JP2013096828A - Doppler radar system and object detection method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve detection accuracy of an object position while suppressing increase of receiving antennas.SOLUTION: A Doppler radar system 100 includes: a transmitting antenna 120; a plurality of receiving antennas 121, a micro-Doppler calculation unit 105 for calculating a plurality of measurement points indicating respective positions of a plurality of reflection points included in a target object 150; a first centroid calculation unit 106a for calculating a first centroid 116a of the measurement points by using a first division frame having a first interval; a second centroid calculation unit 106b for calculating second centroids 116b of the measurement points by using a second division frame having a second interval narrower than the first interval; and a centroid determination unit 107 for determining at least one second centroid 116b out of the plurality of second centroids 116b as the position of the target object 150 on the basis of positional relation between the second centroid 116b and the first centroid 116a.

Description

  The present invention relates to a Doppler radar system that detects the position of a target object and an object detection method.

  In recent years, the emphasis has been placed on the living environment where people can live safely and comfortably and the development of social infrastructure. Along with this, in daily life, for example, it has been desired to realize an assisting device that assists the actions or operations of elderly people, or an autonomous robot that works to reduce human burdens.

  In an actual living environment, there are humans or many other obstacles. In order for a mobile device or a robot to run around in it, it is desired to avoid a collision between the mobile device or the robot and an obstacle. That is, it is strongly desired that these mobile devices or robots ensure safety while efficiently avoiding collisions with obstacles, warning the approach of obstacles, and stopping.

  On the other hand, among obstacles that become obstacles when such a mobile device travels, it is human that it is difficult to predict movement or behavior. In addition, there is a problem that interpersonal collision has a large effect on both pedestrians and passengers.

  Such robots are used in unusual conditions, such as a fire-smoke environment, darkness without streetlights at night, or a space across walls, rather than a space with good visibility. Sensing under conditions is required. That is, it is necessary to be able to detect humans well and predict their behavior and movement trajectory even in the worst environment.

  As a sensing means applicable to such an environment, there is a radio wave sensor (radar). Since radar can measure relative speed and relative distance simultaneously and instantaneously, it has already been put into practical use as a sensing system for automobiles.

  As an attempt to measure the shape and movement of humans such as pedestrians using radar (radio waves), combining speed detection by Doppler shift and direction-of-arrival detection (DOA: Direction-of-Arrival) by interferometry A method called DDOA (Doppler and Direction-of-Arrival) method is known (for example, see Non-Patent Document 1).

  FIG. 24 is a diagram illustrating a configuration of a radar apparatus using the DDOA method.

  A radar apparatus 901 illustrated in FIG. 24 includes a transmitter 910, receivers 920 and 930, a transmission antenna 911, and reception antennas 912 and 913.

  The objects to be detected by the radar device 901 are measured objects 931, 932, and 933. The transmitter 910 emits a detection radio wave having a certain frequency. The receivers 920 and 930 receive the reflected waves in which the detection radio waves are reflected by the measured objects 931 to 933.

  When the measured objects 931 to 933 are moving at a certain line-of-sight speed with respect to the radar apparatus 901, the frequency of the reflected wave received by the receivers 920 and 930 is the frequency of the detection radio wave radiated from the transmitting antenna 911. On the other hand, the frequency is shifted by the frequency corresponding to the visual line direction speed. From the shifted frequency (Doppler frequency), the line-of-sight speeds of the objects to be measured 931 to 933 can be detected.

  Here, the “line-of-sight speed” is a speed component along the direction from the radar device 901 to the object to be measured among the speeds of the objects to be measured 931 to 933. That is, the line-of-sight speed is a relative speed component of the measured objects 931 to 933 with respect to the radar device 901. Specifically, as shown in FIG. 24, assuming that the speeds of the objects to be measured 931 to 933 are V1, V2, and Vi, respectively, the visual line direction speeds of the objects to be measured 931 to 933 are the speeds V1, V2, and Vi. V1f, V2f, and Vif are velocities decomposed along the direction from the radar apparatus 901 to the measured objects 931 to 933.

  That is, the radar apparatus 901 detects the visual line direction velocities V1f, V2f, and Vif of the objects to be measured 931-933 using the frequencies of the reflected waves received by the receivers 920 and 930 with respect to the frequency of the detection radio wave.

  Incidentally, as shown in FIG. 24, the radar apparatus 901 includes two reception systems including a reception antenna and a receiver corresponding to the reception antenna. Further, the respective receiving antennas 912 and 913 are arranged at different locations.

  Thereby, the distance from each of the measured objects 931 to 933 to the receiving antenna 912 is different from the distance from the measured objects 931 to 933 to the receiving antenna 913.

  Thus, since the distances from the measured objects 931 to 933 to the two receiving antennas 912 and 913 are different, the direction of the measured objects 931 to 933 can be detected. Hereinafter, the principle of the direction detection will be specifically described.

  In FIG. 24, for example, the device under test 933 is present at a position closer to the reception antenna 913 than the reception antenna 912, so that the reflected wave from the device under test 933 arrives at the reception antenna 913 earlier than the reception antenna 912. Therefore, when the reflected wave received by the receiving antenna 912 is compared with the reflected wave received by the receiving antenna 913, the phase of the reflected wave received by the receiving antenna 912 is delayed from the phase of the reflected wave received by the receiving antenna 913. Here, it is assumed that the DUT 933 is in the direction of θi from the front of the reception antenna 912 and the reception antenna 913, and the two reception antennas 912 and 913 are arranged apart by a distance d. In this case, the phase difference between the reflected wave received by the receiving antenna 912 and the reflected wave received by the receiving antenna 913 is expressed by (Equation 1). Note that φ1 is the phase of the reflected wave received by the receiving antenna 912, φ2 is the phase of the reflected wave received by the receiving antenna 913, and λ is the wavelength of the detection radio wave emitted from the transmitting antenna 911.

      φ2−φ1 = 2πd × sin θ / λ (Formula 1)

  When this (formula 1) is modified, the following (formula 2) is obtained. That is, the direction θ of the device under test 933 can be detected from the phase difference φ2 to φ1 between the two reflected waves received by the two receiving antennas 912 and 913.

θ = sin ⁻¹ {(φ2−φ1) λ / (2πd)} (Expression 2)

  This is DOA estimation using interferometry.

  Further, in DDOA, the Doppler frequency of each device under test is detected. And by performing said DOA with respect to the signal of the Doppler frequency corresponding to each to-be-measured object among the reflected waves which the receiving antenna received, the direction of the to-be-measured object can be detected. Thus, the conventional radar apparatus 901 can detect the direction of each object to be measured after identifying a plurality of objects to be measured.

  In FIG. 24, since the radar apparatus 901 includes two receiving antennas 912 and 913, it can detect only a one-dimensional direction. For example, in addition to a straight line including the receiving antenna 912 and the receiving antenna 913, another receiving antenna can be used. By arranging, a horizontal / vertical two-dimensional direction can be detected.

  Consider a case where such DDOA is used for human body detection. Here, parts constituting the body such as human arms, legs, torso, and head move at different speeds. Therefore, the Doppler shift reflecting each motion can be observed as a fine structure of the spectrum (micro Doppler). Thus, by analyzing and analyzing the fine structure, in addition to human detection, human motion identification can be performed. Furthermore, humans and other objects can be distinguished by DDOA. As described above, imaging by DDOA has an advantage that three-dimensional imaging is possible by using only three reception antennas.

  However, since DDOA analyzes the Doppler shift for the reflected wave from the object and estimates the arrival direction for the entire observable range, it is susceptible to various interferences. In addition, there is a problem that the influence of this interference particularly on the arrival direction estimation result is large.

  Specifically, in DDOA, after performing Doppler signal processing (frequency analysis) to analyze Doppler frequencies, it is necessary to estimate the direction of arrival for each Doppler frequency using signal phase differences between a plurality of receiving antennas. There is. Therefore, if objects having the same Doppler velocity (that is, the velocity at which the Doppler frequencies are the same) exist in the observation range, signals in which reflected waves interfere with each other are received. Therefore, the phase information of each object cannot be extracted independently from the received signal. As a result, the direction of arrival cannot be estimated correctly.

  As a technique for reducing such an error in direction-of-arrival estimation, there is a method of arraying receiving antennas and performing array signal processing (see, for example, Non-Patent Document 2).

  Normally, when two people in an observation area are measured using a two-element receiving antenna, a false image is always generated when the two people move the same. Therefore, in this system, as shown in FIG. 25, a radar apparatus having a four-element one-dimensional antenna array is configured based on a CW (Continuous Wave) radar. Furthermore, the radar apparatus can separate two pedestrians in a two-dimensional plane by applying the CLEAN algorithm or the RELAX algorithm.

A. Lin and H. Ling, “Frontal imaging of human using three element Doppler and direction-of-arrival radar,” Electronics Letters, vol. 42, no. 11, pp. 660-661, (2006). Shobha Sundar Ram and Hao Ling, “Through-Wall Tracking of Human Moves Using Joint Doppler and Array Processing L. E. 5, NO. 3, JULY 2008

  However, the conventional radar device described in Non-Patent Document 2 has a problem that when the number of objects existing in the observation area increases, it is necessary to increase the number of receiving antenna elements accordingly. As a result, for example, when the wavelength of the radio wave to be used is 1 cm and the distance between the antenna elements is 0.5 cm which is a half wavelength, the minimum size of the interferometer in which the three elements are arranged in a right isosceles triangle is approximately 0.5 cm. If the number of elements in the horizontal direction is 10, the minimum size is 0.5 cm × 4.5 cm.

  In view of the above-described conventional problems, an object of the present invention is to provide a Doppler radar system capable of improving the detection accuracy of an object position while suppressing an increase in the number of receiving antennas.

  In order to achieve the above object, a Doppler radar system according to an aspect of the present invention is a Doppler radar system that detects a position of a target object, and includes a transmission unit that generates a transmission signal, and the transmission signal as a transmission wave. A transmitting antenna for radiating, a plurality of receiving antennas each receiving a reflected wave reflected by a plurality of reflection points included in the target object, and a plurality of reflections received by the plurality of receiving antennas A reception unit that generates a plurality of reception signals corresponding to each of the waves, and a direction calculation unit that calculates directions of the plurality of reflection points with respect to the Doppler radar system using a phase difference between the plurality of reception signals And using a delay amount between the transmission signal and the reception signal, a distance calculation unit that calculates each distance between the Doppler radar system and the plurality of reflection points; A micro-Doppler computing unit that calculates a plurality of measurement points indicating the positions of the plurality of reflection points using the direction and distance of each of the plurality of reflection points, and a first divided frame having a first interval. A first centroid calculating unit that calculates a first centroid of a measurement point included in each first divided region divided by the first divided frame among the plurality of measurement points; and a first centroid calculating unit that is narrower than the first interval. Second centroid calculation for calculating a second centroid of measurement points included in each of the second divided areas divided by the second divided frame among the plurality of measurement points using a second divided frame having two intervals. And a centroid determining unit that determines at least one second centroid as the position of the target object based on a positional relationship between the second centroid and the first centroid from among the plurality of second centroids. Prepare.

  According to this configuration, the Doppler radar system according to one aspect of the present invention can improve the detection accuracy of the position of the object while suppressing an increase in the number of receiving antennas.

  The second interval may be approximately twice the width of the target object, and the first interval may be twice or more the second interval.

