KR20060035804A - Radar device - Google Patents

Abstract

A radar device can separate directions of a plurality of targets obtained by a combination of antenna beams with a high accuracy. The radar device includes: a direction calculator for calculating a primary direction which is the direction of the target from a combination of characteristic amounts calculated by a signal detector from the reception wave of at least two beams partially overlapped among the beams emitted in the plurality of directions; and a direction integrator for calculating the integrating direction which is the true target direction when a plurality of primary directions calculated by the direction calculator exist, the calculation being performed according to the primary direction belonging to an area in which distribution of the plurality of primary directions is equal to or above a predetermined density.

Description

Radar Device {RADAR DEVICE}

The present invention relates to a radar device, and more particularly, to measuring the direction in which the target exists using the radar.

In general, in the radar system, the direction in which the target exists can be obtained by obtaining the direction in which the received signal can be detected. The precision of measuring the direction of a target improves so that beam width is narrow. However, in order to narrow the beam width by making the radar transmission wavelength constant, an antenna having a larger aperture diameter is required. Therefore, instead of a method in which measurement accuracy depends on the beam width, a method of measuring a target direction using information such as amplitude or phase difference between received signals obtained from a plurality of beams having slightly different beam directions is conventionally used. It is becoming. According to this method, an angular resolution higher than the measurement accuracy determined from the beam width can be achieved.

As an example of such a method, sequential-lobing and monopulse methods are known. In these methods, first, beams in two adjacent directions are selected from a plurality of beam directions, and a difference between the amplitude and phase of a received signal observed from these two beams (called a Δ signal) and sum (Σ signal) are obtained. ) Next, the ratio of the Δ signal and the Σ signal is calculated. If this ratio is called Δ / Σ value, the Δ / Σ value corresponds only to the angle of the target, so that the target direction can be estimated from the Δ / Σ value.

However, in these methods, the number of targets is limited to one. In other words, when a plurality of targets exist in the same beam, there is a problem that the direction cannot be calculated accurately. For example, as shown in FIG. 17, beams radiated from the antenna 100 with respect to directions 101 to 105 (directions 101 and 102, directions 102 and 103, directions 103 and 104, and direction 104). And 105 are adjacent to each other, and a radar system for estimating the direction of the target 110 using the beams in each direction is called beams 101 to 105. From the respective combinations of the beams 1 and 2, the beams 2 and 3, the beams 3 and 4, and the beams 4 and 5, a certain angle value is calculated regardless of the presence or absence of a true target. Arrows 111, 112, 113, and 114 shown in FIG. 18 are examples of directions (images) of angle values calculated based on differences between received signals obtained from the beams 101 to 105 in FIG. In the figure, the direction 111 is obtained from the combination with the beams 101 and 102 and the direction 112 is obtained from the combination with the beams 102 and 103. Further, the direction 113 is obtained from the combination with the beams 103 and 104 and the direction 114 is obtained from the combination with the beams 104 and 105. The direction 113 and the direction 114 have a correlation with the received wave reflected by the transmission of the transmission wave by the beams 103 to 105 to the target 110, but the direction 111 and the direction 112 are the target ( 110 is a false image that is irrelevant and does not correspond to a true goal.

Here, if there is only one target, the virtual may be rejected based on the amplitude or power of the received signal. However, when there are a plurality of targets, the correlation between the direction obtained by the combination of the beams and the target becomes complicated, and the rejection by a simple threshold value is required. It can't cope with.

As a method of measuring the direction of multiple targets, I. Ziskind and M. Wax, "Maximum likelihood localization of multiple sources by alternating projection, IEEE Transaction on Acoustics Speech and Signal processing, vol. 36, no. 10, pp. There is a maximum likelihood localization method disclosed in .1553-1560, Oct.1988, etc. According to this method, the direction can be separated even if a plurality of targets exist in the beam, but this method requires a large amount of computation. For this reason, a signal processing device with high computational performance is required, and in particular, when the target number increases, the computation amount increases.

As described above, in the sequential roving method and the monopulse method, the directions of the plurality of targets cannot be separated, and in the maximum likelihood estimation method, the directions of the plurality of targets can be separated, but there is a problem that the computation load is high.

An object of the present invention is to solve the problems as described above with the conventional method of combining a plurality of beams to calculate a target direction.

The radar device according to the present invention includes an antenna for emitting a beam in a plurality of directions and receiving the beam reflected by a target as a reception wave;

A receiver for detecting the received wave received by the antenna and outputting a received signal;

A signal detector for extracting a feature of the received wave from the received signal output by the receiver;

A radar device having a direction calculator for calculating a primary direction that is the target direction from a combination of feature quantities calculated by the signal detector from reception waves of at least two beams partially overlapping among the beams radiated in the plurality of directions. In

When there are a plurality of primary directions calculated by the direction calculator, an integrated direction that is a true target direction is calculated based on the primary direction belonging to the area from a region where the distribution of the plurality of primary directions becomes equal to or greater than a predetermined density. It has a direction integrator.

In this manner, in this radar apparatus, the true target direction is extracted from each of the regions where the distribution density in the primary direction becomes equal to or greater than a predetermined value. Accordingly, even when a plurality of targets exist, the directions of the respective targets can be separated.

1 is a block diagram showing a configuration of a radar apparatus according to Embodiment 1 of the present invention;

2 is a block diagram showing a detailed configuration of a signal processor of a radar apparatus according to Embodiment 1 of the present invention;

3 is a flowchart illustrating an operation of a signal processor of a radar apparatus according to Embodiment 1 of the present invention;

4 is a view for explaining the relationship between the beam and the target of the radar apparatus according to the first embodiment of the present invention,

5 is a view for explaining the relationship between the beam and the primary direction of the radar apparatus according to the first embodiment of the present invention,

6 is a flowchart of cluster formation processing in the first embodiment of the present invention;

7 is a block diagram showing the configuration of a radar device according to a second embodiment of the present invention;

8 is a block diagram showing a detailed configuration of a signal processor of a radar apparatus according to Embodiment 2 of the present invention;

9 is a diagram for explaining frequency analysis in the radar device according to the second embodiment of the present invention;

10 is a block diagram showing the configuration of a radar device according to a third embodiment of the present invention;

11 is a block diagram showing a detailed configuration of a signal processor of a radar apparatus according to Embodiment 3 of the present invention;

12 is an operation timing chart of a handset switch in the radar apparatus according to the third embodiment of the present invention;

FIG. 13 is a view showing a state in which the radar apparatus of Embodiment 4 of the present invention is mounted on a vehicle;

14 is a block diagram showing the configuration of a radar device according to a fourth embodiment of the present invention;

15 is a block diagram showing a detailed configuration of a signal processor of a radar apparatus according to Embodiment 4 of the present invention;

16 is a flowchart showing the operation of the model fitting processor in the radar apparatus according to the fourth embodiment of the present invention;

17 is a diagram showing a relationship between a beam pattern and a target in the related art;

18 is a diagram showing a relationship between a beam pattern and a direction calculated by the conventional technology.

