission wave and a plurality of antennas that receive a reflected wave from a target of the transmission wave. Receiving unit, a beat signal generating unit that generates a beat signal from the transmission wave and the reflected wave, and a frequency decomposition processing unit that calculates complex number data by frequency-decomposing the beat signal into a predetermined number of beat frequencies And a target detector for detecting a peak value from the intensity value of the beat frequency to detect the presence of a target, and a correlation for calculating a correlation matrix from each complex number data of a detected beat frequency at which the target is detected for each antenna The matrix calculator and the targets in the current detection cycle and the past detection cycle are related by distance and relative speed. A target link unit which characterize the correlation matrix for the current detection cycle targets, the correlation matrix filter unit for generating an average correlation matrix weighted averaging and correlation matrix of the same target in the past detection cycle associated, And an azimuth detector that calculates the arrival direction of the received wave from the average correlation matrix.

  In the electronic scanning radar apparatus according to the present invention, when the target connection processing unit associates the target in the detection cycle of the current and the past, the distance and the relative velocity obtained by the detection beat frequency of the current detection cycle are the past. Prediction is made based on the distance and relative speed obtained by the detection cycle, and whether or not there is an association between the targets is detected depending on whether the distance is included in a preset distance range and relative speed range.

  The electronic scanning radar apparatus according to the present invention further includes a storage unit for storing a distance, a relative velocity, and a correlation matrix corresponding to each of the associated targets as targets for one or a plurality of cycles in the past. The target concatenation processing unit generates an average correlation matrix by weighted averaging a correlation matrix between a target in the current detection cycle and a target in a plurality of time-series past detection cycles associated with the current target. In addition, the distance, relative velocity, and correlation matrix of the current target are associated with the distance, relative velocity, and correlation matrix of the associated past target, and are stored in the storage unit.

  The electronic scanning radar apparatus according to the present invention further includes a storage unit that stores one or a plurality of cycles of complex number data of a detected beat frequency corresponding to the associated target, and the current detection When a target of a past detection cycle associated with a target of the cycle is detected, a correlation matrix calculation unit calculates a correlation matrix from complex data of the past detection cycle, and the target connection processing unit The average correlation matrix is generated by weighting and averaging the correlation matrix of the target in the past and the past target associated with the current target, and complex data of the distance, relative velocity, and detected beat frequency of the associated current target is generated. , Target distance in relative past detection cycle, relative And to store in association with time and complex data.

The electronic scanning radar apparatus of the present invention further includes a DBF unit for performing digital beam forming in the antenna arrangement direction based on the complex data for each antenna and detecting the presence and direction of the target, and a current detection cycle. The lateral position is detected by digital beam forming from the beat frequency at, and the target is correlated in the detection cycle between the present and the past by the distance, relative speed, and direction.

  In the electronic scanning radar apparatus according to the present invention, the DBF unit performs digital beam forming using the complex number data to calculate spatial complex number data indicating the intensity of the spectrum for each angle channel, and the intensity of the spectrum of the adjacent angle channel is calculated. When the value exceeds a preset DBF threshold in the range of the preset number of angle channels, the presence of the target is detected (DBF detection target), and the spectrum intensity of the angle channel where the presence of the target is not detected is set to “0”. And a channel deletion unit that outputs as new spatial complex number data, and an IDBF unit that generates reproduced complex number data by inverse DBF of the new spatial complex number data, and the correlation matrix calculation unit includes: Calculating a correlation matrix from the reproduced complex data And features.

  In the electronic scanning radar apparatus of the present invention, when the channel deletion unit detects that there are a plurality of DBF detection targets, the spectrum is divided for each angular channel range corresponding to each DBF detection target, and the number of DBF detection targets is determined. Spatial complex number data is generated, and the IDBF unit performs inverse DBF on the spatial complex number data for each DBF detection target, thereby generating reproduction complex number data for each DBF detection target, and the correlation matrix calculation unit for each DBF detection target. A correlation matrix for each DBF detection target is calculated from the reproduction complex number data.

The electronic scanning radar apparatus according to the present invention is characterized in that the correlation matrix filter unit changes a weighting coefficient corresponding to the relative velocity for each target and performs weighted averaging.
In the electronic scanning radar apparatus of the present invention, the correlation matrix filter unit performs weighted averaging when the amount of change in the lateral position obtained from the past and present azimuth and distance exceeds a preset range. A weighting factor is changed for each target.

  The electronic scanning radar apparatus according to the present invention is characterized in that the number of past cycles used when the target connection processing unit averages is changed in accordance with the relative speed.

A reception wave direction estimation method of the present invention is a method for estimating a reception wave direction by controlling any one of the above-described electronic scanning radar devices mounted on a moving body, and a transmission process of transmitting a transmission wave from a transmission unit; A reception process in which a reception unit including a plurality of antennas receives a reflected wave from the target of the transmission wave; and a beat signal generation process in which a beat signal generation unit generates a beat signal from the transmission wave and the reflection wave; A frequency decomposition processing process in which the frequency resolution processing unit frequency-decomposes the beat signal into a predetermined number of beat frequencies to calculate complex number data, and a target detection unit detects a peak value from the intensity value of the beat frequency. A target detection process for detecting the presence of a target, and a correlation matrix calculation unit that combines a detected beat frequency at which the target is detected for each antenna. A correlation matrix calculation process for calculating a correlation matrix from each of the numerical data, a target connection processing unit in which the target connection processing unit associates targets in the current detection cycle and the past detection cycle with distance and relative speed, and a correlation matrix filter unit A correlation matrix filter process for generating an average correlation matrix by weighted averaging of a correlation matrix of a target of the current detection cycle and a correlation matrix of the same target in the associated past detection cycle ; And an azimuth detection process for calculating the arrival direction of the received wave from the matrix.

The received wave direction estimating program of the present invention is a program for causing a computer to control the operation of estimating the received wave direction by any one of the above-described electronic scanning radar devices mounted on a moving body. A transmission process for transmitting the signal, a reception process for receiving the reflected wave from the target of the transmission wave by a plurality of antennas, and a beat signal for generating a beat signal from the transmission wave and the reflected wave by the beat signal generation unit Generation processing, frequency decomposition processing in which the frequency decomposition processing unit frequency-decomposes the beat signal into beat frequencies of a predetermined number of decompositions to calculate complex number data, and the target detection unit obtains a peak value from the intensity value of the beat frequency A target detection process for detecting the presence of a target and a correlation matrix calculation unit for each antenna. Correlation matrix calculation processing for calculating a correlation matrix from each complex number data of the detected beat frequency at which the get is detected, and target connection in which the target connection processing unit associates the target in the current detection cycle and the past detection cycle with the distance and relative velocity Correlation matrix filter processing in which the correlation matrix filter unit generates a weighted average of the correlation matrix of the target in the current detection cycle and the correlation matrix of the same target in the associated past detection cycle . And an azimuth detection unit that calculates an arrival direction of a received wave from the average correlation matrix.

  As described above, according to the present invention, the target concatenation processing unit performs correlation matrix averaging after associating the same target in the present and the past. In addition, eigenvalue calculation and spectrum calculation (for example, MUSIC) in the azimuth detection processing performed in the subsequent stage can be executed with high accuracy, and the final calculation is performed as compared with the case where the calculation is performed using the correlation matrix of only the current time. Recognition performance in the distance and direction of the target can be improved.

  In addition, according to the present invention, since a plurality of correlation matrices or complex number data is stored in the target unit and averaged using all of them, the final target can be obtained regardless of the variation in the distance to the target. The recognition performance in the distance and direction can be improved.

  Further, according to the present invention, the target distance varies because the number of detection cycles of the correlation matrix used for averaging the correlation matrix is arbitrarily changed based on the relative speed with the target in units of targets. In order to reduce the number of detection cycles when the target is moving and to increase the number of detection cycles when the distance to the target is stable, the appropriate filter characteristics are correlated to each target according to the relative speed state. The matrix filter unit can be provided, and the recognition performance in the final target distance and direction can be improved.