  According to this configuration, the Doppler radar system according to one aspect of the present invention can calculate the second centroid according to the number of target objects.

  The first divided frame includes a third divided frame and a fourth divided frame different from the third divided frame, and the second divided frame is different from the fifth divided frame and the fifth divided frame. And a first center-of-gravity calculation unit that calculates a third center of gravity of the measurement points included in each of the third divided regions divided by the third division frame among the plurality of measurement points. , Out of a plurality of the third centroids, the centroid of the third centroid included in each of the fourth divided regions divided by the fourth division frame is calculated as the first centroid, and the second centroid calculator Of the plurality of measurement points, the fifth centroid of the measurement point included in each fifth divided region divided by the fifth divided frame is calculated, and among the plurality of fifth centroids, divided by the sixth divided frame The center of gravity of the fifth centroid included in each of the sixth divided regions may be calculated as the second centroid.

  According to this configuration, the Doppler radar system according to an aspect of the present invention can reduce the dependency of the division position of the division frame.

  The center-of-gravity determination unit divides the plurality of second centroids into a first group and a second group based on the position of the first centroid, and at least one point from each of the first and second groups. A second center of gravity may be selected, and the selected second center of gravity may be determined as the position of the target object.

  According to this configuration, the Doppler radar system according to an aspect of the present invention can improve the detection accuracy of the position of the object when a false image is generated.

  The center-of-gravity determination unit may calculate the second center of gravity of two points included in the pair for each pair configured by the second center of gravity included in the first group and the second center of gravity included in the second group. A pair centroid that is a centroid is calculated, a pair of the pair centroids closest to the first centroid is selected from the calculated pair centroids, and two second centroids included in the selected pair The position of the target object may be determined.

  Further, the Doppler radar system further includes a tracking prediction unit that predicts a predicted position that is the position of the target object in the future, using the position of the target object in the past determined by the center of gravity determination unit, The tracking prediction unit predicts the number of target objects detected by the Doppler radar system, and the center-of-gravity determination unit is the number of second centroids predicted by the tracking prediction unit among the plurality of second centroids. May be determined as the position of the target object.

  According to this configuration, the Doppler radar system according to an aspect of the present invention can improve the detection accuracy of the position of the object by using the result of tracking prediction.

  In addition, when the first centroid exists between the plurality of predicted positions predicted by the tracking prediction unit, the centroid determination unit is the closest to the predicted position among the plurality of second centroids. When the center of gravity is determined as the position of the target object and the first center of gravity does not exist between the plurality of predicted positions predicted by the tracking prediction unit, the second center of gravity is selected from the plurality of second centers of gravity. Based on the positional relationship between the center of gravity and the first center of gravity, at least one second center of gravity may be determined as the position of the target object.

  According to this configuration, the Doppler radar system according to an aspect of the present invention can perform appropriate processing according to the tracking prediction result.

  The center-of-gravity determination unit divides the plurality of second centroids into a first group and a second group on the basis of the position of the first centroid, and the plurality of the centroids on the basis of the position of the first centroid. Dividing the predicted position into a third group and a fourth group, out of a plurality of second centroids included in the first group, a second centroid closest to each predicted position included in the third group; Of the plurality of second centroids included in two groups, the second centroid closest to each predicted position included in the fourth group may be determined as the position of the target object.

  According to this configuration, the Doppler radar system according to an aspect of the present invention can improve the detection accuracy of the position of the object even when the number of target objects is three or more.

  In addition, when the predicted position is included in both the third group and the fourth group, the center of gravity determination unit includes each of the plurality of second centers of gravity included in the first group included in the third group. The second centroid closest to the predicted position and the second centroid closest to each predicted position included in the fourth group among the plurality of second centroids included in the second group are set as the positions of the target objects. N-1 points included in the first group when the third group includes N (N is an integer of 3 or more) predicted positions and the fourth group does not include a predicted position. And the second center of gravity included in the second group may be determined as the position of the eye object.

  According to this configuration, the Doppler radar system according to an aspect of the present invention can improve the detection accuracy of the position of the object even when the number of target objects is three or more.

  In addition, when a plurality of predicted positions are included in the same distance range, the center-of-gravity determination unit, based on the positional relationship between the second center of gravity and the first center of gravity, from among the plurality of second centers of gravity, One second centroid may be determined as the position of the target object, and when only one predicted position is included in the same distance range, the second centroid may be determined as the position of the target object.

  According to this configuration, the Doppler radar system according to an aspect of the present invention can perform appropriate processing depending on whether or not a false image is generated.

  The present invention can be realized not only as such a Doppler radar system, but also as an object detection method using characteristic means included in the Doppler radar system as a step, and such a characteristic step can be implemented in a computer. It can also be realized as a program to be executed. Needless to say, such a program can be distributed via a non-transitory computer-readable recording medium such as a CD-ROM and a transmission medium such as the Internet.

  Furthermore, the present invention can be realized as a semiconductor integrated circuit (LSI) that realizes part or all of the functions of such a Doppler radar system.

  The present invention can provide a Doppler radar system that can improve the detection accuracy of an object position while suppressing an increase in the number of receiving antennas.

It is a figure which shows the utilization scene of embodiment of this invention. It is a figure which shows the measurement point when there is no interference based on embodiment of this invention. It is a figure which shows the measuring point when there exists interference based on embodiment of this invention. 1 is a block diagram of a radar system according to Embodiment 1 of the present invention. It is a flowchart of the target object detection process which concerns on Embodiment 1 of this invention. It is a figure which shows the structure of the range output which concerns on Embodiment 1 of this invention. It is a figure which shows the accumulated range output based on Embodiment 1 of this invention. It is a figure which shows the structure of the Doppler output which concerns on Embodiment 1 of this invention. It is a figure for demonstrating the calculation process of the arrival direction which concerns on Embodiment 1 of this invention. It is a figure which shows the structure of the direction output which concerns on Embodiment 1 of this invention. It is a figure which shows the structure of the micro Doppler output which concerns on Embodiment 1 of this invention. It is a figure which shows the measurement point in case the reflective power ratio is small based on Embodiment 1 of this invention. It is a figure which shows the measurement point in case the reflection power ratio is large based on Embodiment 1 of this invention. It is a figure which shows distribution of the DOA result in case the reflection power ratio is small based on Embodiment 1 of this invention. It is a figure which shows distribution of the DOA result in case the reflection power ratio is large based on Embodiment 1 of this invention. It is a figure for demonstrating the 1st gravity center calculation process which concerns on Embodiment 1 of this invention. It is a figure for demonstrating the 2nd gravity center calculation process which concerns on Embodiment 1 of this invention. It is a flowchart of the gravity center determination process which concerns on Embodiment 1 of this invention. It is a figure for demonstrating the gravity center determination process which concerns on Embodiment 1 of this invention. It is a figure for demonstrating the gravity center determination process which concerns on Embodiment 1 of this invention. It is a figure which shows the example of a shape of the division frame which concerns on Embodiment 1 of this invention. It is a figure which shows the example of a shape of the division frame which concerns on Embodiment 1 of this invention. It is a figure which shows the example of a shape of the division frame which concerns on Embodiment 1 of this invention. It is a figure which shows the example of a shape of the division frame which concerns on Embodiment 1 of this invention. It is a figure which shows the example of a shape of the division frame which concerns on Embodiment 1 of this invention. It is a figure which shows the example of a shape of the division frame which concerns on Embodiment 1 of this invention. It is a figure which shows the example of a shape of the division frame which concerns on Embodiment 1 of this invention. It is a figure which shows the example of a shape of the division frame which concerns on Embodiment 1 of this invention. It is a figure which shows the example of a shape of the division frame which concerns on Embodiment 1 of this invention. It is a block diagram of the radar system which concerns on Embodiment 2 of this invention. It is a flowchart of the gravity center determination process which concerns on Embodiment 2 of this invention. It is a figure for demonstrating the gravity center determination process which concerns on Embodiment 2 of this invention. It is a figure for demonstrating the gravity center determination process which concerns on Embodiment 2 of this invention. It is a figure for demonstrating the gravity center calculation process which concerns on Embodiment 2 of this invention. It is a flowchart of the gravity center determination process which concerns on the modification of Embodiment 2 of this invention. It is a figure for demonstrating the gravity center determination process which concerns on the modification of Embodiment 2 of this invention. It is a block diagram of the conventional radar apparatus. It is a block diagram of a conventional radar system.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Each of the embodiments described below shows a preferred specific example of the present invention. The numerical values, shapes, materials, constituent elements, arrangement positions and connecting forms of the constituent elements, steps, order of steps, and the like shown in the following embodiments are merely examples, and are not intended to limit the present invention. The present invention is limited only by the claims. Therefore, among the constituent elements in the following embodiments, constituent elements that are not described in the independent claims indicating the highest concept of the present invention are not necessarily required to achieve the object of the present invention. It will be described as constituting a preferred form.

  First, terminology will be described before describing embodiments of the present invention.

  The radar system radiates radio waves from the transmission antenna toward the target object. The radio wave radiated from the transmitting antenna is scattered in various directions according to the shape of the target object. Of the scattered radio waves, radio waves that arrive at the receiving antenna are particularly called “reflected waves”. A point on the target object where the reflected wave is reflected is called a reflection point. The radar system acquires target object information based on the reflected wave.

  The surface of the target object has a large number of scattering points (surface scattering points). The reflection point is a scattering point (hereinafter referred to as a physical reflection point) that is physically determined according to reflection conditions. The radar system acquires target object information by estimating the position of a reflection point and the like.

  When the surface of the target object is a mirror-like sphere or flat plate, one reflection point that satisfies the reflection condition is determined. On the other hand, for example, in the case of a target object having a complicated shape having a large unevenness compared to the wavelength of a radio wave, such as a person, there are a plurality of points on the object that satisfy the reflection condition.

  The position of the physical reflection point and the reflection point obtained by measurement may be significantly different, which is called “false image”. In this specification, a physical reflection point is simply referred to as “reflection point”, and a reflection point obtained from measurement is referred to as “measurement point”.

  FIG. 1 is a schematic diagram showing a situation in which two people (target objects 150a and 150b) are detected and tracked using the radar system 100. FIG. The radar system 100 illustrated in FIG. 1 includes one transmission antenna 120 and three reception antennas 121a, 121b, and 121c. Hereinafter, the receiving antennas 121a, 121b, and 121c may be collectively referred to as the receiving antenna 121.

  The three receiving antennas 121 are installed so as to receive reflected waves in different directions. Specifically, the receiving antenna 121a and the receiving antenna 121b are arranged in the horizontal direction, and the receiving antenna 121a and the receiving antenna 121c are arranged in the vertical direction.

  Thereby, the horizontal direction (namely, azimuth angle) of a reflective point can be measured from the reflected wave received with the receiving antenna 121a and the receiving antenna 121b. Further, the vertical direction (that is, the elevation angle) of the reflection point can be obtained from the reflected waves received by the reception antenna 121a (or the reception antenna 121b) and the reception antenna 121c. Furthermore, the relative velocity between the radar system 100 and the target object can be measured from at least one of the reflected waves received by the receiving antennas 121a, 121b, and 121c.

  For example, the motion speed V (scalar amount in the line-of-sight direction) and the position S (vector amount in spatial coordinates) of the target objects 150a and 150b are represented as <V, S>.