(Example 1)

1 is a block diagram showing the configuration of a radar apparatus according to Embodiment 1 of the present invention. In the drawing, the radar device 1 is a pulse radar device using a sequential roving method, and includes a reference signal generator 2, a transmitter 3, an antenna 4-a, 4-b, a receiver 7, and a signal processor ( 8) The reference signal generator 2 is a portion that generates a weak reference signal with a constant frequency by a local generator (local oscillator). In addition, in this description and the following description, a part shall refer to an exclusive circuit or an element. In some cases, a configuration may be employed in which a computer equipped with a central computing device having a general purpose function executes a corresponding process through a computer program.

The transmitter 3 is composed of an amplifier and a pulse modulator. The transmitter 3 amplifies the reference signal generated by the reference signal generator 2 by the amplifier, and generates a pulse wave from the reference signal by the pulse modulator. The antenna 4-a is an antenna that emits pulse waves generated by the transmitter 3 as beams in a predetermined direction. The beam 5-a is a beam radiated by the antenna 4-a. The target 6 is an object which exists outside the radar apparatus 1 and becomes the measurement object of the radar apparatus 1. The radio wave 5-b is a radio wave generated by reflecting a part of the beam 5-a by the target 6. The antenna 4-b is an antenna for receiving the beam 5-b. The receiver 7 is a part which detects the received wave received by the antenna 4-b and outputs a received signal. The signal processor 8 is a portion which performs signal processing on the received signal output by the receiver 7 and calculates the direction in which the target 6 exists. The antenna driver 9 is a part that mechanically or electronically controls the directions of the antennas 4-a and 4-b, and the antennas 4-a and 4-b are connected to the antenna driver 9 so that they face the same direction. To be controlled.

2 is a block diagram showing the detailed configuration of the signal processor 6. In the figure, the signal detector 11 is a part for obtaining a signal feature amount such as amplitude from an input signal. Further, the Δ / Σ measurer 12 obtains the Δ / Σ value of the feature amount determined by the signal detector 11 and calculates the target direction based on each beam from the Δ / Σ value. In addition, in the following description, the direction of the target computed by the Δ / Σ angle measuring device 12 shall be called a primary direction. The cluster separator 13 is a portion for extracting, as a cluster, an angular range in which the primary direction output from the Δ / Σ angle measuring device 12 is concentrated. If there are a plurality of angular ranges in which the primary directions are concentrated, a cluster is formed in the plural angular ranges. The distribution center calculator 14 is a site | part which calculates the distribution center of a primary direction in each cluster formed by the cluster separator 13, and computes the true target direction for every cluster. Here, "distribution center" shall mean the value obtained by statistically processing the singular or plural primary directions which belong to the area | region where a primary direction exists in predetermined distribution density. In addition, in the following description, it is assumed that the direction of the true target is referred to as the "integration direction".

Incidentally, the Δ / Σ angle measurer 12 is an example of a direction calculator, and the cluster separator 13 and the distribution center calculator 14 are examples of a direction integrator.

Next, the operation of the radar apparatus 1 will be described. First, the reference signal generator 2 generates a weak reference signal, and based on the weak signal, the transmitter 3 generates a pulse transmission wave. The antenna 4-a radiates this pulsed transmission wave as the beam 5-a. As described above, the antennas 4-a and 4-b switch the direction in which the beams are radiated or received mechanically or electronically by the control of the antenna driver 9. As a result, the antenna 4-a faces a plurality of directions so that a part of the beams overlap, and radiates two or more antenna beams sequentially in each direction.

The beams 5-a radiated from the antenna 4-a at each time are reflected by the target 6, and a part thereof is received by the antenna 4-b as a reception wave. The received wave received at the antenna 1 is output to the receiver 7, converted from an analog signal to a digital signal (A / D conversion), and the conversion result is output to the signal processor 8 as a received signal. Since the radar device 1 is a pulse radar device using a sequential roving method, the direction of the target 6 is calculated using the signal processor 8 by combining the antenna beams radiated at different times.

(Method for calculating angle from Δ / Σ value)

3 is a flowchart showing processing of the signal processor 6. In step S101, after the characteristic amount of the input signal is detected by the signal detector 11 of the signal processor 6, the Δ / Σ measurer 12 obtains the Δ / Σ value of the characteristic amount of the input signal, and in the primary direction. Calculate more. As an example of the Δ / Σ value, the amplitude of the input signal is taken as a feature amount, the Δ / Σ value (amplitude comparison method) obtained by comparing the amplitudes in a plurality of beam patterns, and the phase of the input signal is taken as a feature amount. There are Δ / Σ values obtained by comparing phases in a plurality of beam patterns, and it is possible to calculate a target direction even when using any of Δ / Σ values. For example, in the amplitude comparison method, in the received signals for two adjacent beams, the error voltage? Resulting from the target direction is the sum of the amplitudes? Of the amplitudes of the received signals of the two beams. Divided by. In other words, the relationship of ε = Δ / Σ is established.

And if the front direction of antenna 4-a, 4-b is (theta) a, the target direction (theta) o will be

(1)

(One)

Given by As a result, the primary direction of the target can be calculated. Moreover, in Formula (1), the function f is a function which shows the relationship between the error voltage (epsilon) and the amount of the deviation from (theta) a of a target direction.

Next, in step S102, the cluster separator 13 forms a cluster based on the distribution in the primary direction calculated by the Δ / Σ angle measurer 12. 4 and 5 conceptually show the relationship between the true target and the primary direction calculated by the Δ / Σ angle measurer 12 and the cluster formed by the cluster separator 13. 4 shows the direction in which the true targets 35 and 36 exist using beam scanning by the antennas 4-a and 4-b when there are a plurality of targets consisting of the true targets 35 and 36. The case where is calculated is shown. As shown in the figure, the antenna 1 emits beam patterns 41 to 47, but beam patterns 41 and 42, 42 and 43, 43 and 44, 44 and 45, 45 and 46, 46 and 47. Are partially overlapping each other.

5 is a diagram showing a result of calculating, as the primary direction, the direction in which the true targets 35 and 36 exist using the beam patterns 41 to 47. In the figure, the primary direction 51 is the primary direction calculated by the Δ / Σ goniometer 12 based on the beam patterns 42 and 43. Similarly, the primary direction 52 is calculated based on the beam patterns 43 and 44, and the primary direction 53 is calculated based on the beam patterns 44 and 45. In addition, the primary direction 54 is calculated based on the beam patterns 45 and 46, and the primary direction 55 is calculated based on the beam patterns 46 and 47.

What is revealed from FIG. 6 is as follows.

(1) One primary direction is calculated from each of the combinations of the beam patterns.

(2) Even if a plurality of primary directions for a certain true target are calculated, for example, primary directions 51 and 52, their primary directions can be different values.