Further, according to the present invention, the weighting factor used for averaging of the correlation matrix is varied based on the relative speed with respect to the target in units of targets, so that depending on the state of the relative speed for each target. Therefore, the correlation matrix filter unit can have appropriate filter characteristics, and the recognition performance in the final target distance and direction can be improved.
According to the present invention, when the amount of change in the lateral position obtained from the azimuth and distance in the past and the present exceeds a preset range, the weighting coefficient when weighted average is changed for each target. That is, the correlation matrix filter unit can have an appropriate filter characteristic that reduces the number of past cycles to be averaged, or substantially reduces the number of connections by changing the weighting factor, so that the final target distance and Recognition performance in the direction can be improved.

  Further, according to the present invention, in order to store the complex number data of the target frequency, the past data stored in the storage unit is the number of channels of the original plurality of antennas per target × 2 (the complex number of the real part and the imaginary part). Data), and the storage capacity can be reduced as compared with the case where data is simply stored as a correlation matrix.

  Further, according to the present invention, since the DBF means for detecting the azimuth by the DBF from the frequency-resolved beat frequency is provided, not only the prediction range from the distance and the relative speed but also the azimuth range is included. Thus, it is possible to improve the accuracy of association of the current and past correlation matrices.

  Further, according to the present invention, the spectrum is divided for each angular channel range corresponding to the DBF detection target and each spatial complex number data is generated, and the IDBF unit performs inverse DBF on the spatial complex number data for each DBF detection target. Since the correlation matrix for each DBF detection target is calculated from the reproduction complex number data, the correlation matrix used when performing eigenvalue calculation includes only the component of the received wave that arrives for each DBF detection target. On the other hand, even if received waves arrive from more than that number of targets, it is possible to improve the direction and distance recognition performance with high accuracy without erroneously calculating eigenvalues for each DBF detection target.

<First Embodiment>
Hereinafter, an electronic scanning radar apparatus (FMCW millimeter wave radar) according to a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration example of the embodiment.
In this figure, the electronic scanning radar apparatus according to the present embodiment includes a receiving antenna 11-1n, a mixer 211-2n, a transmitting antenna 3, a distributor 4, a filter 51-5n, a SW (switch) 6, an ADC (A / D). Converter) 7, control unit 8, triangular wave generation unit 9, VCO 10, and signal processing unit 20.
The signal processing unit 20 includes a memory 21, a frequency separation processing unit 22, a peak detection unit 23, a peak combination unit 24, a distance detection unit 25, a speed detection unit 26, a pair determination unit 27, a correlation matrix calculation unit 28, a correlation matrix filter. A unit 29, an orientation detection unit 30, a target determination unit 31, and a target connection processing unit 32.

Next, the operation of the electronic scanning radar apparatus according to the present embodiment will be described with reference to FIG.
The receiving antennas 11 to 1n receive a reflected wave, i.e., a received wave, which is reflected from the target, with the transmitted wave reflected by the target.
Each of the mixers 21 to 2n mixes a transmission wave transmitted from the transmission antenna 3 and a signal obtained by amplifying the reception wave received at each of the reception antennas 11 to 1n by an amplifier to cope with each frequency difference. Generated beat signal.
The transmission antenna 3 transmits a transmission signal obtained by frequency-modulating the triangular wave signal generated by the triangular wave generation unit 9 in a VCO (Voltage Controlled Oscillator) 10 to the target as a transmission wave.
The distributor 4 distributes the frequency-modulated transmission signal from the VCO 10 to the mixers 21 to 2n and the transmission antenna 3.

Each of the filters 51 to 5n performs band limitation on the beat signals of Ch1 to Chn corresponding to the receiving antennas 11 to 1n generated in the mixers 21 to 2n, respectively, and beats band-limited to the SW (switch) 6 Output a signal.
The SW 6 sequentially switches the beat signals of Ch1 to Chn corresponding to the receiving antennas 11 to 1n that have passed through the filters 51 to 5n in response to the sampling signal input from the control unit 8, and performs ADC (A / A D converter) 7.
The ADC 7 performs A / D conversion on the beat signals of Ch1 to Chn corresponding to the receiving antennas 11 to 1n, which are input from the W6 in synchronization with the sampling signal, into a digital signal in synchronization with the sampling signal. The signal is converted and sequentially stored in the waveform storage area of the memory 21 in the signal processing unit 20.
The control unit 8 is constituted by a microcomputer or the like, and controls the entire electronic scanning radar apparatus shown in FIG. 1 based on a control program stored in a ROM (not shown).

<Principle to detect distance, relative speed, angle (azimuth)>
Next, the principle of detecting the distance, relative speed, and angle (azimuth) between the electronic scanning radar apparatus and the target used in the signal processing unit 20 in this embodiment will be briefly described with reference to FIG.
FIG. 2 shows a transmission signal obtained by frequency-modulating the signal generated in the triangular wave generation unit 9 of FIG. 1 by the VCO 10 and a state in which the transmission signal is reflected by the target and input as a reception signal. The example of FIG. 2 shows a case where there is one target.
As can be seen from FIG. 2A, a reception signal that is a reflected wave from the target is received with a delay in the right direction (time delay direction) in proportion to the distance from the target with respect to the signal to be transmitted. Furthermore, it varies in the vertical direction (frequency direction) with respect to the transmission signal in proportion to the relative speed with the target. Then, after the frequency conversion (Fourier transform, DTC, Hadamard transform, wave red transform, etc.) of the beat signal obtained in FIG. 2A, as shown in FIG. Each of the ascending region and the descending region has one peak value. Here, in FIG. 2A, the horizontal axis represents frequency and the vertical axis represents intensity.

The frequency resolution processing unit 22 performs discrete frequency analysis on the rising portion (upward) and the falling portion (downward) of the triangular wave from the sampled data of the beat signal stored in the memory 21 by, for example, Fourier transform. Convert frequency.
As a result, as shown in FIG. 2B, a graph of the signal level for each beat frequency obtained by frequency decomposition is obtained in the rising portion and the falling portion.
The peak detector 23 detects the peak value from the signal level for each beat frequency shown in FIG. 2B, detects the presence of the target, and the beat frequency of the peak value (both ascending and descending parts). Is output as the target frequency.

Next, the distance detection unit 25 calculates a distance from the target frequency fu of the rising portion and the target frequency fd of the falling portion input from the peak combination unit 24 by the following formula.
r = {C · T / (2 · Δf)} · {(fu + fd) / 2}
Further, the speed detection unit 26 calculates a relative speed from the target frequency fu of the rising portion and the target frequency fd of the falling portion input from the peak combination unit 24 by the following formula.
v = {C / (2.f0)}. {(fu-fd) / 2}
In the formula for calculating the distance r and the relative velocity v,
C: speed of light Δf: frequency modulation width of triangular wave f0 : Triangular wave center frequency T: Modulation time (rising part / falling part)
fu: Target frequency in the rising part fd : Target frequency in the descending part

Next, the receiving antennas 11 to 1n in the present embodiment are array antennas arranged at intervals d as shown in FIG.
The receiving antennas 11 to 1n are incident on an incoming wave (incident wave, that is, a transmission wave transmitted from the transmission antenna 3) that is incident from an angle θ direction with respect to an axis perpendicular to the plane on which the antennas are arranged. The reflected wave from the target) is input.
At this time, the incoming waves are received at the same angle by the receiving antennas 11 to 1n.
A phase difference “dn ₋₁ · sin θ” obtained by the same angle, for example, the angle θ and the interval d between the antennas, is generated between the adjacent receiving antennas.
Using the above phase difference, the angle θ is detected by signal processing such as digital beam forming (DBF) or super-resolution algorithm that further Fourier transforms the antenna frequency in the time direction for each antenna. can do.