  The head, arm, opposite arm, leg, and opposite leg of the target object 150a are <V11, S11>, <V12, S12>, <V13, S13>, <V14, S14>, <V15, respectively. , S15>, and the head, arm, opposite arm, leg, and opposite leg of the target object 150b are <V21, S21>, <V22, S22>, <V23, S23>, <V24, Observed as S24>, <V25, S25>. In actual observation, a finer motion state can be obtained, but is simplified in the schematic diagram.

  A virtual reflection point is set so as to have a human shape, and a walking motion data is given to each reflection point to construct a model that simulates a moving pedestrian. With respect to this model, measurement points are calculated on a computer by numerical calculation, and the results of observing two people are shown in FIGS. 2A and 2B.

  FIG. 2A shows a case where the target object 150a and the target object 150b exist and there is no interference in the Doppler speed. FIG. 2B shows a case where the target object 150a and the target object 150b have reflection points at which the Doppler velocities are the same, that is, there is interference. 2A and 2B show a result (plan view) obtained by projecting a three-dimensional <V, S> distribution onto a two-dimensional plane, and each point corresponds to a measurement point. The origin indicates the position of the radar system 100.

  As shown in FIG. 2A, when the Doppler velocities of a plurality of humans (the target object 150a and the target object 150b) do not interfere with each other, the distribution of the measurement points of the target object 150a and the target object 150b is clear. Both can be separated.

  However, as shown in FIG. 2B, when a plurality of people includes the same Doppler velocity, the distribution of <V, S> is greatly widened, making it difficult to separate the two. This is a false image due to interference of Doppler velocity.

  Thus, if there is an error in the direction of arrival estimation result, the estimation of the height or width related to the direction in which the object exists or the shape of the object will be wrong, and the imaging result will be disturbed. This is called a false image due to interference of Doppler velocity.

  Embodiments of the present invention will be described below.

(Embodiment 1)
FIG. 3 is a block diagram showing a configuration of radar system 100 according to Embodiment 1 of the present invention. A radar system 100 shown in FIG. 3 is a Doppler radar system that specifies the position of a target object 150. The radar system 100 includes a transmission unit 101a, a reception unit 101b, a distance calculation unit 102, a Doppler analysis unit 103, a direction calculation unit 104, a micro Doppler calculation unit 105, a first centroid calculation unit 106a, 2 centroid calculating unit 106b, centroid determining unit 107, transmitting antenna 120, and receiving antennas 121a, 121b and 121c.

  The radar system 100 shown in FIG. 3 measures the distance, direction, and Doppler velocity of the target object 150. Here, the distance and direction are the distance and direction with reference to the radar system 100. Further, the radar system 100 determines the position of the measurement point on the target object from the distance and direction, and further associates the Doppler speed of the measurement point. Then, the radar system 100 estimates the shape and motion of the target object 150 by acquiring a plurality of measurement points constituting the target object 150.

  First, an outline of the operation of the radar system 100 will be described.

  FIG. 4 is a flowchart of processing for detecting a target object by the radar system 100.

  First, the transmission unit 101a generates a transmission signal. Then, the transmission antenna 120 radiates the transmission signal as a transmission wave (S101).

  Each of the plurality of reception antennas 121 a, 121 b, and 121 c receives a reflected wave in which a transmission wave is reflected by a plurality of reflection points included in the target object 150. Then, the reception unit 101b generates a plurality of reception signals 111 corresponding to each of the plurality of reflected waves received by the plurality of reception antennas 121a, 121b, and 121c (S102).

  Next, the distance calculation unit 102 calculates the distance between the Doppler radar system 100 and a plurality of reflection points using the delay amount between the transmission signal and the reception signal 111 (S103).

  Next, the Doppler analysis unit 103 calculates the Doppler frequency of each received signal by analyzing the frequency component of each received signal (S104).

  Next, the direction calculation unit 104 calculates the directions of the plurality of reflection points with respect to the Doppler radar system 100 using the phase differences between the plurality of received signals for each Doppler frequency (S105).

  Next, the micro Doppler computing unit 105 calculates a plurality of measurement points indicating the positions of the plurality of reflection points using the directions and distances of the plurality of reflection points calculated in steps S103 and S105 (S106).

  Next, the first center-of-gravity calculation unit 106a uses the first divided frame having the first first interval, and the measurement included in each first divided region divided by the first divided frame among the plurality of measurement points. The first centroid 116a of the point is calculated (S107).

  In addition, the second center-of-gravity calculation unit 106b uses the second divided frame having the second interval narrower than the first interval, and is included in each second divided region divided by the second divided frame among the plurality of measurement points. The second centroid 116b of the measurement point to be measured is calculated (S108).

  Finally, the center-of-gravity determination unit 107 sets at least one second center of gravity 116b to the position of the target object 150 based on the positional relationship between the second center of gravity 116b and the first center of gravity 116a among the plurality of second centers of gravity 116b. Determine (S109).

  Hereinafter, each configuration will be described in detail.

<Transmitter 101a>
The transmission unit 101a generates a transmission signal using a carrier wave in a desired frequency band. The desired frequency band is, for example, a microwave or millimeter wave frequency.

  The transmission antenna 120 radiates the transmission signal generated by the transmission unit 101a to the external space as a transmission wave.

<Receiving unit 101b>
The receiving antennas 121a to 121c receive reflected waves that are transmitted waves reflected by the target object 150. Hereinafter, a wave emitted from the transmission antenna 120 and reaching the target object 150 is referred to as a transmission wave, and a wave reflected by the target object 150 is referred to as a reflected wave.

  The reception unit 101b generates a reception signal 111 corresponding to the reflected wave received by each reception antenna 121. Here, the reception signal 111 includes three reception signals corresponding to the reflected waves received by the reception antennas 121a, 121b, and 121c. Hereinafter, a signal and a processing unit corresponding to each reception antenna are also referred to as a reception channel.

  Specifically, the reception unit 101b performs the following processing for each reception signal 111 of each reception channel. The receiving unit 101b converts the reflected wave into a signal of a predetermined frequency band by demodulating the reflected wave using a signal having the same frequency as the carrier wave or the same carrier wave. Then, the reception unit 101b outputs the converted signal as the reception signal 111. Here, the received signal 111 includes fluctuations in signal amplitude corresponding to the presence / absence of reflection and the distance. Therefore, the distance from the received signal 111 to the reflection point can be calculated. The predetermined frequency band is a frequency band including a Doppler frequency centered on a DC component having a frequency of zero.

  More specifically, the receiving unit 101b distributes the carrier wave to the first signal and the second signal. The receiving unit 101b shifts the phase of the first signal by 90 degrees and does not shift the phase of the second signal. Next, the reception unit 101b generates a first reception signal and a second reception signal as the reception signal 111 by demodulating the reflected wave using the first signal and the second signal. The first received signal and the second received signal are orthogonal to each other. The second received signal 111 means an I-phase signal (In-Phase). The first received signal 111 means a Q-phase signal (Quadrature). That is, the reception signal 111 of each reception channel is a complex signal having the second reception signal as a real part and the first reception signal as an imaginary part.

  The radar system 100 includes at least two receiving antennas 121a and 121b. The radar system 100 acquires a distribution of measurement points on a two-dimensional plane or a three-dimensional space using the phase difference between the reflected wave received by the receiving antenna 121a and the reflected wave received by the receiving antenna 121b. Thus, the shape, speed, or position of the target object 150 is measured. Details of this will be described later.

  The radar system 100 may include a plurality of distance calculation units 102 and Doppler analysis units 103 according to the number of reception antennas 121. Specifically, the radar system 100 includes a number of distance calculation units 102 and Doppler analysis units 103 corresponding to the number of channels when a set of two receiving antennas 121 is one channel.

  For example, when the radar system 100 has the receiving antennas 121a, 121b, and 121c, the radar system 100 has three channels. Specifically, there are three sets: a set of the receiving antenna 121a and the receiving antenna 121b, a set of the receiving antenna 121b and the receiving antenna 121c, and a set of the receiving antenna 121a and the receiving antenna 121b. At this time, the radar system 100 includes three Doppler analysis units corresponding to the respective channels.

<Distance calculation unit 102>
The distance calculation unit 102 acquires the reception signal 111 from the reception unit 101b. Then, the distance calculation unit 102 calculates the distance between the radar system 100 and the plurality of reflection points using the delay amount between the transmission signal and the reception signal 111.

  Here, the method of determining the position of the target object 150 differs according to the transmission signal generated by the transmission unit 101a. For example, when the transmission unit 101a uses a pseudo-noise code to generate a transmission signal in which a carrier wave is modulated, a method for measuring the position of the target object 150 is a spread spectrum method.

  On the other hand, when the transmission unit 101a generates a transmission signal in which a carrier wave is chirp-modulated, the method of measuring the position of the target object 150 is an FM-CW (Frequency Modulated Continuous Wave) method. Further, when the transmission unit 101a uses two types of carrier waves having different frequencies and transmits both of them alternately, the method of measuring the position of the target object 150 is a two-frequency CW method.

  In addition, when the transmission part 101a does not modulate a carrier wave at all, it becomes a CW (Continuous Wave) system, but distance measurement cannot be performed in principle in the CW system.

  In this embodiment, a case where a spread spectrum method is used will be described as an example.

  Spread spectrum radar modulates (spreads) a carrier wave using a pseudo-noise code at the time of transmission, and reproduces a pseudo-noise code by demodulating (despreading) a reflected wave at the time of reception. Evaluate the amount of delay. Then, the radar calculates the distance to the target object from the delay amount.

  That is, the distance calculation unit 102 uses a transmission pseudo-noise code (hereinafter referred to as a transmission code) and a reception pseudo-noise code (hereinafter referred to as a reception code) obtained by shifting the transmission code. The delay of the transmission code caused by (that is, reflection) is calculated. Specifically, the distance calculation unit 102 calculates the correlation between the transmission code and the reception code, and calculates the distance from the code shift amount of the reception code when the correlation reaches a maximum value equal to or greater than a predetermined threshold. To do. Further, when there is no reflected wave, there is no maximum value exceeding the threshold value, so the distance is not determined. When reflected waves from different distances arrive, there are as many local maximum values that exceed the threshold, so that the respective distances can be determined.

  When the spread spectrum method is used, the distance calculation unit 102 determines the distance (R) using the following (Equation 3).

  R = cD / 2b (Formula 3)

  Here, b [bit / sec] ([chip / sec]) is a pseudo-noise code rate, D [bit] ([chip]) is a code shift amount at the time of maximum correlation, and c [m / sec ] Is the speed of light.

  A value when the shift amount D = 1 is referred to as range resolution. The distance R is determined discretely in units of range resolution (referred to as range bins), and signal processing can be performed independently for each range bin.

  Here, the range resolution is ΔR. When the maximum detection distance that can be detected by the radar system 100 is Rmax and M is an integer, Rmax = M × ΔR. When i is an integer of 1 ≦ i ≦ M, a distance range R (i) = i × ΔR is determined corresponding to the code shift amount D = i. This index i is called a distance range number. Thus, the distance range R (i) indicates a range of the distance r that satisfies (i−1) × ΔR <r ≦ i × ΔR.

  The output of the distance calculation unit 102 is referred to as a range output 112. Assume that the range output 112 is also in complex form.