(3) As in the primary direction 53 calculated from the beam patterns 44 and 45, in some cases, the primary direction is calculated as if any target exists even though the direction does not exist. Although the true target does not exist in this way, the calculated primary direction is called "primary direction to virtual" (or "direction to virtual").

Thus, in order to solve the problem in particular (2), the cluster separator 13 clusters the primary direction. In addition, the center distribution calculator 14 calculates the direction with respect to the true target from each of the clusters formed by the cluster separator 13, and excludes the direction with respect to the virtual.

(The details of the cluster formation processing)

6 is a detailed flowchart of the cluster formation process in step S102. These processes are executed by the cluster separator 13. In step S111 of the figure, 1 is set as an initial value of the counter variable N. FIG. This counter variable is used to indicate one of a plurality of primary directions calculated by the Δ / Σ measuring device 12. Next, in step S112, the N-th primary direction is substituted into the variable E. Next, as shown in FIG. As a premise of this process, it is assumed that the plurality of primary directions calculated by the Δ / Σ measuring device 12 are arranged in a constant order, and any one primary direction is determined solely by the order from the beginning. That is, each element is uniquely determined by the designations N-th primary direction and N + 1th primary direction.

In step S113, 1 is similarly set as the initial value of another counter variable M. FIG. This counter variable is used to indicate one of those clusters when there are already created clusters. Like the primary direction, the clusters are also arranged in a constant order, and shall be uniquely determined by the order from the head. In step S114, the maximum value of the Mth cluster is stored in the variable MAX, and the minimum value of the Mth cluster is stored in the variable MIN. The maximum value of a cluster shall mean the maximum value of the primary direction which belongs to the cluster. In addition, the minimum value of a cluster shall mean the minimum value of the primary direction which belongs to the cluster.

Next, in step S115, it is checked whether the variable E (the N-th primary direction) is smaller than the value obtained by adding the predetermined value D to the variable MAX (the maximum value of the M-th cluster) (called condition 1). In addition, it is examined whether the variable E (the N-th primary direction is larger than the value obtained by subtracting the predetermined value D from the variable MIN (the minimum value of the M-th cluster) (called condition 2). As a result of the selection, when each primary direction is arranged in the order of the value of the primary direction, the primary directions as two adjacent primary directions, which are separated by a predetermined value or more, belong to different clusters, that is, the neighboring primary directions. If the distance is near, the distribution density in the primary direction is said to be high. If the distance is far, the distribution density in the primary direction is considered to be low.

In addition, it is sufficient to form the clusters around the point where the distribution density in the primary direction is large. Therefore, in addition to the method of forming clusters based on the distance values between the primary directions, clusters may be formed using other statistical indicators such as standard deviation. Of course it is good to form.

If both condition 1 and condition 2 are satisfied, the process proceeds to step S116 (step S115: YES). In this case, the variable E (Nth primary direction) is included in the Mth cluster in step S116. Thereafter, the process proceeds to step S120, but subsequent processing will be described later.

On the other hand, in step S115, when either or both of condition 1 and condition 2 are not satisfied, the process proceeds to step S117 (step S115: no). In step S117, 1 is added to the counter variable M, and it is determined in step S118 whether the counter variable M is larger than the total number of clusters currently formed. If M is larger than the total number of clusters, it means that there are no more clusters to be processed anymore, and the flow proceeds to step S119 (step S118: YES). In step S119, since the cluster to which the variable E (N-th primary direction) should belong also does not exist, a new cluster is created, and the variable E (N-th primary direction) belongs to this cluster. Thereafter, the process proceeds to step S120, but subsequent processing will be described later.

If the counter variable M is equal to or less than the total number of clusters currently formed in step S118, the process returns to step S114 to repeat the processing thereafter. Accordingly, if the variable E (the N-th primary direction) should belong to one of the existing clusters, the variable E (N) may be repeated by repeating steps S114 to S118 or steps S114 to S116. First primary direction) can belong to the cluster.

Next, the processing after step S120 will be described. In step S120, 1 is added to the counter variable N. If N does not exceed the total number of primary directions in step S121, the process returns to step S112 to perform the process of the next primary direction (step S121: NO). On the other hand, when N exceeds the total number of primary directions in step S121, since the primary directions to be processed no longer exist, the process ends. The cluster formation process by step S102 is complete | finished above.

(Distribution-oriented output processing)

Next, in steps S103 to S104, the distribution center calculator 14 calculates the distribution center of each cluster. First, in step S103, the distribution center calculator 14 substitutes 1 for the counter variable M and initializes it. Next, in step S104, the distribution center of the Mth cluster is calculated to be the integration direction of this cluster. Calculation of the distribution center of the cluster is performed as follows. Now, it is assumed that N primary directions θ _i (i = 1, 2,... N) belonging to the M-th cluster belong to a combination of different beams. Moreover, the amplitude value of the received signal used for calculation of the primary direction (theta) _i is made to _ai . At this time, the distribution center calculator 14 calculates the distribution center W _M of the M-th cluster by Formula (2).

(2)

(2)

The distribution center applied by Formula (2) is a value normalized by dividing the total weighted value by the amplitude value of the received signal in the primary direction θ _i by the total sum of the amplitude values of the received signal. That is, in the pulse radar, a peak of the amplitude value of the received signal exists near the distance where the target exists, and the larger the peak, the greater the probability of the target's existence. Therefore, by increasing the evaluation of the primary direction calculated on the basis of the received signal, the degree of agreement between the distribution center and the direction of the true target becomes high, and the angle measurement accuracy can be improved.

Subsequently, in step S105, the distribution center calculator 14 calculates the intensity a _cM of the M-th cluster using Equation (3).

(3)

(3)

In step S106, the distribution center calculator 14 adds 1 to the counter variable M, and checks in step S107 whether M exceeds the total number of clusters. If M is equal to or less than the total number of clusters, there are also clusters to be processed, and the flow returns to step S104 (step S107: no). On the other hand, if M exceeds the total number of clusters, the process proceeds to step S108 (step S107: YES).

In step S108, the distribution center calculator 14 compares the strengths of the clusters calculated by Equation (3) and selects a predetermined number of clusters in the order of the strongest strength from the strongest clusters. Then, the distribution centers of the selected clusters are respectively calculated as the integration directions. As with the virtual 53 of FIG. 5, the virtual is generally independent of other images. Therefore, the number of primary directions which belong to the cluster 57 which the cluster separator 13 formed corresponding to the virtual 53 becomes small. Since the intensity by Equation (3) is determined based on the sum of the amplitude values of the received powers used to calculate each primary direction, the smaller the number of primary directions belonging to the cluster is, the smaller the strength is considered. From this, the virtual can be eliminated by selecting the integration direction with a large intensity between each cluster.

Similarly, the strength may be calculated based on the number of primary directions belonging to the cluster 57. For example, the number itself may be the strength of the cluster.