<Signal Processing for Received Wave in Signal Processing Unit 20>
Next, the memory 21 stores time series data (ascending portion and descending portion) obtained by A / D converting the received signal in the waveform storage area by the ADC 7 for each of the antennas 11 to 1n. . For example, when 256 samples are sampled in each of the ascending portion and the descending portion, data of 2 × 256 × the number of antennas is stored in the waveform storage area.
The frequency resolution processing unit 22 converts each beat signal corresponding to each Ch1 to Chn (each antenna 11 to 1n) into a frequency with a preset resolution by, for example, Fourier transformation, and the like, and a frequency point indicating the beat frequency. The complex number data of the beat frequency is output. For example, if each of the ascending and descending parts has 256 sampled data for each antenna, it is converted into a beat frequency as complex frequency domain data for each antenna, and 128 in each of the ascending and descending parts. Complex number data (2 × 128 pieces × number of antennas data). The beat frequency is indicated by a frequency point.
Here, the difference between the complex number data for each antenna is only the phase difference depending on the angle θ, and the absolute values (reception intensity, amplitude, etc.) on the complex plane of each complex number data are equivalent.

The peak combination unit 24 exceeds the preset numerical value from the peak in the signal intensity (or amplitude, etc.) using the complex number data, with the peak values of the intensity rising and falling areas of the triangular wave of the beat frequency subjected to frequency conversion. By detecting a beat frequency having a peak value, the presence of a target for each beat frequency is detected, and a target frequency is selected.
Therefore, the peak detection unit 23 converts the complex number data of any antenna or the sum of complex number data of all antennas into a frequency spectrum, so that each peak value of the spectrum has a target depending on the beat frequency, that is, the distance. Can be detected as By adding complex number data of all antennas, noise components are averaged, and the S / N ratio is improved.

The peak combination unit 24 combines the beat frequency and the peak value shown in FIG. 4 input from the peak detection unit 23 with the beat frequency and the peak value of the ascending region and the descending region in a matrix form in a brute force manner, That is, all beat frequencies in the ascending region and the descending region are combined and sequentially output to the distance detection unit 25 and the speed detection unit 26. Here, in FIG. 4, the horizontal axis indicates the frequency point of the beat frequency, and the vertical axis indicates the signal level (intensity).
The distance detection unit 25 calculates the distance r to the target based on a numerical value obtained by adding the beat frequencies of the combinations of the rising region and the falling region that are sequentially input.
Further, the speed detection unit 26 calculates the relative speed v with respect to the target based on the difference between the beat frequencies of the combinations of the ascending region and the descending region that are sequentially input.

  The pair determination unit 27 generates the table shown in FIG. 5 based on the input distance r, relative speed v, and descending and ascending peak value levels pu and pd, and each of the ascending region and the descending region corresponding to each target is generated. Appropriate combinations of peaks are determined, the pairs of peaks in the ascending region and the descending region are determined as a table shown in FIG. 6, and the target group number indicating the determined distance r and relative velocity v is output to the target determining unit 31 . In FIG. 6, distance, relative speed, and frequency point (ascending region and descending region) are stored corresponding to the target group number. The tables of FIGS. 5 and 6 are stored in the internal storage unit of the pair determination unit 27. Here, since the direction of each target group is not determined, the position in the horizontal direction parallel to the arrangement direction of the receiving antennas 11 to 1n with respect to the vertical axis with respect to the arrangement direction of the antenna array in the electronic scanning radar apparatus is Not decided.

Here, for example, the pair determination unit 27 gives priority to the value predicted in the current detection cycle from the distance r and the relative speed v with each target finally determined in the previous detection cycle. It is also possible to use a technique such as selecting a combination of.
Further, the correlation matrix calculation unit 28 selects the beat frequency frequency-resolved by the frequency decomposition processing unit 22 based on the frequency point of the beat frequency in the target group whose combination is determined by the pair determination unit 27, A correlation matrix corresponding to the beat frequency of any one of the descending parts (the descending part in the present embodiment) is generated and output to the correlation matrix filter unit 29 and the target connection processing unit 32.

Next, the target connection processing unit 32 combines the distance r, the relative speed v, the frequency point of FIG. 6 input from the pair determination unit 27 and the target in the past detection cycle stored in the memory 21, The past correlation matrix for each target is output to the correlation matrix filter 29.
The correlation matrix filter unit 29 is obtained by multiplying each of the past correlation matrix input from the target concatenation processing unit 32 and the current correlation matrix by a weighting coefficient, and then averaging these correlation matrices. The averaged correlation matrix is output to the direction detection unit 30.
The azimuth detection unit 30 detects the azimuth of the corresponding target from the averaged correlation matrix using MUSIC or the like of the super-resolution algorithm, and outputs it to the target determination unit 31.
Further, the target connection processing unit 32 attaches the identification information of the distance, the relative speed, and the azimuth output from the target determination unit 31 to the current correlation matrix, and stores them in the memory 21.

<Super-resolution algorithm for estimating direction of arrival of received waves>
Next, the super-resolution algorithm for estimating the arrival direction of the received wave in the correlation matrix calculation unit 28, the correlation matrix filter unit 29, and the azimuth detection unit 30 will be described with reference to FIG. 7 taking MUSIC as an example. FIG. 7 is a flowchart showing a general MUSIC process flow. Since the MUSIC process itself is generally used (for example, Non-Patent Documents 1 and 2, or Patent Documents 3 to 6), only the necessary parts in this embodiment will be described.
The frequency resolution processing unit 22 reads the beat signal based on the received wave stored in the memory 21 (step S101), and converts the frequency of the beat signal for each antenna (step S102).
Then, as already described, the correlation matrix calculation unit 28 calculates the frequency-resolved complex frequency domain data (hereinafter referred to as complex number data) corresponding to the target frequency point of the descending domain whose combination is determined by the pair determination unit 27. The frequency matrix is selected and read from the frequency resolution processing unit 22, and a correlation matrix indicating the correlation for each antenna is generated in the descending region (step S103).

For example, there are methods shown in FIGS. 8A and 8B in the generation of the correlation matrix in the correlation matrix calculation unit 28 in step S103, which will be briefly described below.
In the method shown in FIG. 8A, the correlation matrix calculation unit 28 generates a correlation matrix (complex correlation matrix) as complex data (step S103_1), and only forward spatial average (Forward spatial average method) or forward / backward. Processing is performed by spatial averaging (Forward-Backward spatial averaging method) (step S103_2).
The spatial average is obtained by dividing the number of antennas in the original array of receiving antennas into subarrays having a smaller number of antennas and averaging the subarrays. The basic principle of this spatial averaging method is that the phase relationship of correlated waves differs depending on the reception position. Therefore, if the correlation matrix is obtained by appropriately moving the reception point, the correlation of the correlation interference wave is suppressed by the average effect. . In general, a plurality of subarrays having the same arrangement are extracted from the entire array of receiving antennas without moving the array of receiving antennas, and the correlation matrix is averaged.

For example, as shown in FIG. 9, considering an array of nine receiving antennas 11 to 1n (n = 9), the correlation matrix calculation unit 28 uses the following correlation matrix CR of the following equation (1). against ^f _f, seeking backward correlation matrix ^CR _{b f} of the rear (2), seeking backward correlation matrix ^CR _{b f} of the rear (2), (1) the correlation matrix of the equation (2 ): The element corresponding to the backward correlation matrix in the equation is expressed as follows: CR ^fb _f = (CR ^f _f + CR ^b _f ) / 2
Is the average processing of the elements in the front / rear.
In this way, the correlation matrix CR ^fb _f obtained by the forward / backward averaging process is divided into subarrays and averaged to obtain a correlation matrix Rxx used for estimating the arrival direction of the received wave. That is, the correlation matrix obtained by the forward / backward spatial averaging process is expressed by the following equation.
Rxx = (CR ^fb1 _f + CR ^fb2 _f + CR ^fb3 _f ) / 3
Here, the correlation matrix calculation unit 28 divides the array of the nine receiving antennas 11 to 19 into three subarrays of seven antennas 11 to 17, 12 to 18, and 13 to 19, and calculates the matrix of each subarray. The correlation matrix Rxx is obtained by averaging the corresponding elements.