  If the spread spectrum method is not used, the distance range is assumed to be composed of only a single R (1), and the distance measured by each method may be assigned to R (1). In addition, when there is no distance information as in the CW method, it is preferable to assign a physically meaningless numerical value such as R (1) = ∞.

  In addition, it is good to store the processing result of the signal performed in the distance calculation part 102, the Doppler analysis part 103 mentioned later, and the direction calculation part 104 in a memory | storage part (not shown). Furthermore, each processing unit may be configured to add, update, and delete processing results with respect to the storage unit. As described above, the time width (that is, the window length) of the observation can be freely set for the time series data of the range output 112 through the storage unit.

  FIG. 5A is a diagram illustrating the configuration of the range output 112. As shown in FIG. 5A, the range output 112 includes an IQ signal for each distance range number. Further, each IQ signal includes three IQ signals (sa (i), sb (i), sc (i)) corresponding to each of the receiving antennas 121a to 121c.

  FIG. 5B is a diagram illustrating the range output 112 stored in the storage unit. As shown in FIG. 5B, the time-series range output 112 is accumulated in the storage unit.

<Doppler analysis unit 103>
The Doppler analysis unit 103 analyzes the frequency component of the range output 112, calculates the Doppler frequency component included in the range output 112, its power and phase, and outputs the calculation result as the Doppler output 113.

  The frequency analysis method is, for example, fast Fourier transform (FFT). When analyzing a phenomenon limited to a set window length, a short-time Fourier transform (STFT) is used. Here, this STFT is simply expressed as FFT. Further, in order to explain the meaning of the frequency analysis processing in an easy-to-understand manner, this FFT is also referred to as Doppler signal processing.

  Hereinafter, a case where a specific distance range R (i) included in the range output 112 is extracted and frequency analysis is performed will be described.

  Here, as shown in FIG. 5B, the range output 112 is stored in the storage unit in time series. When N is an integer equal to or greater than 1, the Doppler analysis unit 103 selects N sets from the latest one of the plurality of range outputs 112 stored in the storage unit. For example, the range output (IQ signal) in the distance range R (i) is represented by the symbol s (i), and the IQ signal for the sampling number h (1 ≦ h ≦ N) is represented by s (i, h). A time-series signal of N samples in R (i) is obtained. However, the indexing is performed so that h = 1 is the latest data and h = N is the oldest data. The Doppler analysis unit 103 performs FFT on the N-sample time-series signal s (i, h) (1 ≦ h ≦ N). This corresponds to performing FFT with a window length of N for each distance range.

  Note that since the capacity of the storage unit is finite, it is preferable to delete old IQ signals that are no longer needed from the storage unit.

  Here, the range output 112 is a complex number, and the FFT result (Doppler spectrum) includes a power spectrum representing the frequency characteristic of the power in the range output 112 and a phase spectrum representing the frequency characteristic of the phase.

  Therefore, if the analyzed Doppler spectrum is used, the peak of the spectrum can be detected, and the power and phase with respect to the peak position (Doppler frequency) can be read.

  And the Doppler analysis part 103 stores the Doppler output 113 containing these information in a memory | storage part.

  FIG. 5C is a diagram illustrating a configuration of the Doppler output 113. As shown in FIG. 5C, the Doppler output 113 includes a Doppler spectrum peak frequency (ie, Doppler frequency) F, power P, and phase φ. Moreover, since the peak is not necessarily single, it is represented as a set {F, P, φ}.

  In addition, when the Doppler analysis unit 103 detects a peak in the Doppler spectrum, the power P at a point that is not a peak is rewritten to zero or a negative number (for example, −1 or −P), thereby determining whether the power P is a positive number. Whether it is a peak or not can be determined. The unit of Doppler frequency is called Doppler bin, and the index is called “Doppler frequency number”.

  At this time, the relationship between the Doppler frequency number k and the Doppler frequency F is expressed by a linear function. For example, when the FFT frequency resolution is ΔF and the Doppler frequency number N / 2 is a DC component (frequency zero), the Doppler frequency corresponding to the Doppler frequency number k is F = (k−N / 2) ΔF.

  Further, the Doppler frequency F and the Doppler velocity Vd are in a proportional relationship and can be converted to each other. For example, Vd = γF is represented by a proportional coefficient γ. That is, the discussion about whether the Doppler velocities are the same is the same as the discussion about whether the Doppler frequencies are the same, and will be described below using the “Doppler frequency”. However, “interference of Doppler velocity” is expressed by “Doppler velocity”.

  The Doppler analysis unit 103 reveals the line-of-sight direction component of each motion speed for a plurality of motion sites constituting the target object 150.

  The Doppler outputs 113 exist in parallel for the number of reception channels. For example, when the number of reception channels is 3, they are distinguished as Doppler outputs 113a, 113b, and 113c, and correspond to the reception antennas 121a, 121b, and 121c, respectively. Note that when only the Doppler output 113 is expressed, it is assumed that all are included without distinction of reception channels.

  Also, a set K of elements constituting the Doppler output 113 (generally having a plurality of peaks) is distinguished as Ka, Kb and Kc so as to correspond to them, and Ka = {Fa, Pa, φa}, Kb = { Fb, Pb, φb}, Kc = {Fc, Pc, φc}.

  Further, using the Doppler frequency number k (an integer satisfying 1 ≦ k ≦ N) and the distance range number i (1 ≦ i ≦ M), the element Ka (i, k) = (Fa (i, k)) , Pa (i, k), φa (i, k)). It should be noted that the element index is omitted, and is simply expressed as Ka (i, k) = (Fa, Pa, φa). This is because the kth frequency Fa (i, k) and the power Pa (i, k) corresponding to the kth frequency Fa (i, k) in the result of Doppler signal processing for the i-th distance range of the range output 112 of the signal received by the receiving antenna 121a. , Phase φa (i, k) is shown. The same applies to other cases.

<Direction calculation unit 104>
The direction calculation unit 104 calculates the directions of the plurality of reflection points with respect to the radar system 100 using the phase differences between the plurality of received signals. Specifically, the direction calculation unit 104 receives Doppler outputs 113a, 113b, and 113c.

  The direction calculation unit 104 calculates the phase difference between the reception channels using the phases φa, φb, and φc included in the Doppler outputs 113a, 113b, and 113c. At this time, the phase differences between the reception channels are Δφab = φb−φa and Δφac = φc−φa. Forming a pair with the Doppler output 113 and comparing the phases of the pair in this way means that “the interferometer is configured”.

  In order to simplify the description, the index (i, k) based on the distance range number i and the Doppler frequency number k is omitted here.

  The direction calculation unit 104 calculates the arrival direction using this phase difference.

  FIG. 6 is a diagram for explaining the calculation process of the arrival direction.

  First, the direction calculation unit 104 calculates the phase difference Δφab from the phase φa corresponding to the receiving antenna 121a and the phase φb corresponding to the receiving antenna 121b. Then, the direction calculation unit 104 calculates the arrival direction θ by substituting this phase difference Δφab into Δφ in (Expression 4) below.

θ = sin ⁻¹ (λΔφ / 2πd) (Formula 4)

  Here, λ is the wavelength of the transmission wave, and d is the distance between the reception antenna 121a and the reception antenna 121b. Further, λΔφ is a path difference in an interferometer configured by the receiving antenna 121a and the receiving antenna 121b.

  Then, direction calculation section 104 outputs arrival direction θ calculated using phase difference Δφab as azimuth angle A.

  Further, the direction calculation unit 104 outputs the arrival direction θ calculated using (Equation 4) and the phase difference Δφac as the elevation angle E.

  FIG. 7A is a diagram illustrating the configuration of the direction output 114. As shown in FIG. 7A, the direction output 114, which is the output of the direction calculation unit 104, includes information on the azimuth angle A and the elevation angle E with respect to the measurement point of the Doppler frequency F. That is, the direction output 114 includes A (i, k) and E (i, k) for F (i, k).

  At this time, the contents of the direction output 114 (elements of the set) are expressed as L (i, k) = (A (i, k), E (i, k)). This is simply expressed as L (i, k) = (A, E).

<Micro Doppler operation unit 105>
The micro Doppler computing unit 105 calculates measurement points that are positions of the plurality of reflection points using the directions and distances of the plurality of reflection points.

  Specifically, the range output 112, the Doppler output 113, and the direction output 114 are input to the micro Doppler calculation unit 105.

  The micro Doppler computing unit 105 includes information {R} on the distance R included in the range output 112, information {F, P, φ} on the frequency F, power P, and phase φ included in the Doppler output 113, and direction output. The information {A, E} of the azimuth angle A and the elevation angle E included in 114 is correlated with each other by the index of the distance range number and the extracted peak number, so that the distance R and power P of the measurement point having the Doppler frequency F are correlated. Determine the azimuth angle A and elevation angle E.

  More specifically, the micro Doppler computing unit 105 performs range output R (i), Doppler output K (i, k) = (F, P, φ), and direction output L (i, k) = (A , E) and accessing each output with index numbers 1 ≦ i ≦ M and 1 ≦ k ≦ N, the measurement point Qp (i, k) = (F, P, R, A, E) To decide. Then, the micro Doppler computing unit 105 outputs this Qp as the micro Doppler output 115.

  FIG. 7B is a diagram showing a configuration of the micro Doppler output 115.

  Note that the micro Doppler computing unit 105 may convert the measurement variables (R, A, E) forming polar coordinates into measurement variables (X, Y, Z) in an orthogonal coordinate system as necessary, and output them. . In this case, measurement point Qr (i, k) = (F, P, X, Y, Z). X, Y, and Z are coordinate values in the left-right direction (lateral direction, spread), the front-rear direction (vertical direction, depth), and the vertical direction (height direction) with respect to the front direction of the radar.

  When the number of reception channels is 2, since only two-dimensional positioning can be performed, the above description can be applied as it is by regarding all elevation angles E as E = 0.

  The micro Doppler output 115 is processed by the first centroid calculating unit 106a and the centroid determining unit 107, and finally the centroid position of the target object 150 is output. Further, the number of barycentric positions corresponding to the number of target objects is output.

  Note that the micro Doppler computing unit 105 may interpolate a distance equal to or less than the range resolution using a waveform obtained by redisplaying the reflected power in the distance range direction for each Doppler frequency. By doing so, the distance determination accuracy can be improved.

  Before describing the details of the center-of-gravity calculation unit 106 and the center-of-gravity determination unit 107, the features of the present embodiment based on the knowledge found by the present inventors will be described with reference to FIGS. 8A, 8B, 9A, and 9B. To do.

  8A and 8B are diagrams illustrating measurement points when two reflection points having the same Doppler frequency are observed in the same range. The size of the reflection point in FIGS. 8A and 8B represents the reflection power (intensity P).

  FIG. 8A is a diagram illustrating measurement points when there are two reflection points 161a and 161b having substantially the same reflection power. Here, substantially the same reflection power can be read as substantially the same scattering cross section. That is, for example, the reflection points 161a and 161b are two trunks of different people.

  In such an example, it is experimentally known that the position of the measurement point 131 is an intermediate position between the reflection points 161a and 161b.

  On the other hand, FIG. 8B is a diagram illustrating measurement points when there are two reflection points 162a and 162b having a large difference in reflection power. For example, the reflection point 162a is the left leg, and the reflection point 162b is the right arm.