In addition, the virtual 53 is not an image corresponding to a true target, and the amplitude of the received signal which formed the virtual 53 is smaller than the amplitude of the received signal of the received wave reflected and received by the true target. Therefore, the intensity may be determined based on the average value of the amplitude of the received signal of the beam unit used to calculate the primary direction belonging to each cluster, rather than the equation (3) affected by the number of primary directions belonging to the cluster.

Instead of calculating the intensity based on the amplitude, the intensity may be calculated based on the power value of the received signal of the received wave of each beam.

As is apparent from the above, according to the radar apparatus of the first embodiment of the present invention, similarly to the radar apparatus of the sequential roving system, it is possible to measure a plurality of target angles by using a radar originally used for the single angle measurement angle.

In addition, it is possible to suppress the generation of imaginary and to achieve an accurate angle measurement.

Moreover, in Example 1 of this invention, the distribution center value of the cluster was calculated | required by Formula (5), and the distribution center value was made into the integration direction of each cluster. By this, the primary direction calculated based on each beam is weighted and averaged, and the effect of reducing the error of each angle value can be expected. However, instead of this method, the primary direction calculated from the beam having the maximum amplitude or the maximum power among the beams based on calculating the primary direction of each cluster may be used as the unifying direction. When the SN ratio of the received signal is sufficiently high, even if the primary direction calculated from the received signal having the highest reception amplitude or power in the cluster is used as the cluster angle value as it is, sufficient side angle accuracy can be obtained.

In addition, in the first embodiment, the pulse radar apparatus of the sequential roving method was described as an example. However, even in the pulse radar device according to the monopulse method, only the configuration of the antenna and the power feeding method are different, and the configuration of the site for processing the received signal is no different from that of the sequential roving. Embodiment 1 of the present invention The principle of operation of can be applied. Therefore, the scope of the present invention is not limited to the pulse radar device of the sequential roving method.

In addition, the features of the present invention shown in the first embodiment of the present invention can be applied to a pulse radar device other than the pulse method, but, in particular, such as a radar device of a frequency modulation continuous wave (FMCW) method, up to distance information In the case of obtaining a cluster, if the cluster is formed, the cluster is not formed only on the condition that the angle value (primary direction) is an approximation, but the cluster is formed by combining the condition that the distance of each image is approximated. , The precision can be further improved.

(Example 2)

In Example 1, a cluster radar was formed on the basis of the distribution density in the primary direction, and the centers of the clusters were sorted, and a method of separating the directions of the plurality of targets was described. However, instead of the method for forming a cluster, a primary direction in which the characteristic amount of the received signal is maximized (with a locally large value) may be selected as the integration direction. The radar device according to Embodiment 2 of the present invention has such a feature.

7 is a block diagram showing the configuration of a radar device according to a second embodiment of the present invention. In the drawing, the radar 21 is a Doppler radar device using a sequential roving method, and among the components, the components given the same conformity as those in FIG. 1 are the same as those in the radar apparatus of the first embodiment. Omit the description. The receiver 22 performs frequency conversion of the received wave (analog signal) received by the antenna 4-b into a video signal (or an intermediate frequency signal) using a reference signal output from the reference signal generator 2. to be. In addition, the receiver 22 performs A / D conversion on the frequency-converted received signal to convert it into a digital signal. The signal processor 23 is configured to perform signal processing on the received signal digitized by the receiver 22, the detailed configuration of which is indicated by FIG.

In the figure, the frequency analyzer 24 is a portion for converting a received signal output from the receiver 22 into a signal in the frequency domain, that is, a spectrum, for example, by Fast Fourier Transform (FFT). The beam radiated by the antenna 4-a is reflected by the target 6, but when the target 6 is moving, frequency shift occurs due to the Doppler effect. The characteristic of the Doppler radar is that the deviation of the frequency is extracted and the speed of the target 6 is measured. The signal detector 25 is a portion that detects a signal component corresponding to the target reflected wave from the spectrum output by the frequency analyzer 24 and detects the characteristic amount (amplitude, power, etc.). [Delta] / [Sigma] measuring instrument 12 is a portion which calculates the primary direction from the feature amounts of the received signal as in the first embodiment.

In addition, the amplitude density distribution calculator 26 is a site for calculating the amplitude density distribution with respect to the primary direction calculated by the Δ / Σ angle measuring device 12. The maximum value calculator 27 is a portion for calculating the maximum value of the amplitude density distribution calculated by the amplitude density distribution calculator 26.

Next, the operation of the radar apparatus 21 will be described. Since the operation from the reference signal generator 2 to the receiver 22 is the same as in the first embodiment, description thereof is omitted. Since the radar device 21 is a Doppler radar only, the Doppler effect occurs because the target 6 is moving, and shifts between the frequency of the transmission wave 5-a and the frequency of the reflected reception wave 5-b. Note that this is happening.

Subsequently, the receiver 22 frequency-converts the received wave input as an analog signal to a video signal (or an intermediate frequency signal) using the reference signal output from the reference signal generator 2. Further, A / D conversion is performed on the received signal obtained by the frequency conversion, and the digital signal is output to the signal processor 23.

Inside the signal processor 23, the frequency analyzer 24 performs a Fast Fourier Transform on the received signal, and outputs a distribution of the frequency deviation and the received amplitude from the received signal for each period for each beam. Specifically, as shown in Fig. 9, the received signal (Fig. 9 (a)) is cut out at regular intervals, and frequency analysis is performed for each (Fig. 9 (b)). In this example, the distribution between the frequency shift and the reception amplitude is output. However, it goes without saying that the distribution between the frequency shift and the reception power may be output. In addition, the reception amplitude and the reception power are examples of the feature quantities of the received signal, and may be a distribution of the frequency deviation and other feature quantities.

Subsequently, the signal detector 25 extracts, from the frequency-receive amplitude distribution output by the frequency analyzer 24 at predetermined intervals, the frequency (deviation of the peak) of the received amplitude and the received amplitude, and Δ / Output to the angle measuring device 12; In the Δ / Σ measuring device 12, similarly to the first embodiment, the Δ / Σ value is calculated from the reception amplitude in the combination of adjacent beam patterns, and the primary direction is output.

The amplitude density distribution calculator 26 obtains an amplitude density distribution by smoothing amplitude values distributed discretely with respect to the plurality of primary directions calculated by the Δ / Σ angle measuring device 12. Specifically, assuming that the reception amplitude value of the antenna beam used to calculate the primary direction θ _k is a _k , the amplitude density distribution calculator 26 calculates the reception amplitude density distribution A (θ) by equation (4). do.

(4)

(4)

Here, δ (θ) is the δ function of the disk, and W (θ) is the window function used for the smoothing process.