On the other hand, in the case of the forward spatial average, the matrix of V ₁₁ to V ₉₉ may be the matrix of the formula (1) from W ₁₁ to W ₉₉ , so that V ₁₁ = (W ₁₁ + W ₉₉₎ / 2 The average of each element indicated by is unnecessary.
In applications that estimate the direction of arrival of received waves in radar, the transmitted waves transmitted from all incoming received waves are reflected waves reflected by the target, so the received wave data received for each antenna is strongly correlated. Correlation will be shown. Therefore, the result of eigenvalue calculation in the subsequent stage does not appear correctly. Therefore, the spatial average has the effect of suppressing the cross-correlation, extracting the auto-correlation, and correctly estimating the arrival wave direction.

Next, the correlation matrix calculation unit 28 performs unitary conversion for converting the correlation matrix of the complex number data spatially averaged by the above-described processing into a real number correlation matrix.
Here, by converting to a real number correlation matrix, the eigenvalue calculation with the heaviest calculation load in the subsequent steps can be calculated only for the real number, and the calculation load can be greatly reduced.
On the other hand, FIG. 8B is a type in which eigenvalue calculation in the next step is also performed with complex numbers without performing conversion to a real correlation matrix by unitary transformation as shown in FIG. 8A.
In step S103, each correlation matrix Rxx obtained in S103_3 in FIG. 8A and S103_2 in FIG. 8B is further set based on the maximum value of the correlation matrix (or the diagonal component of the correlation matrix). The element value may be normalized (= divided by the maximum value).

Next, the azimuth detecting unit 30 calculates the eigenvalues of the correlation matrix Rxx (actually, the correlation matrix after correlation matrix filter processing described later) obtained in step S103 and the corresponding eigenvectors,
Rxxe = λe
Are calculated as an eigenvalue λ and an eigenvector e that satisfy the eigen equation (step S104).
Then, the azimuth detecting unit 30 estimates the number of incoming waves necessary for removing the signal component vector from the obtained eigenvalue λ (step S105).
Next, the azimuth detection unit 30 creates an angle spectrum by performing an inner product calculation of a vector including only a noise component, excluding the signal vector, and a direction vector for each azimuth angle set in advance. (Step S106). Thereby, directivity null can be matched with the arrival direction of a received wave.

Then, the azimuth detecting unit 30 detects a peak exceeding a preset threshold from the angle spectrum, thereby detecting the peak and calculating the incoming wave direction (angle θ) (step S107).
Further, the azimuth detecting unit 30 is positioned laterally with respect to the vertical axis of the antenna array in the electronic scanning radar apparatus by the angle (= the arrival direction of the received wave) and the distance calculated by the distance detecting unit 25. It can also be converted.
The above is a standard MUSIC, but in the MUSIC spectrum calculation in step S106, a method called Root-MUSIC that finds a solution from a root of a polynomial instead of a type that searches by a direction vector can also be used.

Further, after step S107 in FIG. 7, processing for calculating received power and deleting unnecessary waves (unnecessary received wave data) may be added.
That is, the azimuth detection unit 30 compares the power appearing in the diagonal component of the matrix S in the following equation with a preset threshold value, detects whether the power exceeds the threshold value, When it exceeds the threshold, it is determined that the received wave is necessary. On the other hand, when the power is equal to or lower than this threshold, it is determined that the received wave is unnecessary.
S = (A ^H A) ⁻¹ A ^H (Rxx−σ ² I) A (A ^H A) ⁻¹
Here, S is a correlation matrix of a received wave signal, A is a direction matrix, A ^H is a conjugate transpose matrix of A, I is a unit matrix, Rxx is a correlation matrix calculated by the correlation matrix calculation unit 28, and σ ² is noise The variance of the vector.
By adding the processing of the received power calculation and unnecessary wave deletion described above, when estimating the number of received waves in the estimation of the received wave number in step S105, it is possible to delete received waves that arrive unnecessarily by this processing. it can.

<Current and Past Averaging Processing in Correlation Matrix Filter Unit 29>
Next, the process of averaging the correlation matrix with the present and the past in the present embodiment will be described. This averaging process is a process mainly performed by the correlation matrix calculation unit 28, the correlation matrix filter unit 29, and the target connection processing unit 32 in FIG.
In order to perform the averaging process in the correlation matrix filter unit 29, the target connection processing unit 32 is predicted from the current target group (t) and the determined past target data for each target in the table shown in FIG. The following processing is performed to connect the target (t) and the targets (t-1, t-2, t-3) determined in the past.

In FIG. 10, t-1 is the result of the detection cycle one cycle before (immediately before), t-2 is the result of the detection cycle two cycles before, and t-3 is the result of three cycles before.
As a result of each detection cycle, the distance r, vertical position long_d (perpendicular to the antenna arrangement direction), and lateral position late_d (position parallel to the antenna arrangement direction) are determined for each determined target. , Relative velocity to the target (velo) (ie, v), down peak frequency point f_dwn, and down peak frequency correlation matrix mat_dwn (ie, Rxx) are stored in the memory 21 in the table format of FIG. In the case of using downstream data, more precisely, the storage area of mat_dwn is larger than the others, but is the same for convenience of illustration in the table). Here, the vertical position long_d and the horizontal position late_d of the target are obtained from the angle with respect to the target (the angle in the arrival direction of the received wave) and the distance r. When the angle is θ and the distance is r, the vertical position long_d is calculated by r · cos θ, and the horizontal position is calculated by long_d · sin θ.

Further, the target connection processing unit 32 determines the distance r, the vertical position long_d, and the horizontal position of each target in the current cycle from the target distance r, the vertical position long_d, the horizontal position late_d, and the relative velocity velo that have been determined in the past. Predict late_d, relative speed, and peak frequency points. For example, in the prediction of the vertical position long_d, the horizontal position late_d, and the peak frequency point, a movable range in the time after the detection cycle period is obtained based on the previous distance r, the vertical position long_d, the horizontal position late_d, and the relative speed. The relative speed can be predicted by calculating the slope of the change in the relative speed value transition in the past several cycles.
For example, the target connection processing unit 32 sets a movable range that is set in advance corresponding to the distance r, the vertical position long_d, the horizontal position late_d, the peak frequency point, and the relative speed, which are predicted from the results determined in the past. A frequency point range and a relative speed range are provided, and a connection is made based on whether or not each value calculated in the current cycle falls within the range. When the value is outside the range, it is determined that the target is different.

Then, in the table of FIG. 10, the target connection processing unit 32 moves the result of t−2 to t−3 when the target in the current detection cycle is associated with the past target, and sets the result of t−1 to t− -2 to move the result of the current detection cycle to the result of t-1, and calculate the prediction result of the next cycle.
On the other hand, the target connection processing unit 32 sets a current target that is not linked to any past target result as a new target, and performs the direction estimation without filtering with the past correlation matrix. It is output to the direction detection unit 30.
Further, when there is a past target that is not associated with the current target group result, the target connection processing unit 32 clears all the information on the past target.
Therefore, when the target enters a distance affected by multipath and a detection cycle in which no peak is detected at the beam frequency is reached, the filter effect using the result of the past target group is reset. In the case of the present embodiment shown in FIG. 10, the results of the targets of the past three detection cycles are stored in the memory 21.

Further, as another example, the target connection processing unit 32 may determine that a past target result that has been determined is a predetermined cycle even when a past target that is not associated with a target in the current detection cycle is detected. You may make it last only the number.
And the target connection process part 32 updates also the prediction result estimated from the past result one by one, and even if the target in the present detection cycle is not detected by the influence of multipath, it is tied after the next detection cycle. In this case, past data other than the number of cycles in which no peak is detected due to the influence of multipath can be used for the filter processing.
In addition, as in the extrapolation method in tracking, it is also possible to continue the presence state of the target by using the prediction result as the result in the current detection cycle in the detection cycle times where the peak value is not detected.

Then, as shown in FIG. 11, the correlation matrix filter unit 29 performs an averaging process of the correlation matrix Rxx in the process of S103_3 in FIG. 8A, and a correlation matrix for estimating the arrival direction of the received wave. Is a process performed when generating
Here, the correlation matrix filter unit 29 calculates one past detection cycle in the table of FIG. 10 stored in the memory 21 after the current real correlation matrix calculation (unitary transformation, step S103_3_1) calculated from the complex number data in the current detection cycle. A weighted average process with a correlation matrix of minutes or more is performed (step S103_3_2).