  When the movement is viewed in time series, the body is directed almost in one direction and the fluctuation of the reflected power (scattering cross section) is small, whereas the directions of the legs and arms change every moment. Therefore, their scattering cross sections appear to change. Therefore, it is empirically known that when the difference in reflected power is large, the position of the measurement point 132 becomes unpredictable due to such time-series fluctuation of the scattering cross section.

  In the first place, what happens qualitatively when a reflected wave with the same Doppler frequency arrives will be described using a simple model (formula).

  The reflected waves arriving at the receiving antenna from the individual reflection points are denoted by s1 (t), s2 (2), s3 (t),.

  At this time, the waveform of the radio wave actually received by the receiving antenna is obtained by superimposing (combining) these. If only the components of the same Doppler frequency that interfere with each other are focused, the frequency term becomes common, so that only the phase term needs to be focused. This phase term is expressed by (Equation 6).

  From this (Equation 6), the combined phase Φ is as shown in (Equation 7).

  The arrival direction DOA is expressed as (Equation 8) using the phase difference ΔΦ between the receiving antennas of the combined phase Φ and the constant K. Here, since the DOA is simple, it is assumed that the number of reception channels is two.

  Subsequently, using these equations, it was verified how the DOA is affected by the ratio of the reflected power, that is, the fluctuation of the reflected power.

  9A and 9B are diagrams illustrating the results of calculating the influence of the power difference between two reflection points on the arrival direction estimation. It is assumed that the two reflection points are in the directions of + π / 4 and −π / 4, respectively, and K = 1 / π (corresponding to the case where the interval between the reception antennas is half of the transmission wavelength).

  FIG. 9A shows a case where the ratio of the reflected power of the two reflection points varies randomly at 3 dB or less, and FIG. 9B shows a case where the ratio of the reflected power varies randomly at 26 dB or less. The power fluctuations at the two reflection points may be considered as results within 3 dB and 26 dB, respectively. FIG. 9A and FIG. 9B are diagrams showing the distribution of DOA results calculated for the power fluctuation in a histogram. The number of trials is 5000.

  As shown in FIG. 9A, when the difference between the powers of the two reflection points is small, measurement is performed within a range of ± 0.05 rad at 0 rad near the middle of ± π / 4, that is, within a range of 1 m wide at a point 10 m away. The points are concentrated.

  On the other hand, as shown in FIG. 9B, when the difference in power between the two reflection points is large, the measurement points are widely spread and distributed in a range of 5 m wide at a point 10 m away.

  From this, it can be seen that the distribution of the measurement points is large in the motion region where the power of the reflection point (scattering cross section) varies greatly. Furthermore, when the state of the reflection point changes from moment to moment, the position of the measurement point moves from moment to moment, resulting in an indeterminate state. This is called an “unstable measurement point”.

  In other words, if there are two reflection points with relatively high power (such as two human torso) whose reflection powers are substantially the same, a measurement point with high power appears at an intermediate position between the reflection points. I understand that. This is referred to as a “metastable measurement point”.

  The present invention provides means for estimating the position of the original reflection point of the target object from the false image generated by the interference of the Doppler velocity by using this metastable measurement point.

  Considering the principle and significance of metastable measurement points, it is assumed that there are two equivalent measurement points. For this reason, the present embodiment is characterized in that when a false image is generated, the number of target objects is regarded as a maximum of two, so that the position of the object can be easily estimated and the two can be separated.

<Center of gravity calculation unit 106>
Next, the center-of-gravity calculation unit 106 will be described with reference to FIGS. 10 and 11.

  The centroid calculating unit 106 includes a first centroid calculating unit 106a and a second centroid calculating unit 106b.

  The first centroid calculating unit 106a calculates the first centroid 116a of the reflection points included in each divided frame among the plurality of reflection points, using the first divided frames having the first interval. The second center-of-gravity calculation unit 106b calculates the second center of gravity 116b of the reflection points included in each division frame among the plurality of reflection points using the second division frame having the second interval narrower than the first interval. .

  That is, the centroid calculating unit 106 divides the observation area into several areas and calculates the centroid for each divided area. A partition for dividing an area is expressed as a “division frame”. In addition, each region divided by the division frame is expressed as a “divided region”.

  The division frame is an arithmetic concept. For example, the whole observation region is represented by −2a ≦ x ≦ 2a using a real number x representing the position in the observation region. When this entire observation region is divided into four regions (1) -2a ≦ x <−a, (2) -a ≦ x <0, (3) 0 ≦ x <a, (4) a ≦ x ≦ 2a, x = 0, ± a, and ± 2a indicate division frames.

  Although the above example is a case of a one-dimensional divided frame, a two-dimensional divided frame can be formed by using two real numbers (x, y). For example, if a divided frame is formed by x = 0, ± a, ± 2a,..., Y = 0, ± 2a, ± 4a,. It can be divided into a plurality of divided areas consisting of a rectangle of the side 2a.

  In this embodiment, an example in which a square monotonous divided frame (square divided frame) in an orthogonal coordinate system is used will be described.

  When the square division frame is used, there is an advantage that a plurality of target objects can be separated on the basis of the lattice by setting the lattice interval to a size suitable for the target object. The reason for this will be described later in the method of setting the lattice spacing.

  First, the description of the pattern of the division frame will be described.

  A basic lattice pattern of a square division frame having a length a on one side is provided, which is represented as SQ [a]. Here, the basic lattice pattern is, for example, a lattice pattern provided so that the lattice intersection of the divided frame coincides with the origin of the measurement system. A lattice pattern obtained by translating the basic lattice pattern SQ [a] by Δ1 × a in the x direction and Δ2 × a in the y direction is represented as SQ [a | Δ1, Δ2].

For example, when expressed as SQ [a | 0.5, 0.5], a lattice pattern obtained by translating the basic lattice pattern SQ [a] by + 0.5a in the x direction and + 0.5a in the y direction is shown.
Note that SQ [a | 0,0] = SQ [a].

  Further, when a specific divided area is indicated from among a plurality of divided areas, an identification number and an identification address are added as identifiers as in SQ [a | 0,0] (“identifier”).

  For example, if the identification symbol “S1” is assigned to the divided area of the address (1, −3) of the divided frame SQ [a | 0,0], this is represented by SQ [a | 0,0] (S1 ) Or SQ [a | 0,0] (1, -3).

  In this embodiment, four patterns of square division frames in which different side lengths and different positional relationships (translation) are combined are used.

  That is, there are four patterns of SQ [D1 | 0,0], SQ [D1 | 0.5,0.5], SQ [D2 | 0,0], SQ [D2 | 0.5,0.5]. . Note that D1 ≠ D2.

  Here, the divided frames constituted by the lattice pattern are sequentially divided into a first divided frame A (third divided frame), a first divided frame B (fourth divided frame), and a second divided frame A (fifth divided frame). Frame) and second divided frame B (sixth divided frame).

  The first divided frames A and B are used by the first centroid calculating unit 106a, and the second divided frames A and B are used by the second centroid calculating unit 106b.

  FIG. 10A, FIG. 10B, FIG. 11A, and FIG. 11B show examples of processing using the first divided frames A and B and the second divided frames A and B, respectively. FIG.

  As shown in FIG. 10A, the first centroid calculating unit 106a uses the first divided frame A, and the first centroid A (the third centroid) that is the centroid of the measurement point of the division unit (within the divided region) of the divided frame. (Centroid) is calculated. Next, as shown in FIG. 10B, the first centroid calculating unit 106a uses the first divided frame B, and the first centroid B (first centroid 116a) that is the centroid of the first centroid A in the divided region. Is calculated.

  Here, if the center of gravity is calculated using only one divided frame, the result depends on the dividing position of the divided frame. On the other hand, by using two patterns of divided frames A and B that are translated from each other, the dependence of the separation position can be reduced. For example, when only one pattern of divided frames is used, if a single target object is divided by the divided frames, two centroids may be obtained. However, by using two patterns of divided frames, the two centroids are merged into one again. The first centroids A and B are calculated at a plurality of points according to the number and size of the target objects.

  Similarly, as shown in FIG. 11A, the second centroid calculating unit 106b uses the second divided frame A to calculate the second centroid A (fourth centroid) that is the centroid of the measurement point for each divided region. To do. Next, as illustrated in FIG. 11B, the second centroid calculating unit 106b uses the second divided frame B, and the second centroid B (second centroid 116b) that is the centroid of the second centroid A in the divided region. Is calculated. The second centroids A and B are calculated at a plurality of points according to the number and size of target objects.

  An equation for calculating the center of gravity is shown in (Equation 9). As described above, the calculation is performed for each divided region without distinguishing the Doppler frequency.

  As shown in (Equation 9), the center-of-gravity calculation unit 106 has a micro Doppler measurement point group (set) obtained from the target object Qp = {F, P, R, A, E} (polar coordinate system) or Qr = { For F, P, X, Y, Z} (orthogonal coordinate system), a power weighted average is obtained for each region. The center-of-gravity calculation unit 106 may determine the power weighted average for both of them, not limited to either Qp or Qr. Here, since the power weighted average is an average amount reflecting the power of the reflection point or the size of the scattering cross section, it represents a macro motion of the target object. In other words, this calculation method is equivalent to an operation that regards a part having a large scattering cross-sectional area such as a body part as a human movement center, and is a calculation method that gives a result consistent with an actual human movement trajectory.

  The center of gravity is a name including not only the average of the positions but also the average value of the Doppler frequency and the average value of the reflected power. Here, the center of gravity is an average value of position, velocity (Doppler frequency), and reflected power, but only the average value of the position may be used as the center of gravity, and at least one of the average value of velocity and reflected power. The average value of the positions may be used.

  Next, the setting of the lattice spacing (side length) of the divided frames will be described.

  In this embodiment, a human such as a pedestrian is assumed as the target object. In the first embodiment, it is preferable to set D2 = 2Wh and D1 = 6Wh (D1 / D2 = 3) when the human lateral width is Wh. For example, a horizontal value of 0.5 m may be adopted and a fixed value of Wh = 0.5 m may be used.

  With this setting, as a result of the second centroid calculation process shown in FIG. 11, the second centroid B is determined for each D2 / 2, that is, for each Wh square area. In other words, this means that the position of the center of gravity can be obtained in units of the size of a person assumed as the target object. Therefore, even if there are a plurality of target objects (humans), the center of gravity corresponding to the number of target objects is calculated.

  In the first center-of-gravity calculation process shown in FIG. 10, it is preferable to use the first divided frame having an interval three times that of the second divided frame. If discussed in the same manner as in FIG. 11, the first center of gravity B is determined for each point of D1 / 2, that is, 3 Wh square. This corresponds to, for example, a setting in which the center of gravity can be separated when two pedestrians are separated from each other.

  That is, the first center-of-gravity calculation unit 106a aims to obtain the above-mentioned “metastable measurement point”. Therefore, a frame for placing another one-point centroid between the centroids of two people is required. When a false image is generated due to interference of Doppler velocity, a margin of “one person” is necessary as a frame for calculating a result of “metastable measurement point”. Therefore, it can be said that it is desirable to set D1 / 2 = 3Wh. However, it is not necessary to strictly satisfy D1 / D2 = 3, and it may be set in a range of D1 / D2> 2.