Next, the local maximum calculator 27 calculates the local maximum of A (θ) from the received amplitude density distribution A (θ) calculated by Equation (4). A (θ) of Formula (4) becomes large in the vicinity of the direction where the primary direction is concentrated. In addition, the primary direction calculated from the received signal of the beam that does not interfere with the reflected waves from the plurality of targets becomes almost the same value as the direction of the true target regardless of the combination of the beams. In contrast, the primary direction calculated from the received signal of the beam interfering with the reflected waves from the plurality of targets differs depending on the beam used for the calculation. This direction becomes an imaginary direction, and since the distribution of a primary direction becomes sparse in an imaginary, the value of A ((theta)) becomes small. A (θ) in the vicinity of the imaginary does not become a maximum value, or even a maximum value thereof is small. On the other hand, it becomes a value more than a fixed value with respect to the direction of a true target. In this way, the direction with respect to the true target can be separated.

Subsequently, the maximum value calculator 27 detects the maximum value (peak) of A (θ) by a predetermined number and outputs it as an integrated direction.

As is clear from the above, according to the radar apparatus of Example 2, virtual separation can be prevented, separating a true target by calculating amplitude density distribution and its maximum value.

In addition, in this Example 2, although cluster separation was not carried out, similarly to Example 1, the component corresponding to the cluster separator 13 was provided, cluster separation is carried out, and amplitude density distribution by Formula (4) for every cluster is carried out. May be calculated. In this case, the intensity of each cluster may be calculated in the same manner as in Example 1, the clusters may be selected based on this intensity, and then the amplitude density distribution may be calculated.

(Example 3)

The radar apparatuses according to Examples 1 and 2 select the integration direction from the primary directions forming the predetermined distribution density, or determine the integration direction by calculating the distribution center to separate the directions of the plurality of targets.

In addition, after obtaining the integration direction by the method described so far, the model fitting process may be performed with respect to the integration direction, and the calculation precision of the target direction may be improved. The radar apparatus according to Embodiment 3 of the present invention has such a feature.

10 is a block diagram showing the configuration of a radar apparatus according to Embodiment 3 of the present invention. In the figure, the radar device 61 is a Doppler radar device by the monopulse method. Since the reference signal generator 2 and the transmitter 3 are the same as those of the substantial part of the second embodiment, description thereof is omitted. The distributor 62 is a circuit or element that distributes the transmission signal generated by the transmitter 3 to a plurality of output destinations. In the example of the figure, it is comprised so that it may distribute to two output destinations. The transmitter / receiver 63-a is a switch which directly connects the antenna 64-a to either the distributor 62 or the receiver 22, has a movable terminal A, contacts B and C, and controls from the outside. The connection destination can be changed by a signal. Regarding the transmitter / receiver 63-b, it is a switch that connects the antenna 64-b to either the distributor 62 or the receiver 22, similarly to the transmitter / receiver 63-a. It has contacts B and C. In addition, the movable terminal mentioned here is not limited to being mechanically movable, You may select the connection terminal electronically.

The antennas 64-a and 64-b simultaneously radiate the transmission wave 65-1 and the transmission wave 65-2, respectively, and receive the reflected waves 65-3 and 65-4, respectively. to be. The beam patterns of the transmission waves 65-1 and 65-2 are radiated so that a part of them overlaps. The reflected wave 65-3 is a reflected wave generated by reflecting a portion of the transmission wave 65-1 to the target 6, and the reflected wave 65-4 is a portion of the transmitted wave 65-2 to the target 6; Reflected wave generated by reflection. Since the receivers 66-a and 66-b are the same as the receiver 22 of the second embodiment, description thereof is omitted. The signal processor 67 combines the received signals of at least two systems of the antenna 64-a, the receiver 66-a, and the antenna 64-b, and the receiver 66-b to perform side angle processing. The detailed configuration is shown by the block diagram of FIG.

In FIG. 11, the frequency analyzer 68 is a site | part corresponded to the frequency analyzer 24 in Example 2, and analyzes the frequency shift by the Doppler effect. The signal detector 69 is a portion for extracting the feature amount of the received signal in the deviation of the frequency analyzed by the frequency analyzer 68. Here, since the radar device 61 is a pulsed Doppler radar device, not only the frequency shift but also the arrival delay time of the pulse wave must be taken into consideration. That is, the pulsed waves radiated from the antennas 64-a and 64-b reach the target 6, are also reflected by the target 6, and return to the antennas 64-a and 64-b again. The time elapses as the radio wave moves the round trip distance between the antennas 64-a and 64-b and the target 6. Thus, since the distance information is also included in the arrival time of the pulse wave in the received signal in a pulsed Doppler radar apparatus, these are extracted as needed.

Since the Δ / Σ angle measuring device 12, the cluster separator 13, and the distribution center calculator 14 are the same as those in the first embodiment, the description thereof is omitted. The model fitting processor 70 integrates the model defining the relationship between the reflectance and the direction and the feature amount of the received signal, and the feature amount of the received signal calculated by the distribution center calculator 14 and the integration direction. It is the part which improves the precision of a direction.

The model fitting processor 70 also constitutes a part of the direction integrator.

Next, the operation of the radar device 61 will be described. First, the reference signal generator 2 generates a weak reference signal with a constant frequency by a local generator built therein. The transmitter 3 amplifies this weak reference signal and performs pulse modulation to generate a transmission signal. This transmission signal is transmitted by the distributor 62 to the handset switching devices 63-a and 63-b.

Transmitters 63-a and 63-b connect the movable terminal A to the contact B. FIG. As a result, since the distributor 62 and the antennas 64-a and 64-b are directly connected, the transmission signal (pulse signal) generated by the transmitter 3 is transmitted to the antennas 64-a and 64-b. And are simultaneously radiated from the antennas 64-a and 64-b as the transmission waves 65-1 and 65-2, respectively. Since the radar device 61 is a radar device of the monopulse method, by simultaneously radiating beams by at least two beam patterns, a combination of beams can be obtained by only one pulse wave radiation. However, in parallel with this, the antenna driver 9 has the antennas 64-a and 64-b so as to obtain a beam pattern in a direction larger than the number of element arrays by the antennas 64-a and 64-b. Changes the direction in which the beam is emitted.

When the antennas 64-a and 64-b emit pulsed waves, the hand-turn devices 63-a and 63-b connect the movable terminal A to the contact C. FIG. Thereby, the antenna 64-a and the receiver 66-a, and the antenna 64-b and the receiver 66-b are directly connected. In the meantime, the beams 65-1 and 65-2 are reflected by the target 6 and arrive at the antennas 64-a and 64-b again as the reflected waves 65-3 and 65-4. The antennas 64-a and 64-b receive the respective reflected waves as receive waves and output the received waves to the receivers 66-a and 66-b via the handset converters 63-a and 63-b. do. In this way, the radar device 61 switches the transmission and reception of the antennas 64-a and 64-b by the transmission and reception switches 63-a and 63-b. The timing at which the antennas 64-a and 64-b emit pulsed waves is arbitrary, whereas the timing at which the antennas 64-a and 64-b receive reflected waves is determined by the relative positional relationship with the target 6. Since the position of the target 6 is determined to be negative in many cases, the timing at which the water switching devices 63-a and 63-b switch the movable terminal A is connected to the contact C as compared with the time connected to the contact B. FIG. Often times are long. Fig. 12 is a timing chart for switching the movable terminal A in the water transmitters 63-a and 63-b.