Alternatively, it is stored in the memory 21 in the state of the complex correlation matrix in step S103_1 in FIG. 8A, and the above-described averaging process is performed before the spatial averaging process in step S103_2 and the unitary transformation in step S103_3. Can obtain the same value, but since the amount of data stored in the memory 21 increases, a method of storing and using a unitary transformed real number correlation matrix is preferable.
For example, FIG. 12 shows an example in which a 5 × 5 correlation matrix (real number) subarrayed by the spatial averaging process is created.
In this example, the past detection cycle results stored in the memory 21 are the results up to three cycles before, and can be expressed by the following equations.
Rxx ′ (t ′) = K1 · Rxx (t) + K2 · Rxx (t−1)
+ K3.Rxx (t-2) + K4.Rxx (t-3)

When the weighting coefficient K1 (current) = K2 (one cycle before) = K3 (two cycles before) = K4 (three cycles before) = 0.25 in the above equation, the average of the four correlation matrices is simply obtained. .
It is also possible to change the weighting factor for each cycle. For example, the weighting factor can be increased as it is closer to the current cycle, or cycles that are not included in the average can be multiplied by a zero factor. In particular, the number of past cycles is not limited, but considering the effect of the filter and the capacity of the memory, the value up to three previous cycles in the present embodiment is considered an appropriate value.
The correlation matrix filter unit 29 outputs the above-described correlation matrix Rxx ′ to the direction detection unit 30 as a correlation matrix used for estimating the arrival direction of the received wave.

On the other hand, when the averaging process is performed using the complex correlation matrix illustrated in FIG. 8B, the process of the step illustrated in FIG. 13 is performed in step S103_1.
The correlation matrix filter unit 29 calculates the current complex correlation matrix calculated from the complex number data of the current detection cycle (step S103_1_1), and then weights the complex correlation matrix of one past cycle or more stored in the memory 21. An averaging process is performed (step S103_1_2). In FIG. 13, the correlation matrix filter processing is performed with the complex correlation matrix before spatial averaging as an embodiment different from the example of FIG. 11, but of course, this is also after spatial averaging shown in step S103_2 of FIG. If the correlation matrix filter processing is performed using the complex number correlation matrix, the amount of calculation and the amount of data stored in the memory can be reduced.
Here, when the relative speed with respect to the target is very high, a change in the distance for each detection cycle of the target becomes large.
For this reason, the range (= distance range) of the beat frequency filtered by the correlation matrix filter unit 29 is widened, and the change in the angle θ may be large between the cycles of the connected targets.

  As a countermeasure in this case, as shown in FIG. 14, when connecting the correlation matrix of the same target in the past detection cycle, the correlation matrix to be stored is determined by determining the beat frequency point range that can be connected from the result of the current detection cycle. It is possible to select the past number of cycles to be used for averaging, or to change the weighting factor to substantially reduce the number of connections.

  FIG. 14B is an example in which the target is approaching due to a high relative speed. In this case, since the beat frequency peak moves quickly, the cycle of t−3 is outside the range of data to be averaged. The past detection cycle number of the correlation matrix stored as the table of FIG. 10 is left as it is, the past cycle number used for averaging is selected, or the weighting factor is variable (for example, the weighting factor of the correlation matrix is “ The number of connections can be substantially reduced.

Further, a threshold value may be provided for the relative speed itself between the result in the target in the past detection cycle and the result in the target in the current detection cycle.
In this case, the correlation matrix filter unit 29 calculates the average when any of the relative speeds of the result in the target in the past detection cycle and the result in the target in the current detection cycle is equal to or higher than the set relative speed. Alternatively, the number of past cycles may be reduced, or the number of connections may be substantially reduced by changing the weighting factor.

  Further, as yet another configuration of the correlation matrix filter unit 29, the change of the lateral position value determined in the target of the past detection cycle is directly calculated, and it is detected whether or not it is larger than a preset specified value, When it is detected that the number is large, the number of past cycles to be averaged may be reduced, or the weighting coefficient may be varied to substantially reduce the number of connections. For example, a threshold Δlate_d is provided for each cycle in the current cycle and the past cycle. For example, when a lateral movement of Δlate_d or more is detected between t−1 and t−2, t−2 and t It is conceivable to set the weight coefficient of the data -3 to 0. After the azimuth is detected by the azimuth detection unit 30, the target determination unit 31 stores the target distance, vertical position, horizontal position, relative speed, down peak frequency point, and correlation matrix via the target connection processing unit 32. As described above, the table shown in FIG. 10 is stored as t-1 information for the next detection cycle, and the t-3 information is deleted.

Further, the signal processing unit 40 shown in FIG. 15 has a configuration corresponding to the signal processing unit 20 of FIG. 1, and in the result of each target group shown in the table of FIG. The complex number data may be stored. The same reference numerals are assigned to configurations that perform similar processing, and only differences from the configuration in FIG. 1 will be described.
The frequency decomposition processing unit 42 outputs frequency data, which is complex number data after frequency decomposition, to the target connection processing unit 52. The frequency resolution processing unit 42 is the same as the frequency resolution processing unit 22 shown in FIG.

Then, the target concatenation processing unit 52 replaces the complex number data corresponding to this correlation matrix with the correlation matrix in the distance, vertical position, horizontal position, relative speed, downlink peak frequency point and correlation matrix determined in the target determination unit 31. Store in table 10. The target connection processing unit 52 is the same as the target connection processing unit 32 shown in FIG.
With the configuration shown in FIG. 15 described above, the data stored in the memory 21 is smaller than the correlation matrix. However, when the correlation matrix filter unit 29 averages the current correlation matrix and the past correlation matrix, the correlation matrix filter unit 29 recalculates the correlation matrix Rxx from the complex number data in the past detection cycle result. There is a need. FIG. 16 shows the flow of processing corresponding to FIG. 11, and FIG. 17 shows the flow of processing corresponding to FIG. 13. In this case, FIG. 16 is also similar to the example of FIG. If the averaging process using the correlation matrix filter is performed before the calculation, the amount of calculation may be small.

<Second Embodiment>
Next, an electronic scanning radar apparatus according to a second embodiment of the present invention will be described with reference to FIG. FIG. 18 is a block diagram illustrating a configuration example of an electronic scanning radar apparatus according to the second embodiment.
In the second embodiment, as in the first embodiment, the direction estimation is performed only by the super-resolution algorithm. The same components as those in the first embodiment shown in FIG. 1 are denoted by the same reference numerals, and only differences from the first embodiment will be described below.
The frequency resolution processing unit 22B converts the beat signals of the ascending region and the descending region for each antenna into complex number data, and outputs a frequency point indicating the beat frequency and complex number data to the peak detecting unit 23B.
Then, the peak detection unit 23B detects the peak value of each of the ascending region and the descending region and the frequency point where the peak value exists, and outputs the frequency point to the frequency resolution processing unit 22B.
Next, the frequency resolution processing unit 22B outputs the corresponding complex number data for each of the ascending region and the descending region to the next correlation matrix calculating unit 28.
The correlation matrix calculation unit 28 generates a correlation matrix from the input complex number data.

This complex number data becomes the respective target groups (beat frequencies having peaks in the ascending region and the descending region) in the ascending region and the descending region.
In the target connection processing unit 32B, since it is necessary to link the target determined in the past with both the upstream and downstream target groups, the memory 21 stores a table shown in FIG. The table shown in FIG. 19 includes the frequency points (peak frequencies) in the ascending region (up) and the descending region (down) and the frequency points in both the ascending region and the descending region in each target group in addition to the configuration of FIG. The correlation matrix corresponding to is stored.
The target connection unit 32B performs a connection process between the current detection cycle and the past detection cycle by the same process as the target connection unit 32 of FIG.