  The interval between the divided frames does not need to be a fixed value, and may be increased or decreased dynamically. Specifically, the center of gravity reflecting the state of the object can be calculated by changing the interval between the divided frames according to the state of the object in the observation region. For example, if there are two people, increase the interval so that the two people can be handled at once, or increase the interval even if the false image is too remarkable, and deliberately reduce the number of calculated centers of gravity. be able to.

  Further, here, a case where the target object is a human is described as an example, but the target object may be other than a human. In this case, Wh is the size of the target object. That is, D2 is approximately twice the width of the target object, and D1 is preferably at least twice as large as D2. Moreover, “substantially twice” means, for example, 1.5 to 2.5 times.

  As described above, in the centroid calculating unit 106, the first centroid calculating unit 106a calculates a combined centroid of a plurality of (especially two) objects or a metastable measurement point at the time of Doppler velocity interference, and the second centroid. The calculation unit 106b calculates individual centroids of a plurality of objects or individual centroid candidates at the time of Doppler velocity interference.

  Hereinafter, actual processing will be described with an example.

  First, the calculation process of the 1st gravity center 116a is demonstrated using FIG. The divided areas are simply described using A01 and B03 shown in the figure. For example, SQ [6Wh | 0,0] (A01) is simply referred to as region A01, and SQ [6Wh | 0.5,0.5] (B03) is simply referred to as region B03.

  As shown in FIG. 10A, since the measurement points are distributed in the regions A03 and A04, the first centroid calculating unit 106a is the first centroid in the region A03 and the first centroid in the region A04. Each center of gravity A is calculated. At this time, the first center of gravity A is represented as G (A03) and G (A04).

  Next, as shown in FIG. 10B, the first center of gravity A has G (A03) and G (A04) in the region B01, and there is no other. Therefore, the first centroid calculating unit 106a calculates the first centroid B in the region B01 using the centroids G (A03) and G (A04). The first centroid calculating unit 106a uses (Equation 9) as a calculation formula, and substitutes the value of the first centroid A into the (Equation 9) as the value of the measurement point. Thus, G (B01) is calculated as the first center of gravity B.

  In this example, only one result of the first center of gravity B is calculated, but the result naturally varies depending on the distribution of measurement points.

  In addition, since the first divided frame used in the first centroid calculating unit 106a is wider than the second divided frame used in the second centroid calculating unit 106b, the first centroid B is referred to as a “wide-area centroid” in an easily understandable manner. There is.

  Next, the second centroid calculation process will be described with reference to FIG.

  The calculation process of the second centroid is the same as the calculation process of the first centroid except that the lattice pattern to be used is different from the calculation process of the first centroid. Therefore, the calculated center of gravity results are shown in the figure by indicating the second center of gravity A as g (a1), the second center of gravity B as g (b1), and the like. g (b1) is the center of gravity of g (a1) and g (a2), g (b2) is the center of gravity of g (a3) and g (a4), and g (b3) is g (a5) The same.

  Further, the second centroid B may be referred to as an “individual centroid” in contrast to the “wide centroid” of the first centroid B.

  As described above, the first centroid calculating unit 106a outputs, for example, G (B01) as the first centroid 116a. The second centroid calculating unit 106b outputs, for example, g (b1), g (b2), and g (b3) as the second centroid 116b.

<Center of gravity determination unit 107>
The center-of-gravity determination unit 107 receives a first center of gravity 116a (wide-area center of gravity) and a second center of gravity 116b (individual center of gravity). The center-of-gravity determination unit 107 extracts at least one center of gravity as the position of the target object from the second center of gravity 116b based on the positional relationship between the second center of gravity 116b and the first center of gravity 116a. That is, the center-of-gravity determination unit 107 has a function of selecting a center of gravity corresponding to the number of true target objects using the first center of gravity 116a and the second center of gravity 116b.

  The center-of-gravity determination method will be described using the flowchart of FIG.

  First, the center-of-gravity determination unit 107 determines whether the first center of gravity 116a and the second center of gravity 116b are one each in the entire region (S201).

  When there is one first centroid 116a and one second centroid 116b in all regions (Yes in S201), the centroid determination unit 107 determines that the target object (person) is one, and the second centroid 116b is determined as the position of the target object (S211).

  On the other hand, when there are a plurality of first centroids 116a and second centroids 116b in the entire region (No in S201), the centroid determination unit 107 performs the following process for each distance range. Here, since interference does not occur when the distance ranges are different, it is not necessary to consider the influence of interference between the distance ranges.

  First, the centroid determining unit 107 separates the first centroid 116a and the second centroid 116b for each distance range (S202).

  Next, the center-of-gravity determination unit 107 selects one of the distance ranges including the first center of gravity 116a and the second center of gravity 116b (S203).

  Next, the center-of-gravity determination unit 107 determines whether there is one each of the first center of gravity 116a and the second center of gravity 116b included in the distance range selected in step S203 (S204).

  When there is one each of the first centroid 116a and the second centroid 116b included in the distance range (Yes in S204), the centroid determination unit 107 indicates that only one target object (person) is included in the distance range. The second centroid 116b is determined as the position of the target object (S209).

  On the other hand, when there are a plurality of first centroids 116a and second centroids 116b included in the distance range (No in S204), the centroid determination unit 107 sets the second centroids as two groups (left group) based on the first centroids. And the right group) (S205). Then, the center-of-gravity determination unit 107 selects the second center of gravity at least one point from each of the first and second groups, and determines the selected second center of gravity as the position of the target object 150.

  13A and 13B are diagrams for explaining the processing in steps S205 to S208.

  In the example shown in FIG. 13A, the selected distance range includes one first centroid G (B01) and three second centroids g (b1), g (b2), and g (b3). Yes. In this case, the center-of-gravity determination unit 107 determines the second center of gravity g (b1), g (b2), and g (b3) based on a straight line that passes through the first center of gravity G (B01) and is parallel to the y-axis. Dividing into a left group located on the left side of the straight line and a right group located on the right side of the straight line. In this example, the second group includes the second centroid g (b1), and the right group includes the second centroids g (b2) and g (b3).

  Next, the center-of-gravity determination unit 107 selects a maximum of two points with high power from each group (S206). Here, since the right group includes one point, the left group includes two points, and the second centroid, the centroid determination unit 107 determines the second group centroid g (b1) of the left group and the second centroid g of the right group. Select (b2) and g (b3).

  Next, the center-of-gravity determination unit 107 selects one point from the left group and one point from the right group among the second centroids selected in step S206, and for each selected pair, two points included in the pair The center of gravity M (pair center of gravity) is calculated (S207). Here, as shown in FIG. 13B, the center of gravity M12 of g (b1) and g (b2) and the center of gravity M13 of g (b1) and g (b3) are calculated.

  Next, the center-of-gravity determination unit 107 selects a pair that minimizes the distance between the first center of gravity and the center of gravity M, and determines two second centers of gravity included in the selected pair as the positions of the target objects (S208). Here, as shown in FIG. 13B, since the centroid M closest to the first centroid G (B01) is the centroid M12, the centroid determination unit 107 includes the second centroids g (b1) and g included in the pair of centroids M12. Select (b2). That is, in this case, the center-of-gravity determination unit 107 determines that the target objects exist at the second center of gravity g (b1) and the second center of gravity g (b2), respectively.

  Here, when there are two target objects, the measurement points indicated by the micro Doppler output 115 include “metastable measurement points” corresponding to the bodies of the two target objects, and the two target objects. “Unstable measurement points” corresponding to arms and legs are mixed. The first centroid 116a is the centroid of the mixed measurement points. In other words, the first centroid 116a corresponds to the centroid of the metastable measurement point (false image). The second centroid 116b is a reflection point corresponding to a metastable measurement point among the mixed measurement points, that is, a candidate for the centroid of the body of the two target objects.

  In this example, three second centroids 116b are calculated. In addition, the process of selecting the pair closest to the first centroid described above selects the two reflection points corresponding to the metastable measurement points from the three second centroids 116b, that is, the centroids of the two target objects. It corresponds to the processing.

  If the processes in steps S203 to S209 have not been completed for all distance ranges (No in S210), another distance range is selected in step S203, and the processes in and after step S204 are performed for the selected distance range. Is done. In addition, when the processes in steps S203 to S209 are completed for all distance ranges (Yes in S210), the centroid determination unit 107 ends the centroid determination process. That is, in each distance range, the second center of gravity selected in step S208 or S209 is determined as the position of the target object.

  The range resolution in the spread spectrum radar is preferably set to the same level as the above Wh. For example, the range resolution is preferably 0.3 m to 0.6 m. Note that the range resolution can be changed by adjusting the code rate b of the pseudo-noise code.

  Further, when it is necessary to set the range resolution finely, for example, 0.1 m in order to improve the distance resolution, the center of gravity determination unit 107 sets the distance finely, and then the center-of-gravity determination unit 107 sets a plurality of distance ranges (eg, 3 ranges or 6 ranges). Range) may be integrated, and the processing of steps S203 to S209 may be performed for each integrated virtual range.

  In steps S205 to S208, the second centroid 116b may be corrected using the first centroid 116a instead of using the second centroid 116b itself. Specifically, the centroid determining unit 107 equally distributes the Doppler frequency and power of the first centroid 116a to the second centroid 116b included in the left group and the second centroid 116b included in the right group. Then, the center-of-gravity determination unit 107 may calculate the corrected frequency of the second center of gravity 116b using the distributed Doppler frequency and the original frequency of the second center of gravity 116b. By doing so, the Doppler frequency component (that is, the Doppler velocity component) due to the interference lost from the second centroid 116b when there is a false image can be compensated. As a result, information close to true can be reflected.

  In the present embodiment, the case of using a square lattice pattern has been described. However, various lattice patterns as shown in FIGS. 14A to 15C may be applied.

  FIG. 14A shows a two-dimensional rectangular grid, which is a special case.

  FIG. 14B shows a one-dimensional rectangular grid. This one-dimensional rectangular lattice is obtained by omitting the distance direction demarcation, specializing in that false images due to Doppler velocity interference occur due to direction estimation errors. In the spread spectrum radar, the range resolution can be used as a separation in the distance direction.

  In addition, the rectangular lattice shown in FIG. 14A and FIG. 14B has a uniform horizontal (x-direction) interval regardless of the distance, and therefore corresponds to the size (width) of the entity regardless of the distance of the target object. Unique delimiters are possible. Therefore, these rectangular grids have an advantage that they are easy to handle when a plurality of objects are separated using the size of the target object.

  On the other hand, FIGS. 14C and 14D show two-dimensional and one-dimensional fan lattices, respectively. Since the radar measurement system is a polar coordinate system, if the grid is a sector, the processing affinity is high.

  FIG. 15A shows a lattice pattern in which a one-dimensional rectangular lattice and a fan-shaped lattice in the distance direction are combined. This lattice has a structure that takes into account the object separation reflecting the size of the object by the rectangular lattice and the affinity of the measurement system (false image generation mechanism) by the fan-shaped lattice.

  Further, as shown in FIGS. 15B and 15C, a lattice interval in which the separation width is narrowed with respect to the distance or direction in which accuracy is desired to be increased may be set.

(Modification 1)
Further, the second divided frame may be determined according to the first center of gravity 116a. For example, as shown in FIG. 16A, one point of the first centroid 116a is determined for each distance range, and the intersection coordinates of the second divided frame A are set to coincide with the first centroid 116a. By doing so, it is possible to dynamically calculate the second centroid 116b and determine the centroid while taking into account the left-right positional relationship with respect to the first centroid 116a.