Next, similarly to the receiver 22 in the second embodiment, the receivers 66-a and 66-b convert the reference signal and the received wave into digital signals and output them to the signal processor 67.

In the signal processor 67, the frequency analyzer 68 performs a Fast Fourier Transform on the received signal, and outputs a distribution between the frequency shift and the received signal amplitude from the received signal at predetermined intervals for each beam. In the signal detector 69, a frequency at which the received signal amplitude becomes a peak is extracted in the distribution between the frequency shift and the received signal amplitude. Further, the Δ / Σ measurer 12 obtains a Δ / Σ value based on the received signal amplitude extracted by the signal detector 69, and calculates a primary direction from the Δ / Σ value. Since these processes are the same as the processes in the frequency analyzer 24, the signal detector 25, and the Δ / Σ angle measuring device 12 in Example 2, detailed descriptions are omitted.

Subsequently, the cluster separator 13 forms a cluster based on the distribution in the primary direction, and the distribution center calculator 14 calculates the distribution center in each cluster. These processes are the same as those in the first embodiment.

Next, the model fitting processor 70 performs a model fitting process with respect to the distribution center (integration direction of each cluster) calculated by the distribution center calculator 14. In the following description, the number of beams used for observation is m, the true direction is θ, and the j-th (where j = 1, 2, ... m) beam directing characteristics are α _j (θ) and j-th beams. The received signal actually observed is s _j . In addition, assuming that the target number is n, and the reflectance, angle, and distance of the target of the i-th (where i = 1, 2, ..., n) are γ _i , θ _i , r _i , respectively, the estimated value S 'of the received signal. _j is represented by equation (5).

(5)

(5)

However, C is a coefficient determined by the performance of the radar device.

So, using the least square method

(6)

(6)

Γ _i , θ _i are estimated so that is the minimum. Here, when γ _i and θ _i are simultaneously estimated, since the least square method is a nonlinear least square method, γ _i and θ _i are estimated by repeated improvement. In addition, the value of Formula (6) is known as a residual square sum.

In the estimation process by the iterative improvement, if the initial value selection method is not good, there is a problem that the amount of calculation required for the estimation process increases and the processing time takes. Therefore, in the radar device 61, it is assumed that the distribution center of the cluster calculated by the distribution center calculator 14 is used as the initial value of θ _i . Since the error of the center of distribution of a cluster is small and the precision is high enough, an estimated value converges early when using the center of distribution of a cluster as an initial value. Therefore, the calculation amount can be reduced, and the calculation can be performed sufficiently efficiently.

On the other hand, the reflectance γ _i has a strong linearity in the least squares method and has a low initial value dependency in the estimation process. Therefore, the initial value of the reflectance γ _i need not be given a value that is too close to the true value. For example, an arbitrary constant value may be used as the initial value of the reflectance γ _i .

In addition, when γ _i and θ _i are estimated by model fitting, since the reflectance becomes almost zero in the direction where the target does not actually exist, the reflectance decreases in the direction in which the imaginary appears. Therefore, if the distribution center of the cluster which has a reflectance of a predetermined value or less is rejected, the virtual can be prevented from being adopted.

Finally, the model fitting processor 70 outputs θ _i estimated in this way as the integration direction.

As is apparent from the above, according to the radar apparatus of the fourth embodiment of the present invention, the high-density direction estimation is performed by using the model fitting method, while the center of distribution of the cluster is used as the initial value of the model fitting. Since the estimation is made, the number of iterations in the model fitting can be reduced, and the computation amount can be reduced.

In the model fitting processor 70, the reflectance γ _i is assumed, but instead of the reflectance γ _i , γ ' _i = γ _i / r _i ⁴ may be estimated. In Equation (5), in order to make the reflectance γ _i and the direction θ _i as estimation parameters, other variables must be known, but on the contrary, the distance r _i is often unknown. Even in such a case, by estimating the γ '= γ _{i i} / r _i ^4, the γ _i and r _i When not viewed as an independent variable, even if the distance r _i is unknown, estimation processing by model fitting can be applied.

In addition, the model fitting processor 70 is supposed to perform model fitting with respect to the received signal s _i (s _i has amplitude and phase diagram parameters). However, in addition to this, model fitting may be performed on the amplitude value and the power value of the received signal. In order to perform model fitting using the amplitude value of the received signal, in the formula (6), instead of s _i and s ' _i , | s _i | and | s' _i | may be used. It may be to use a | addition to the model fit using a power value of a received signal, in the formula (6), s _i and s _'i in place of | s _i |, and | s' _i. By model fitting the amplitude value and the power value of the received signal, it is not necessary to deal with the amplitude and phase, and the computation amount can be further reduced.

In addition, although the distribution center was used as the initial value of the direction, when the SN ratio is high as described in the first embodiment, the primary direction calculated from the received signal having the highest reception amplitude or power in the cluster may be the initial value. In addition, as described in the second embodiment, the primary direction in which the characteristic amount of the received signal is maximum in the local distribution in the primary direction may be an initial value.

(Example 4)

According to the third embodiment described above, even when the direction estimation by the model fitting is performed, the calculation amount required for the estimation of the reflectance and the direction can be reduced by setting the distribution center of the cluster as the initial value. However, the total amount of estimation operations by the least square method may be reduced by providing a plurality of initial values in the direction and combining the respective estimation operations. The radar device according to the fourth embodiment has this feature.

The radar device according to the fourth embodiment of the present invention is a radar device (vehicle mounted radar) mounted on an automobile. For example, as shown in Fig. 13, the mounted radar 71 mounted on the front of the vehicle is used for collision prevention and auto cruise. It is used for inter-vehicle control of the dragon. In addition, the vehicle-mounted radars 72 and 73 mounted on the side surface are used for side stone detection and overtaking vehicle detection. And the vehicle-mounted radar 74 mounted in the back is used for back obstacle detection. As these radar devices can detect the direction of the target, for example, if the on-board radar 71 mounted on the front of the vehicle detects a car traveling in the side lane, there is no automatic deceleration. . In addition, if the vehicle-mounted radar (72, 73) mounted on the side, it is possible to accurately determine the position of the overtaking car to perform appropriate safety control.

Since in-vehicle radars often need to measure not only the position of a target but also the speed, a frequency modulated continuous wave (FMCW) radar system capable of measuring the relative distance and the relative speed of the own vehicle is often adopted. Therefore, the radar apparatus according to the fourth embodiment of the present invention is assumed to be a radar apparatus using the FMCW method, and the method of applying the present invention to the FMCW radar will be described.