The correlation matrix filter unit 29 averages the correlation matrix in the current detection cycle and the correlation matrix in the past detection cycle in each of the ascending region and the descending region, and outputs the result to the azimuth detecting unit 26.
Next, the direction detection unit 30 detects the angle θ for each of the correlation matrix of the ascending region and the correlation matrix of the descending region, and outputs it to the peak combination unit 24B as a table shown in FIG.
Then, the peak combination unit 24B performs combinations having the same angle based on the information in the table shown in FIG. 20, and combines the beat frequencies of the rising region and the falling region with the distance detection unit 25 and the speed detection unit 26. Output to.

Similar to the first embodiment, the distance detection unit 25 calculates the distance based on the beat frequencies of the ascending region and the descending region of the combination.
Moreover, the speed detection unit 26 calculates the relative speed based on the beat frequencies of the ascending region and the descending region of the combination, as in the first embodiment.
Here, each of the distance detection unit 25 and the speed detection unit 26 does not need to filter the values of the distance and the relative speed by averaging the current detection cycle and the past detection cycle as in the direction detection. Therefore, the calculation is performed with a combination of the rising region and the falling region of the beat frequency of the current detection cycle.
The target determination unit 31B determines the correlation matrix of the ascending region and the descending region, the frequency point, the distance, and the relative speed in the ascending region and the descending region as the current state.
Then, the target connection processing unit 32B receives the frequency points of the ascending region and the descending region, the correlation matrices of the ascending region and the descending region, the distance and the vertical position for each target, which are input from the target determination unit 31B. The lateral position and the relative speed are stored in the table of FIG. 19 by the same processing as in the first embodiment.

Also in this embodiment, it is possible to store not the correlation matrix but the complex data of the beat frequency at which the peak value is detected in the table of FIG. Here, in the present embodiment, not only the complex number data corresponding to the falling region in the first embodiment, but also the complex number data in both the rising region and the falling region are associated with the frequency points of the respective beat frequencies. It memorize | stores in the table of the said FIG.
In the configuration for storing the complex number data, when the correlation matrix filter unit 29 performs the averaging process of the correlation matrix of the current detection cycle and the correlation matrix of the past detection cycle for each target group, the correlation matrix calculation unit 28 calculates the correlation matrix from the complex number data of the past detection cycle read from the memory 21.

<Third Embodiment>
Next, an electronic scanning radar apparatus according to a third embodiment of the present invention will be described with reference to FIG. FIG. 21 is a block diagram illustrating a configuration example of an electronic scanning radar apparatus according to the third embodiment.
In the third embodiment, unlike the first embodiment, the direction is first estimated using DBF (Digital Beam Forming) having a resolution lower than that of a super resolution algorithm such as MUSIC, and the correlation is averaged later. This is a configuration in which azimuth estimation from a matrix is performed by a super-resolution algorithm. The same components as those in the first embodiment shown in FIG. 1 are denoted by the same reference numerals, and only differences from the first embodiment will be described below.
As shown in FIG. 21, a DBF processing unit 40 is provided between the frequency decomposition processing unit 22 and the peak detection unit 23 in the first embodiment of FIG. The point which detects the azimuth | direction which arrives differs from 1st Embodiment.

Similarly to the first embodiment, the frequency decomposition processing unit 22 performs frequency decomposition (time-axis Fourier transform) on the input beat signal, and sends the frequency point indicating the beat frequency and the complex number data to the DBF processing unit 40. Output.
Next, the DBF processing unit 40 performs Fourier transform on the input complex number data in the antenna arrangement direction, that is, performs spatial axis Fourier transform.
Then, the DBF processing unit 40 calculates spatial complex number data for each angle channel corresponding to the angle, that is, corresponding to the angular resolution, and outputs it to the peak detection unit 23 for each beat frequency.

Accordingly, the spectrum indicated by the spatial complex number data (beat frequency unit) for each angle channel output from the DBF processing unit 40 depends on the arrival direction estimation of the received wave by the beam scanning resolution.
In addition, since the Fourier transform is performed in the antenna arrangement direction, the same effect as adding complex number data between angle channels can be obtained, and the S / N ratio is improved for complex number data for each angle channel. Therefore, the accuracy in detecting the peak value can be improved in the same manner as in the first embodiment.
Both the complex number data and the spatial complex number data described above are calculated in both the rising and falling regions of the triangular wave, as in the first embodiment.

Next, after the processing by the DBF processing unit 40, the peak detection unit 23 detects a peak for each angle channel based on the DBF result, and sends the detected peak value of each channel to the next peak combination unit 24 as an angle channel. Output every time (in the case of spatial axis Fourier transform with 16 resolutions, 15 angle channels).
Similarly to the first embodiment, the peak combination unit 24 combines the beat frequency having the peak value in the ascending region and the descending region with the peak value, and sends it to the distance detection unit 25 and the speed detection unit 26 for each angle channel. Output to.

Then, the pair determination unit 27 generates the table of FIG. 5 for each angle channel based on the distance r and the relative speed v sequentially input from the distance detection unit 25 and the speed detection unit 26, respectively, in the first embodiment. In the same manner as described above, an appropriate combination of peaks in the ascending region and the descending region corresponding to each target is determined for each angle channel. Here, in the resolution in the DBF, since the target indicates the existence over a plurality of angle channels, the ascending region and the descending region for each angle channel are considered for each angle channel in consideration of the coincidence with neighboring angle channels (matrix). Appropriate combinations of peaks can be made. Then, the pair of peaks in the ascending region and the descending region is determined, and the target group number indicating the determined distance r and relative speed v is output to the target determining unit 31, and the table shown in FIG. 22 is created.
Further, since the pair determination unit 27 can obtain not only the distance r and the relative speed v but also the information on the angle channels of the respective targets, the pair determination unit 27 can obtain the vertical position and the horizontal position. The table shown in FIG. 22 is generated, which includes the results corresponding to each target group of the current detection cycle, including the vertical position and the horizontal position.

Then, the target linking unit 32 uses the information in the table of FIG. 22 to perform the process of associating the target in the current detection cycle with the target in the past detection cycle of FIG. Since the vertical position and the horizontal position are used in addition to the distance, the relative speed, and the peak frequency point, it is possible to perform the linking process with higher accuracy.
Furthermore, the reliability of direction detection is improved by estimating with the AND theory of the azimuth information from the azimuth detection unit 30 and the azimuth information from the DBF, and the azimuth information is shared, for example, the angular resolution at a short distance Can be rough, so that the angle information of DBF can be used.

Also in this embodiment, it is possible to store not the correlation matrix but the complex data of the beat frequency at which the peak value is detected in the table of FIG.
In the configuration for storing the complex number data, as in the first embodiment, the correlation matrix filter unit 25 averages the correlation matrix of the current detection cycle and the correlation matrix of the past detection cycle for each target group. When performing the processing, the correlation matrix calculation unit 24 calculates the correlation matrix from the complex number data of the past detection cycle read from the memory 21.

<Fourth Embodiment>
Next, an electronic scanning radar apparatus according to a fourth embodiment of the present invention will be described with reference to FIG. FIG. 23 is a block diagram illustrating a configuration example of an electronic scanning radar apparatus according to the fourth embodiment.
In the fourth embodiment, unlike the first embodiment, the azimuth estimation is first performed using DBF (Digital Beam Forming) having a lower resolution than the super resolution algorithm such as MUSIC, and the target angle range is narrowed down. IDBF (inverse DBF, that is, inverse space axis Fourier transform) is performed to return to complex data on the time axis, and the accuracy of azimuth estimation performed by a super resolution algorithm performed later is improved. The same components as those in the third embodiment shown in FIG. 21 are denoted by the same reference numerals, and only differences from the third embodiment will be described below.
In this embodiment, a Ch (channel) deletion unit 41 and an IDBF processing unit 43 are added to the third embodiment.