  Further, as shown in FIG. 16B, the one-dimensional rectangular lattice shown in FIG. 14B is combined with the distance range determined by the spread spectrum method, and the separation position of the rectangular lattice is set for each distance range. May be shifted left and right so as to match.

(Modification 2)
When the second centroid calculating unit 106b calculates the second centroid 116b, the number of measurement points in the divided area, that is, the number of measurement points contributing to the centroid may be recorded. Then, in the center-of-gravity determination process (FIG. 12), the center of gravity included in the divided region having the maximum number of measurement points contributing to the center of gravity from the second center of gravity 116b, or the center of gravity closest to the center of gravity may be selected. .

  In this way, it is possible to select a centroid that takes into account the concentration and spread of the measurement point group, that is, a centroid that reflects the geometric center of the object shape.

  12 is calculated by calculating both the center of gravity selected as the physical center shown in the flow of FIG. 12 and the center of gravity selected as the geometric center as in the second modification, and examining whether or not they match. It is also possible to determine the certainty of the center of gravity.

  As described above, the radar system 100 according to the present embodiment can display the center-of-gravity output 117 that is the output of the center-of-gravity determination unit 107 and the acquired image by the micro Doppler output 115 that is the imaging result on the display screen. it can. Thereby, either the shape image of the target object and the position of the target object can be monitored at the same time or as necessary.

  Thereby, even if the shape image of the target object is unclear, the radar system 100 can separately display at least macro information such as the number, position, and speed of the objects. This is important information when interpreting the shape.

  As described above, the radar system 100 according to Embodiment 1 of the present invention can estimate macro information for each of a plurality of objects without increasing the number of antennas even in a situation where a false image is generated due to interference of Doppler velocity. Thus, the radar system 100 can improve the detection accuracy of the position of the object.

(Embodiment 2)
FIG. 17 is a block diagram showing a configuration of the radar system 200 according to the second embodiment. A radar system 200 shown in FIG. 17 includes a tracking prediction unit 208 in addition to the configuration of the radar system 100 described in the first embodiment. Further, the center of gravity determination unit 207 is different in that the prediction output 218 that is the output of the tracking prediction unit 208 is used. The radar system 200 according to the second embodiment can also be applied when the number of target objects is three or more as in the first embodiment.

  In addition, regarding the common part with Embodiment 1, the same code | symbol as Embodiment 1 is attached | subjected and description is abbreviate | omitted.

<Tracking prediction unit 208>
The tracking prediction unit 208 receives the centroid output 217 output from the centroid determination unit 207. The center of gravity determining unit 207 will be described later, but the output of the center of gravity 217 is the same as that of the first embodiment.

  The tracking prediction unit 208 uses the past target object position determined by the center-of-gravity determination unit 207 to predict a predicted position that is the future target object position. Specifically, the tracking prediction unit 208 sequentially acquires the centroid output 217 in time series as measurement data. In addition, the tracking prediction unit 208 includes a prediction filter that can estimate (referred to as “prediction”) the barycentric position of the next step (future). Such a prediction filter includes a Kalman filter. The estimated value from the past to the present by such a prediction filter is used for tracking (tracking) of the target object because the measurement result is reflected each time and is determined gradually while being corrected.

  Further, the tracking prediction unit 208 predicts the number of target objects detected by the radar system 100. Specifically, the prediction filter may cause the target object to enter or exit from the non-detection area (area that cannot be detected by the system due to detection performance or physical occlusion) (that is, whether the number of objects increases or decreases), or multiple objects It is possible to predict whether there is an approach to the same distance range (that is, whether a false image is generated). Accordingly, the tracking prediction unit 208 corrects the number of objects using the prediction of the number variation, and determines the number of target objects in advance.

  Specifically, the tracking prediction unit 208 sets the number determination to 2 if the number of objects in the past is 2 and the prediction filter determines that a false image has occurred at the current or next step. As a result, even if three or more centroid candidates (second centroid 116b) are output due to the influence of the false image, it is clear that only two of them should be selected.

  As described above, the predicted output 218 includes the number N of target objects and information on the respective predicted values (Xpi, Ypi, Zpi) or (Rpi, Api, Epi) (where i = 1, 2,... N ) And false image determination information FI. For example, when N = 2, two sets of data (Xp1, Yp1, Zp1) and (Xp2, Yp2, Zp2) are output.

  Further, the false image determination information FI is, for example, when Rp1 = Rp2 and FI = 1 (false) when the range numbers of (Xp1, Yp1, Zp1) and (Xp2, Yp2, Zp2) are Rp1 and Rp2, respectively. When Rp1 ≠ Rp2, FI = 0 (no false image).

  Alternatively, the determination is further performed in multiple stages. When | Rp1-Rp2 | = 0, FI = 2 (fake image determination), and when | Rp1-Rp2 | = 1, FI = 1 (false image preliminary stage). Yes, otherwise it may be FI = 0 (no false image).

  Note that the predicted output 218 is preferably processed by being divided for each distance range. By doing so, the distance range in question can be distinguished and handled, so that it is possible to determine an appropriate center of gravity according to the presence or absence of a false image.

<Center of gravity determination unit 207>
The center of gravity determination unit 207 receives the first center of gravity 116a, the second center of gravity 116b, and the predicted output 218. Unlike the case of the first embodiment, the center-of-gravity determination unit 207 determines the center of gravity based on the number and positions of target objects indicated by the prediction output 218. Specifically, the center-of-gravity determination unit 207 determines the number of second centroids 116b predicted by the tracking prediction unit 208 among the plurality of second centroids 116b as the position of the target object.

  FIG. 18 is a flowchart of the center of gravity determination process. 19 and 20 are diagrams for explaining a method of selecting the center of gravity in the center of gravity determination unit 207.

  In the following description, it is assumed that only information of the same distance range is extracted and the information is rounded into a one-dimensional space. The horizontal axis is the A value (azimuth angle) or the X value (X position).

  First, it is assumed that the number of objects indicated by the predicted output 218 is N = 2. The method of selecting the center of gravity differs depending on the positional relationship between the predicted position indicated by the predicted output 218 and the first center of gravity 116a.

  First, the center-of-gravity determination unit 207 determines whether or not generation of a false image is indicated in the predicted output 218 (S301). Specifically, the center-of-gravity determination unit 207 determines that a false image is generated when F = 1 or 2 (Yes in S301), and determines that no false image is generated when F = 0. (No in S301).

  When a false image is generated (Yes in S301), the centroid determining unit 207 determines whether or not the first centroid 116a exists between the two predicted positions (S302). In other words, the centroid determining unit 207 determines whether or not predicted centroids exist on both the left and right sides of the first centroid 116a.

  As shown in FIG. 19, when the first centroid 116a (first centroid B) exists between two predicted positions (Yes in S302), the centroid determination unit 207 selects each predicted position from the second centroid 116b. The point closest to is selected (S303). Specifically, when the predicted position indicated by the predicted output 218 is N points, the centroid determining unit 207 determines the distance between each of the N points of the plurality of second centroids 116b and each of the N predicted positions. The sum of (or the square of the distance) is calculated. Then, the centroid determining unit 207 selects the second centroid 116b included in the combination that minimizes the calculated sum.

  In the example of FIG. 19, the center-of-gravity determination unit 207 calculates the following six distances d and selects a combination of minimum distances.

d12 = (Xg1-Xp1) ² + (Xg2-Xp2) ²
d13 = (Xg1-Xp1) ² + (Xg3-Xp2) ²
d21 = (Xg2-Xp1) ² + (Xg1-Xp2) ²
d23 = (Xg2-Xp1) ² + (Xg3-Xp2) ²
d31 = (Xg3-Xp1) ² + (Xg1-Xp2) ²
d32 = (Xg3-Xp1) ² + (Xg2-Xp2) ²

  For example, if d12 is the minimum distance, the centroid determining unit 207 selects the second centroids Xg1 and Xg2 as the centroid output 217.

  In addition, the tracking prediction unit 208 manages the target object by labeling (with an identification number) in order to perform tracking. For example, the predicted position with respect to the object 1 is Xp1, and the predicted position with respect to the object 2 is Xp2. Therefore, in the centroid output 217, the centroid Xg1 is associated with the object 1 and the centroid Xg2 is associated with the object 2.

  Although the description has been given here with the position dimension being reduced to one dimension, it can be easily expanded if evaluation in two dimensions and three dimensions is necessary.

  Further, by examining not only the position information but also the minimum distance including the Doppler frequency F (Fg, Fp), evaluation including the similarity of motion is possible, which is more preferable.

  For example, when i and j are indices and w is weighted (including unification of physical dimensions), the distance d may be calculated using the following (Equation 10).

d = Σi, j {(Xgi−Xpj) ² + w × (Fgi−Fpj) ² } (Equation 10)

  FIG. 20 is a diagram illustrating a case where the first centroid 116a is not present between the predicted positions. When the first centroid 116a does not exist between the two predicted positions (No in S302), the centroid determination unit 207 does not use the prediction output 218 for the determination, and determines the centroid according to the determination flow of the first embodiment shown in FIG. Select (S304).

  If the predicted output 218 indicates that no false image is generated (No in S301), the plurality of second centroids B calculated by the second centroid calculation unit 106b as they are as shown in FIG. The position is selected (S305). In this case, the first centroid calculating unit 106a may not calculate the first centroid 116a.

  Although the case where there are two objects has been described here, the same processing can be performed when three or more objects are predicted by the prediction output 218.

(Modification 1 of Embodiment 2)
The center-of-gravity determination unit 207 may perform the following process instead of the above-described process.

  FIG. 22 is a flowchart of the center-of-gravity determination process according to the first modification of the second embodiment. Further, in the predicted output 218, an outline when the number of objects is 3 or more (N ≧ 3) is shown in FIG. 23 by taking N = 3 as an example. When N ≧ 3, the center-of-gravity determination unit 207 determines the center of gravity according to the following flow. The following processing is performed for each distance range.

  First, the center-of-gravity determination unit 207 divides the second center of gravity into left and right groups (first group and second group) using the first center of gravity 116a as a reference (S321).

  The center-of-gravity determination unit 207 divides the prediction candidates into left and right groups (third group and fourth group) using the first center of gravity 116a as a reference (S322).

  Next, the center-of-gravity determination unit 207 determines whether the predicted position is included in both the left and right groups (S323).

  When the predicted position is included in both the left and right groups (Yes in S323), the centroid determining unit 207 selects the combination of the second centroids closest to each predicted position for each group as the position of the target object ( S324). That is, the center-of-gravity determination unit 207 includes, among a plurality of second centroids included in the left group, a second centroid closest to each predicted position included in the left group and a plurality of second centroids included in the right group, The second centroid closest to each predicted position included in the right group is determined as the position of the target object. Specifically, when the predicted position included in the left group is N points, the centroid determining unit 207 determines each of the N points of the plurality of second centroids 116b included in the left group and the predicted position of the N points. The sum of the distance (or the square of the distance) with each is calculated. Then, the center-of-gravity determination unit 207 selects the center of gravity with a combination that minimizes the calculated sum. The same applies to the processing for the right group.