14 is a block diagram showing the configuration of a radar apparatus according to Embodiment 4 of the present invention. In the figure, the radar device 81 is an FMCW radar device. In the radar device 81, the reference signal generator 82 is provided with a voltage controlled oscillator (VCQ), and is a portion that generates a weak reference signal whose frequency gradually increases or decreases at a predetermined period. The transmitter 3 has an amplifier similarly to the first embodiment, and is a site for amplifying a weak reference signal to a transmission signal. The transmitter-receiver 83 outputs the transmission signal output from the transmitter 3 to the antenna 4-a, and simultaneously outputs the reflected wave output from the antenna 4-a to the receiver 85 mentioned later. to be.

In the antenna 4-a and the antenna driver 9, since it is the same as that of Embodiment 1, description is abbreviate | omitted. Since the radar device 81 is an FMCW radar, a portion of the beam 84-a, which is an up phase and a down phase (or an up chirp and a down chirp), is radiated to the target 6, and a part thereof. Is received by the target 6 and receives the reception wave 84-b arriving at the antenna 4-a. The receiver 85 is provided with a mixer (mixer) and generates a bit signal between the received wave and the reference signal generated by the reference signal generator 82, and converts the bit signal into a digital signal and outputs the bit signal. to be. The bit signal is generated for each target in each phase. Therefore, in one aspect, when bit signals are obtained from the reflected wave by irradiating beams to N targets, N bit signals are generated. The signal processor 86 processes the bit signal to calculate the relative distance and relative speed, which are the characteristics of the FMCW radar, and to measure the direction of the target, which is the object of the present invention. Indicated by the figure.

In Fig. 15, the frequency analyzer 87 is a part for frequency analysis of the bit signal, and the signal detector 88 is a part of the bit signal such as the frequency and amplitude of the bit signal from the frequency distribution analyzed by the frequency analyzer 87. It is a site for extracting feature quantities. The up / down hinge connector 89 is a site for generating a pair of the bit signal in the up phase (up-up) and the bit signal in the down phase (down-up).

Thereafter, the Δ / Σ angle measurer 12, the cluster separator 13, and the distribution center calculator 14 are the same as those in the first embodiment, and thus description thereof is omitted. Finally, the model fitting processor 90 is a site for model fitting with respect to the distribution center of the cluster. In addition, in FIG. 15, the structure required in order to calculate the direction of a target using a FMCW radar is shown, The illustration and description are abbreviate | omitted about the structure required in calculating a relative distance or a relative speed.

Next, the operation of the radar device 81 will be described. Although the radar apparatus 81 is a radar apparatus of the FMCW system, since the operation from the reference signal generator 82 to the receiver 85 is the same as the operation of a substantial portion in the pulse radar apparatus shown in the first embodiment, it is explained. Omit. However, as described above, the receiver 85 differs in that it outputs the bit signal to the signal processor 86.

In the signal processor 86, the frequency analyzer 87 generates a frequency distribution of the bit signal by fast Fourier transform or the like, and the signal detector 88 extracts the frequency of the bit signal from this frequency distribution. The up / down fold connector 89 generates a pair of the beat signal frequency of the up phase and the beat signal frequency of the down phase extracted by the signal detector. This is for the following reason.

Although the radar device 81 is an FMCW radar device, the FMCW radar device is said to calculate the relative speed and the relative distance when a pair of up-phase bit signals and down-phase bit signals can be obtained. Specifically, if the bit signal frequency of the up phase is U, the bit signal frequency of the down phase is D, the frequency sweep width is B, the modulation time is T, the light flux is c, and the wavelength of the transmission wave is lambda, the desired relative distance. R and relative velocity V are given by equations (7) and (8), respectively.

(7)

(7)

(8)

(8)

Thus, in order to calculate R and V, it is necessary to determine U and D. By the way, when there are N targets, there are N bit signals in each of the upphase and the downphase, and the combination of the bit signals of some upphase and the bit signals of some downphase is expressed by Equation (7). It can be understood that R and V become completely different values depending on whether or not (8) is substituted. Therefore, when there are a plurality of targets, it is important to obtain an accurate combination of the upphase bit signal and the downphase bit signal in order to calculate the respective relative distance and the relative speed. The up / down fold connector 89 does this processing. Here, in the FMCW radar, when there are a plurality of targets, there are already some known methods (e.g., Japanese Patent Application Laid-Open No. 5-142337) regarding a method of obtaining an accurate combination of an upphase bit signal and a downphase bit signal. Since the publication "Millimeter-wave radar distance velocity measuring apparatus" etc. is introduced, description is abbreviate | omitted here.

Next, in the Δ / Σ measuring instrument 12, the Δ / Σ value is obtained based on the amplitude value of the bit signal, and the target primary direction is calculated. Subsequently, the processing in the cluster separator 13 and the distribution center calculator 14 are the same as those in the first embodiment, and thus description thereof is omitted.

Subsequently, in the model fitting processor 90, model fitting of the received signal s _j and the estimated value s _j of the distribution center of the cluster is performed. 16 is a flowchart showing model fitting processing in the model fitting processor 90. First, in step S401 of FIG. 16, the model fitting processor 90 initializes the counter variable M to one. In subsequent processing, the counter variable M is used to identify the cluster.

Next, in step S402, the model fitting processor 90 sets the angle estimation value by a predetermined number in the range near the distribution center of the cluster with respect to the M-th cluster. Here, it is assumed that N angles (N is a natural number) are selected. As a method of setting the angle estimate value, for example, a method of selecting N primary directions as the angle estimate value in order of decreasing distance from the distribution center among the primary directions belonging to the M-th cluster.

In step S403, the counter variable I and the variable MIN are initialized. 1 is applied to the initial value of the counter variable I. This counter variable I is used to identify any one of the N angle estimates. The variable MIN is a storage area used for calculating the minimum value, and any value may be used as the initial value as long as it is a certain large value that cannot be taken as the minimum value. Subsequently, in step S404, to the angle estimated value of the I-th θ _i, to calculate the s' _j by the equation (5), in the least square method using the equation (6), to improve only the reflection factor γ _j repeat We do model fitting. As described in Example 3, since the reflectance γ _j has a strong linearity, the estimated value converges at an early time. As a result, the reflectance which minimizes the value (residual square sum) of Formula (6) by repeating improvement is calculated | required. In step S405, whether the minimum value found in step S404 is smaller than MIN is checked. When the minimum value is smaller than MIN, it means that the minimum value by the current angle estimate value is smaller than the minimum value obtained by using the previous angle estimate value in this cluster. In this case, the process proceeds to step S406 (step S405: YES), the current residual square sum calculated by the formula (6) is set in the variable MIN, and the current angle estimated value is stored in the memory. In addition, the minimum value of Formula (6) and the angle estimation value which obtained the minimum value are reserved for each cluster, and are preserve | saved. After that, the flow advances to step S407.