As in the third embodiment, the DBF processing unit 40 performs spatial axis Fourier transform, and outputs spatial complex number data to the peak detection unit 23 and also outputs it to the Ch deletion unit 41.
Here, as shown in FIG. 24A, the DBF processing unit 40 performs a spatial axis Fourier transform in the arrangement direction of the receiving antennas in the present embodiment, for example, with a resolution of 16 points, resulting in 15 angles. A spectrum in a channel angle unit is generated and output to the Ch deletion unit 41.
Then, the Ch deletion unit 41 continuously adjoins the spectrum level of the spatial complex data corresponding to the peak frequency point (for example, the descending portion) of the DBF target determined by the pair determination unit 27 within a preset angle range. In addition, it is detected whether or not a preset DBF threshold level is exceeded, a process for replacing the spectrum of the angle channel that does not exceed the DBF threshold with “0”, and narrowed-down spatial complex data is output.
In the processing described above, for example, the Ch deleting unit 41 has one or more targets in the range when adjacent four-angle channels are continuously at a level exceeding the DBF threshold as shown in FIG. As a result, the spectrum of these angle channels is left, and the intensities of the spectra of other angles are replaced with “0”.

Then, the IDBF processing unit 42 narrows down the spectrum, that is, leaves only the data of the angle channel region that continuously exceeds the DBF threshold in the set number of angle channels, and replaces the intensity of other regions with “0”. The spatial complex number data is subjected to inverse spatial axis Fourier transform, converted back to time axis complex number data, and output to the correlation matrix calculating unit 28.
Then, since the correlation matrix calculation unit 28 calculates the correlation matrix from the input complex number data, it is possible to obtain a correlation matrix with good orthogonality by removing roadside objects and the like and reducing noise components. FIG. 24 (c) creates a correlation matrix by using the above method for the target group at the DBF resolution shown in FIG. 24 (b) (actually, there is a possibility that there are two or more targets). This is an example in which the target is further separated by the super-resolution algorithm.
As shown in FIG. 25 (a), when receiving a reception wave including reflection components from a plurality of target groups, the spatial complex data output from the DBF processing unit 40 includes DBFs in continuous angle channels. There will be multiple angular channel ranges that exceed the level.

Then, when the spectrum level of the adjacent angle channel continuously exceeds the DBF threshold level in the set angle channel range in the input spatial complex data, the Ch deletion unit 41 determines the angle channel that has exceeded Each region is extracted, the intensity of the spectrum other than the angle channel region is replaced with “0”, and separate spatial complex number data identified in the angle channel region as shown in FIGS. 25 (b) and 25 (c) Divide into
Here, as in the third embodiment, the pair determination unit 27 obtains the distance, the relative speed, the vertical position, and the horizontal position, and outputs the distance, the relative speed, and the horizontal position to the Ch deletion unit 41 and the target connection processing unit 32.
The Ch deletion unit 41 selects spatial complex number data corresponding to the frequency point of the DBF target, performs the above-described Ch deletion, and outputs the data to the IDBF processing unit 42.

Then, the IDBF processing unit 42 performs inverse spatial Fourier transform on the input spatial complex number data, and outputs the obtained time axis complex number data to the correlation matrix calculation unit 28.
As a result, the correlation matrix calculation unit 28 calculates a correlation matrix from the input complex number data, and outputs the correlation matrix to the correlation matrix filter unit 29 as a correlation matrix in the current detection cycle.
The target connection processing unit 32 extracts the correlation matrix of the past detection cycle corresponding to the input distance, relative speed, and vertical position and horizontal position from the table of FIG. 10 in the memory 21 and outputs it to the correlation matrix filter unit 29. .
The correlation matrix filter unit 29 performs an averaging process on the input correlation matrix of the current detection cycle and the correlation matrix (data subjected to IDBF processing in the past) associated with the past detection cycle. The converted correlation matrix is output to the direction detection unit 30.

With the above-described processing, the detection direction range can be narrowed when calculating the spectrum in the MUSIC of the azimuth detecting unit 30, and the resolution can be further increased compared to the first to third embodiments.
Furthermore, with the configuration described above, the azimuth detector 33 virtually receives the received wave divided into the reflection components for each target group in the correlation matrix used for eigenvalue calculation. Even if received waves including reflection components from many targets larger than the number of receiving antennas and sub-arrays are received, calculation can be performed without error in eigenvalue calculation.

Also in the present embodiment, it is possible to store not the correlation matrix but the IDBF complex number data of the beat frequency at which the peak value is detected, instead of the correlation matrix.
In the configuration for storing the complex number data, as in the first embodiment, the correlation matrix filter unit 25 averages the correlation matrix of the current detection cycle and the correlation matrix of the past detection cycle for each target group. When performing the processing, the correlation matrix calculation unit 24 calculates the correlation matrix from the complex number data of the past detection cycle read from the memory 21.

The first to fourth embodiments have been described based on the configuration example used for the FMCW radar shown in FIG. 1, but can be applied to other antenna configurations of the FMCW scheme.
Also, the present invention can be applied to other systems other than the FMCW system such as multi-frequency CW and pulse radar.
Furthermore, in the present embodiment, the super resolution algorithm MUSIC has been described as an example of the azimuth detection unit. Similarly, a correlation matrix (or covariance matrix) is created, and detection accuracy increases as white noise is removed in this portion. As long as the algorithm is based on the principle of improving, it can be applied to the present invention.

  A program for realizing the functions of the signal processing unit 20 in FIGS. 1, 15, 18, 21, and 23 is recorded on a computer-readable recording medium, and the program recorded on the recording medium is recorded. The signal processing for detecting the azimuth from the received wave may be performed by being read into the computer system and executed. Here, the “computer system” includes an OS and hardware such as peripheral devices. The “computer system” includes a WWW system having a homepage providing environment (or display environment). The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Further, the “computer-readable recording medium” refers to a volatile memory (RAM) in a computer system that becomes a server or a client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. In addition, those holding programs for a certain period of time are also included.

  The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

1 is a block diagram illustrating a configuration example of an electronic scanning radar apparatus according to a first embodiment of the present invention. It is a conceptual diagram explaining the production | generation of the beat signal in the rising area and falling area of a triangular wave with a transmission wave and a received wave. It is a conceptual diagram explaining the received wave in a receiving antenna. It is the result of frequency-decomposing the beat signal, and is a graph showing the beat frequency (horizontal axis) and its peak value (vertical axis). It is a table | surface which shows the matrix of the beat frequency of the raise area | region in the combination part 24, and the fall area | region, and the distance and relative velocity in the intersection of the matrix, ie, the combination of the beat frequency of an raise area | region and a fall area. It is a table which shows the distance and relative speed for every target in the present detection cycle. It is a flowchart for demonstrating the process of MUSIC. It is a flowchart which shows the substep performed in step S103 of the flowchart of FIG. It is a conceptual diagram explaining the process at the time of calculating the spatial average of a correlation matrix. It is a conceptual diagram which shows the table structure in which the correlation matrix was described corresponding to the distance and relative speed of the past detection cycle used when matching with the present detection cycle and the past detection cycle. It is a flowchart which shows the substep of step S103_3 in Fig.8 (a). It is a conceptual diagram explaining the averaging process of the present detection cycle and the past detection cycle (plurality). It is a flowchart which shows the substep of step S103_1 in FIG.8 (b). It is a conceptual diagram explaining the detection cycle number of matching of a detection cycle. It is a block diagram which shows the structural example of the electronic scanning radar apparatus by the modification of the 1st Embodiment of this invention. It is a flowchart which shows the substep of step S103_3 in Fig.8 (a). It is a flowchart which shows the substep of step S103_1 in FIG.8 (b). It is a block diagram which shows the structural example of the electronic scanning radar apparatus by the 2nd Embodiment of this invention. It is a conceptual diagram showing a table configuration in which a correlation matrix is described in association with the distance, the vertical position and the horizontal position, and the relative speed of the past detection cycle, which is used when associating the current detection cycle and the past detection cycle. is there. It is a table which shows a response | compatibility with each angle for every target in the present detection cycle, and a frequency point. It is a block diagram which shows the structural example of the electronic scanning radar apparatus by the 3rd Embodiment of this invention. It is a table which shows the distance for every target in the present detection cycle, a vertical position, a horizontal position, and a relative speed. It is a block diagram which shows the structural example of the electronic scanning radar apparatus by the 4th Embodiment of this invention. It is a conceptual diagram explaining the process of the intensity | strength of the spectrum in each angle channel. It is a conceptual diagram explaining the process of the intensity | strength of the spectrum in each angle channel.