  In the example illustrated in FIG. 23, the second centroids Xg1, Xg2, and Xg3 are included in the left group, and the second centroids Xg4 and Xg5 are included in the right group. Further, the predicted positions Xp1 and Xp2 are included in the left group, and the predicted position Xp3 is included in the right group. Therefore, the center-of-gravity determination unit 207 determines the distance (or the square of the distance) between each second center of gravity and each predicted position for six patterns that are combinations of three second center of gravity and two predicted positions included in the left group. The sum of is calculated. Here, among the six combinations, the sum of the combinations including the pair of Xg1 and Xp1 and the pair of Xg2 and Xp1 is the smallest. Therefore, the centroid determining unit 207 selects Xp1 as the second centroid corresponding to the predicted position Xp1, and selects Xp2 as the second centroid corresponding to the predicted position Xp2.

  In addition, the center-of-gravity determination unit 207 uses the second center of gravity Xg4 closest to the predicted position Xp3 among the second centers of gravity Xg4 and Xg5 included in the right group as the second center of gravity corresponding to the predicted position Xp3 included in the right group. select.

  On the other hand, when the predicted position is included only in one of the left and right groups (Yes in S323), the center-of-gravity determination unit 207 determines that one point from the group that does not include the predicted position and N− from the group that includes the predicted position. One second centroid is selected (S325). For example, when the predicted position is included only in the left group, the center-of-gravity determination unit 207 selects N−1 from among the N predicted positions included in the left group. For example, the center-of-gravity determination unit 207 selects N-1 predicted positions excluding the predicted position closest to the first center of gravity B among the N predicted positions included in the left group. Then, the center-of-gravity determination unit 207 selects the N−1 second centroids that minimize the sum of the distances for the selected predicted positions of the N−1 points. The center-of-gravity determination unit 207 selects the second center of gravity closest to the first center of gravity 116a among the second centers of gravity included in the right group.

  In this way, by using the first centroid 116a as a reference, the number of combinations of pairing between the predicted position and the second centroid 116b can be reduced. Furthermore, even if a false image is generated due to interference of Doppler velocity, it is possible to extract data that is as effective as possible. As described above, the radar system 200 can realize more stable tracking.

  As described above, in the radar system 200 according to Embodiment 2 of the present invention, even in a situation where there is a false image due to interference of Doppler velocity, macro information of each target object can be obtained from the measurement points without increasing the number of antennas. Can be estimated and separated. Thereby, the radar system 200 can realize stable tracking.

  Although the radar system according to the embodiment of the present invention has been described above, the present invention is not limited to this embodiment.

  Further, a part or all of the processing units included in the radar system according to the above embodiment may be realized as an LSI that is an integrated circuit. These may be individually made into one chip, or may be made into one chip so as to include a part or all of them.

  Further, the circuit integration is not limited to LSI, and may be realized by a dedicated circuit or a general-purpose processor. An FPGA (Field Programmable Gate Array) that can be programmed after manufacturing the LSI or a reconfigurable processor that can reconfigure the connection and setting of circuit cells inside the LSI may be used.

  Moreover, you may implement | achieve part or all of the function of a radar system based on embodiment of this invention, when processors, such as CPU, run a program.

  Furthermore, the present invention may be the above program or a non-transitory computer-readable recording medium on which the above program is recorded. Needless to say, the program can be distributed via a transmission medium such as the Internet.

  Moreover, you may combine at least one part among the functions of the radar system which concerns on the said Embodiment 1-2, and those modifications.

  Moreover, all the numbers used above are illustrated for specifically explaining the present invention, and the present invention is not limited to the illustrated numbers.

  In addition, division of functional blocks in the block diagram is an example, and a plurality of functional blocks can be realized as one functional block, a single functional block can be divided into a plurality of functions, or some functions can be transferred to other functional blocks. May be. In addition, functions of a plurality of functional blocks having similar functions may be processed in parallel or time-division by a single hardware or software.

  In addition, the order in which the above steps are executed is for illustration in order to specifically describe the present invention, and may be in an order other than the above. Also, some of the above steps may be executed simultaneously (in parallel) with other steps.

  Further, various modifications in which the present embodiment is modified within the scope conceivable by those skilled in the art are also included in the present invention without departing from the gist of the present invention.

  The present invention can be applied to a Doppler radar system. The present invention is useful for a crime prevention system or a collision avoidance system using such a Doppler radar system.

100, 200 Radar system 101a Transmitter 101b Receiver 102 Distance calculator 103 Doppler analyzer 104 Direction calculator 105 Micro Doppler calculator 106 Center of gravity calculator 106a First center of gravity calculator 106b Second center of gravity calculator 107, 207 Center of gravity determiner 120 Transmitting antenna 121, 121a, 121b, 121c Receiving antenna 111 Received signal 112 Range output 113 Doppler output 114 Direction output 115 Micro Doppler output 116a 1st center of gravity 116b 2nd center of gravity 117, 217 Center of gravity output 131, 132 Measurement points 150, 150a, 150b Target object 161a, 161b, 162a, 162b Reflection point 208 Tracking prediction unit 218 Prediction output 901 Radar device 910 Transmitter 911 Transmit antenna 912, 913 Receiving Antenna 920 and 930 receivers 931, 932, 933 DUT

Claims (11)

A Doppler radar system for detecting the position of a target object,
A transmission unit for generating a transmission signal;
A transmission antenna that radiates the transmission signal as a transmission wave;
A plurality of receiving antennas each receiving a reflected wave reflected by a plurality of reflecting points included in the target object;
A receiving unit that generates a plurality of received signals corresponding to each of a plurality of reflected waves received by the plurality of receiving antennas;
A direction calculation unit that calculates the direction of each of the plurality of reflection points with respect to the Doppler radar system using a phase difference between the plurality of reception signals;
A distance calculation unit that calculates each distance between the Doppler radar system and the plurality of reflection points using a delay amount between the transmission signal and the reception signal;
A micro-Doppler computing unit that calculates a plurality of measurement points indicating the positions of the plurality of reflection points using the direction and distance of each of the plurality of reflection points;
First centroid calculation for calculating a first centroid of measurement points included in each first divided region divided by the first divided frame among the plurality of measurement points using a first divided frame having a first interval. And
A second center of gravity of the measurement points included in each of the second divided regions divided by the second divided frame among the plurality of measurement points using a second divided frame having a second interval narrower than the first interval. A second centroid calculating unit for calculating
A Doppler radar comprising: a center-of-gravity determining unit that determines at least one second center of gravity as the position of the target object based on a positional relationship between the second center of gravity and the first center of gravity among the plurality of second centers of gravity; system. The second interval is approximately twice the width of the target object;
The Doppler radar system according to claim 1, wherein the first interval is at least twice as long as the second interval. The first divided frame includes a third divided frame and a fourth divided frame different from the third divided frame,
The second divided frame includes a fifth divided frame and a sixth divided frame different from the fifth divided frame,
The first center-of-gravity calculation unit
Of the plurality of measurement points, calculate a third center of gravity of the measurement points included in each third divided region divided by the third division frame,
Of the plurality of third centroids, calculate the centroid of the third centroid included in each fourth divided region divided by the fourth division frame as the first centroid,
The second centroid calculating unit
Of the plurality of measurement points, calculate a fifth center of gravity of the measurement points included in each fifth divided region divided by the fifth division frame,
3. The Doppler radar system according to claim 1, wherein among the plurality of fifth centroids, a centroid of a fifth centroid included in each sixth divided region divided by the sixth division frame is calculated as the second centroid. The center-of-gravity determination unit
Dividing the plurality of second centroids into a first group and a second group based on the position of the first centroid;
Selecting a second centroid from each of the first and second groups at least one point;
The Doppler radar system according to claim 1, wherein the selected second center of gravity is determined as the position of the target object. The center-of-gravity determination unit
For each pair composed of the second centroid included in the first group and the second centroid included in the second group, a pair centroid that is the centroid of the two second centroids included in the pair is calculated. ,
Of the plurality of calculated pair centroids, select the pair of the pair centroids closest to the first centroid,
The Doppler radar system according to claim 4, wherein two second centroids included in the selected pair are determined as positions of the two target objects. The Doppler radar system further includes:
A tracking prediction unit that predicts a predicted position that is the position of the target object in the future using the position of the target object in the past determined by the center of gravity determination unit;
The tracking prediction unit predicts the number of target objects detected by the Doppler radar system,
The center of gravity determination unit determines the number of second centroids predicted by the tracking prediction unit among the plurality of second centroids as the position of the target object. Doppler radar system. The center-of-gravity determination unit
When the first centroid exists between the plurality of predicted positions predicted by the tracking prediction unit, the second centroid closest to each predicted position among the plurality of second centroids is determined as the position of the target object. Decided on
When the first centroid does not exist between the plurality of predicted positions predicted by the tracking prediction unit, based on the positional relationship between the second centroid and the first centroid from among the plurality of second centroids. The Doppler radar system according to claim 6, wherein at least one second center of gravity is determined as the position of the target object. The center-of-gravity determination unit
Dividing the plurality of second centroids into a first group and a second group based on the position of the first centroid;
Dividing the plurality of predicted positions into a third group and a fourth group based on the position of the first center of gravity;
Of the plurality of second centroids included in the first group, the second centroid closest to each predicted position included in the third group, and the second centroid included in the second group. The Doppler radar system according to claim 6, wherein a second center of gravity closest to each predicted position included in four groups is determined as the position of the target object. The center-of-gravity determination unit
When the predicted position is included in both the third group and the fourth group, the second centroid closest to each predicted position included in the third group among the plurality of second centroids included in the first group. And the second centroid closest to each predicted position included in the fourth group among the plurality of second centroids included in the second group is determined as the position of the target object,
When the third group includes N (N is an integer of 3 or more) predicted positions, and the fourth group does not include a predicted position, the second of N−1 points included in the first group. The Doppler radar system according to claim 8, wherein a center of gravity and a second center of gravity of one point included in the second group are determined as the positions of the eye objects. The center-of-gravity determination unit
When a plurality of predicted positions are included in the same distance range, at least one second centroid is determined from the plurality of second centroids based on a positional relationship between the second centroid and the first centroid. Determine the position of the object,
The Doppler radar system according to claim 6, wherein when only one predicted position is included in the same distance range, the second centroid is determined as the position of the target object. An object detection method for detecting the position of a target object using a Doppler radar system,
Generating a transmission signal; and
Radiating the transmission signal as a transmission wave by a transmission antenna;
Receiving the reflected waves reflected by a plurality of reflection points included in the target object at each of a plurality of receiving antennas;
Generating a plurality of received signals corresponding to each of a plurality of reflected waves received by the plurality of receiving antennas;
Calculating a direction of each of the plurality of reflection points with respect to the Doppler radar system using a phase difference between the plurality of received signals;
Calculating each distance between the Doppler radar system and the plurality of reflection points using a delay amount between the transmission signal and the reception signal;
Using a direction and a distance of each of the plurality of reflection points to calculate a plurality of measurement points indicating positions of the plurality of reflection points;
Calculating a first centroid of measurement points included in each first divided region divided by the first divided frame out of the plurality of measurement points using a first divided frame having a first interval;
A second center of gravity of the measurement points included in each of the second divided regions divided by the second divided frame among the plurality of measurement points using a second divided frame having a second interval narrower than the first interval. Calculating steps,
Determining at least one second centroid as the position of the target object based on a positional relationship between the second centroid and the first centroid from among the plurality of second centroids.