On the other hand, in step S405, if the minimum value obtained in step S404 is not smaller than MIN, since the minimum value is not applicable as the minimum value of the entire cluster, the process proceeds directly to step S407 (step S405: No).

Next, in step S407, 1 is added to I. In a subsequent step S408, it is determined whether I exceeds N, and if it is equal to or less than N, the flow returns to step S404 (step S408: no). On the other hand, when N is exceeded, since there are no angle estimates in this cluster any more, the process transitions to the next cluster. Specifically, the process proceeds to step S409 (step S408: YES) and 1 is added to M. In step S410, it is determined whether M exceeds the total number of clusters. As a result, when M is equal to or less than the total number of clusters, there are also unprocessed clusters, and the flow returns to step S402 to process the next cluster (step S410: no). In addition, when M exceeds the total number of clusters, since there are no more unprocessed clusters, the process proceeds to step S411 (step S410: YES).

In step S411, the model fitting processor 90 outputs an angle estimation value for minimizing equation (6) required for each cluster.

As is apparent from the above, according to the radar apparatus of the fifth embodiment of the present invention, it is possible to eliminate the iterative improvement operation on the angular components by fixing the angular components that are difficult to estimate unless used in the nonlinear least-squares method to a plurality of angular estimated values. This makes it possible to further reduce the amount of computation.

In addition, only the reflectance may be model-fitted by setting the number N of the angle estimation values to 1, and as a result, only the cluster suitable for the predetermined condition may be selected. For example, if only one primary direction is selected in each cluster (this is an angle estimate), the angular component of Equation (5) is fixed in this primary direction, and only the reflectance is model-fitted, and the reflectance becomes above the reference value. You may employ | adopt this angle estimation value as an integration direction. It goes without saying that the distribution center may be used instead of the primary direction.

Similarly, only one distribution center or one primary direction is selected as the angle estimate, and only the reflectance is model-fitted, so that when the value of equation (6) does not fall below a predetermined value, the cluster rejects the primary direction of the cluster. You may. When the formula (6) (residual square sum) is not sufficiently small, it means that the assumed direction, that is, the precision of the cluster angle value is low, so that the angle angle precision can be improved by invalidating the value.

As mentioned above, the radar apparatus which concerns on this invention is useful for the use which measures the direction of several target.

Claims (23)

An antenna for emitting a beam in a plurality of directions and receiving the beam reflected by a target as a reception wave; A receiver for detecting the received wave received by the antenna and outputting a received signal; A signal detector for extracting a feature of the received wave from the received signal output by the receiver; A radar device having a direction calculator for calculating a primary direction which is a direction of the target from a combination of feature quantities calculated by the signal detector from reception waves of at least two beams partially overlapping among the beams radiated in the plurality of directions. To When there are a plurality of primary directions calculated by the direction calculator, an integrated direction that is a true target direction is calculated based on the primary direction belonging to the area from a region where the distribution of the plurality of primary directions becomes equal to or greater than a predetermined density. With directional integrator Radar device, characterized in that. An antenna for emitting a beam in a plurality of directions and receiving the beam reflected by a target as a reception wave; A receiver for detecting the received wave received by the antenna and outputting a received signal; A signal detector for extracting a feature of the received wave from the received signal output by the receiver; A radar device having a direction calculator for calculating a primary direction which is a direction of the target from a combination of feature quantities calculated by the signal detector from reception waves of at least two beams partially overlapping among the beams radiated in the plurality of directions. To In the case where there are a plurality of primary directions calculated by the direction calculator, the distribution of the characteristic amounts of the received waves used for the calculation of the plurality of primary directions is based on a target direction belonging to the region from the region where the distribution of the characteristic amounts becomes equal to or greater than a predetermined density. With a direction integrator that calculates the direction of integration that is the direction of the true target. Radar device, characterized in that. The method according to claim 1 or 2, And the direction integrator forms a cluster in a primary direction belonging to a region of the predetermined density or more, and calculates the integration direction in units of the cluster. The method of claim 3, wherein And the direction integrator associates the directions of the two targets with the other clusters when the angle difference between the two primary directions becomes a predetermined value or more. The method of claim 3, wherein And the direction integrator obtains distribution centers of a plurality of primary directions belonging to the cluster, and outputs a target center as an integration direction of the cluster. The method of claim 5, wherein And the direction integrator obtains the distribution center by weighting each of the primary directions belonging to the cluster by the feature amount of the received wave used for the calculation of the primary direction. The method of claim 6, And the directional integrator weights using the received amplitude of the received wave as the feature amount of the received wave. The method of claim 6, And the directional integrator is weighted using the received power of the received wave as a feature amount of the received wave. The method of claim 3, wherein And the direction integrator sets the direction of the target in which the reception amplitude of the received wave is used to calculate the primary direction belonging to the cluster as the integration direction of the cluster. The method of claim 3, wherein And the direction integrator sets the primary direction in which the received power of the received wave to be used for the calculation of the primary direction belonging to the cluster is the maximum. The method of claim 2, The direction integrator obtains the distribution density of the reception amplitude of the received wave used for the calculation of the primary direction, and outputs the angle at which the distribution density becomes maximum as the integration direction. The method of claim 11, And the directional integrator obtains the distribution density by setting a window function that smoothes the reception amplitude of the received wave. The method of claim 1, And the direction integrator obtains the strength in the integration direction and outputs the integration direction when the strength satisfies a predetermined condition. The method of claim 13, And the direction integrator obtains the sum of the received amplitudes of the received waves used in the calculation of the integration direction as the strength. The method of claim 13, And the direction integrator obtains the average value of the reception amplitude of the received wave used in the calculation of the integration direction as the strength. The method of claim 13, And the direction integrator obtains, as the strength, an average value of the received power of the received wave used in the calculation of the integration direction. The method of claim 13, And said direction integrator outputs said integration direction when said intensity becomes equal to or greater than a predetermined value. The method of claim 3, wherein The direction integrator obtains the strength of the integration direction of the cluster based on the number of primary directions belonging to the cluster, and outputs the integration direction when the strength satisfies a predetermined condition. . The method of claim 3, wherein And the direction integrator selects a predetermined number of the integration directions in order of the strength, and outputs the selected integration directions. The method of claim 3, wherein The direction integrator is a model of a model received signal that is set in advance using the calculated integration direction as an initial value of an angular component, assuming the angle and reflectance of the target, and the received signal used by the direction calculator to calculate the primary direction. Estimating the direction of the target by fitting. The method of claim 20, The direction integrator selects a predetermined number of primary directions from the primary directions belonging to the cluster and performs model fitting for estimating the reflectance by the least-square method using any one of the selected primary directions as the angular component. Ether device. The method of claim 21, And the direction integrator estimates the direction of the target in the cluster when the reflectance estimated by model fitting becomes equal to or greater than a predetermined value. The method of claim 21, And the direction integrator rejects the integration direction calculated from the cluster such that the minimum value of the sum of residuals of the model fitting becomes a predetermined value or more.

Images