Explanation of symbols

11, 1 n, receiving antenna 21, 2 n, mixer 3, transmitting antenna 4, distributor 51, 5 n, filter 6, SW
7 ... ADC
8 ... Control unit 9 ... Triangle wave generation unit 10 ... VOC
DESCRIPTION OF SYMBOLS 20 ... Signal processing part 21 ... Memory 22, 42 ... Frequency decomposition processing part 23 ... Peak detection part 24 ... Peak combination part 25 ... Distance detection part 26 ... Speed detection part 27, 27B ... Pair determination part 28 ... Correlation matrix calculation part 29 ... Correlation matrix filter unit 30 ... Direction detection unit 31, 31B ... Target determination unit 32,52 ... Target connection processing unit 40 ... DBF processing unit 41 ... Ch deletion unit 43 ... IDBF processing unit

Claims (12)

An electronic scanning radar device mounted on a moving body,
A transmission means for transmitting a transmission wave;
A receiving unit comprising a plurality of antennas for receiving reflected waves from the target of the transmission wave;
A beat signal generation unit that generates a beat signal from the transmission wave and the reflected wave;
A frequency decomposition processing unit that calculates complex number data by frequency-decomposing the beat signal into beat frequencies of a predetermined number of decompositions;
A target detection unit for detecting the peak value from the intensity value of the beat frequency and detecting the presence of the target;
A correlation matrix calculating unit that calculates a correlation matrix from each complex number data of a detected beat frequency at which the target is detected for each antenna;
A target linking processor for associating targets in current and past detection cycles with distance and relative speed;
A correlation matrix filter unit that generates an average correlation matrix that weights and averages the correlation matrix of the target of the current detection cycle and the correlation matrix of the same target in the associated past detection cycle;
An electronic scanning radar apparatus comprising: an azimuth detecting unit that calculates an arrival direction of a received wave from the average correlation matrix.   When the target connection processing unit associates the targets in the current and past detection cycles, the distance and relative speed obtained by the detection beat frequency of the current detection cycle are the distance and relative speed obtained by the past detection cycle. 2. The electronic scanning radar apparatus according to claim 1, wherein the presence / absence of association between the targets is detected based on whether they are included in a preset distance range and relative velocity range. A storage unit that stores a distance, a relative velocity, and a correlation matrix corresponding to each of the associated targets as targets for one or more cycles in the past;
The target concatenation processing unit generates an average correlation matrix by weighted averaging a correlation matrix between a target in a current detection cycle and a target in a plurality of time series past detection cycles associated with the current target. 3. The current target distance, relative velocity, and correlation matrix are stored in the storage unit in correspondence with the past target distance, relative velocity, and correlation matrix associated with each other. An electronic scanning radar device according to claim 1. A storage unit that stores complex data of the detected beat frequency corresponding to the associated target for one or a plurality of cycles, respectively.
When the target of the past detection cycle that is associated with the target of the current detection cycle is detected, the correlation matrix calculation unit calculates the correlation matrix from the complex number data of the past detection cycle,
The target concatenation processing unit generates an average correlation matrix that is a weighted average of a correlation matrix of a target in a current detection cycle and a past target associated with the current target, and a distance between the associated current targets. The complex data of the relative velocity and the detected beat frequency are stored in association with the target distance, the relative velocity, and the complex data in the associated past detection cycle. Electronic scanning radar device. A DBF unit that performs digital beam forming in the antenna arrangement direction based on the complex number data for each antenna, and detects the presence and orientation of the target;
5. The azimuth is detected by digital beam forming from the beat frequency in the current detection cycle, and the association of the target in the detection cycle between the present and the past is performed by the distance, the relative velocity, and the azimuth. An electronic scanning radar device according to claim 1. The DBF unit performs digital beam forming using the complex number data to calculate spatial complex number data indicating the intensity of the spectrum for each angle channel, and the range of the angle channel number in which the intensity of the spectrum of the adjacent angle channel is set in advance. When the preset DBF threshold value is exceeded, the presence of the target is detected (DBF detection target), the spectrum intensity of the angle channel where the presence of the target is not detected is replaced with “0”, and new spatial complex data is obtained. The channel deletion section to output,
An IDBF unit for generating reproduced complex data by performing inverse DBF on the new spatial complex data; and
The electronic scanning radar apparatus according to claim 5, wherein the correlation matrix calculation unit calculates a correlation matrix from the reproduced complex number data. When the channel deletion unit detects that there are a plurality of DBF detection targets, the spectrum is divided for each angle channel range corresponding to each DBF detection target, and spatial complex number data of the number of DBF detection targets is generated.
The IDBF unit performs inverse DBF on the spatial complex number data for each DBF detection target, thereby generating reproduction complex number data for each DBF detection target,
7. The electronic scanning radar apparatus according to claim 6, wherein the correlation matrix calculation unit calculates a correlation matrix for each DBF detection target from reproduction complex number data for each DBF detection target.   The electronic scanning type according to any one of claims 2 to 7, wherein the correlation matrix filter unit changes a weighting factor corresponding to the relative speed and weighted average for each target. Radar device.   The correlation matrix filter unit changes a weighting factor for weighted averaging for each target when the amount of change in lateral position obtained from the past and present azimuth and distance exceeds a preset range. The electronic scanning radar apparatus according to claim 2, wherein the electronic scanning radar apparatus is characterized in that   8. The electronic scanning radar apparatus according to claim 2, wherein a past cycle number used when the target connection processing unit averages is changed in correspondence with the relative speed. It is a received wave direction estimation method by the electronic scanning radar apparatus in any one of Claims 1-10 mounted in a moving body,
A transmission process of transmitting a transmission wave from the transmission means;
A reception process in which a reception unit including a plurality of antennas receives a reflected wave from the target of the transmission wave;
A beat signal generation process in which a beat signal generation unit generates a beat signal from the transmission wave and the reflected wave;
A frequency resolving process in which the frequency resolving unit frequency-decomposes the beat signal into beat frequencies of a predetermined resolving number to calculate complex data;
A target detection process in which the target detection unit detects the peak value from the intensity value of the beat frequency and detects the presence of the target;
A correlation matrix calculation process in which a correlation matrix calculation unit calculates a correlation matrix from each complex number data of a detected beat frequency at which the target is detected for each antenna;
A target connection processing unit in which a target connection processing unit associates the target in the current detection cycle and the past detection cycle with the distance and the relative speed;
A correlation matrix filter process in which a correlation matrix filter unit generates an average correlation matrix by weighted averaging the correlation matrix of the target of the current detection cycle and the correlation matrix of the same target in the associated past detection cycle;
A method for controlling an electronic scanning radar apparatus, comprising: an azimuth detecting process in which an azimuth detecting unit calculates an arrival direction of a received wave from the average correlation matrix. A program for causing a computer to control an operation of estimating a received wave direction by the electronic scanning radar apparatus according to any one of claims 1 to 10, which is mounted on a moving body,
A transmission process in which the transmission means transmits a transmission wave;
A reception process in which a reception unit receives a reflected wave from a target of the transmission wave by a plurality of antennas;
Beat signal generation processing in which a beat signal generation unit generates a beat signal from the transmission wave and the reflected wave;
A frequency resolving process in which a frequency resolving unit frequency-decomposes the beat signal into beat frequencies of a predetermined resolving number to calculate complex data;
A target detection process in which the target detection unit detects the peak value from the intensity value of the beat frequency and detects the presence of the target;
A correlation matrix calculating unit that calculates a correlation matrix from each complex number data of a detected beat frequency at which the target is detected for each antenna;
A target connection processing unit in which the target connection processing unit associates the target in the current detection cycle and the past detection cycle with the distance and the relative speed;
A correlation matrix filter processing in which a correlation matrix filter unit generates an average correlation matrix obtained by weighted averaging the correlation matrix of the target of the current detection cycle and the correlation matrix of the same target in the associated past detection cycle;
A received wave direction estimation program, comprising: an azimuth detecting unit that calculates a direction of arrival of a received wave from the average correlation matrix.

Images