JP2019027973A - Target determination device - Google Patents

Abstract

To improve detection accuracy of guard rails.SOLUTION: A target determination device to be loaded to a vehicle is configured to acquire reflection point data on a reflection point group including the reflection point reflecting a laser wave applied within a detection range in S110; cluster the reflection point consisting of the reflection group in S120; and output a cluster satisfying a guard rail condition of clusters generated by clustering as a guard rail target in S130. Herein, in the guard rail condition, are included, for example, whether a plurality of pieces of cluster exists; whether a size of the cluster falls within a range of predetermined size, whether a position of the cluster falls within a range of a predetermined position, and whether an approximate curve expressing a distribution of the cluster matches a predetermined shape.SELECTED DRAWING: Figure 2

Description

  The present disclosure relates to a target determination device.

  A radar device used for target detection is to radiate a radar wave such as a laser wave or a millimeter wave as a transmission wave over a predetermined angular range around the vehicle and receive a reflected wave with respect to the transmission wave. Thus, a stationary obstacle that is an obstacle to the traveling of the host vehicle existing on the road on which the vehicle is traveling is detected.

  In this type of system, it is necessary to distinguish and recognize whether the detected target is a guardrail or a stationary obstacle present on the road. For example, when the road shape of the road on which the vehicle is traveling is curved, if a guardrail existing in front of the vehicle along the curved shape is detected as a stationary obstacle, the stationary obstacle on the road Therefore, an operation corresponding to the erroneous detection result is performed.

  On the other hand, for example, Patent Document 1 discloses a technique of recognizing a guard rail by detecting a lateral plate constituting the guard rail.

Japanese Patent Laid-Open No. 9-222477

  However, depending on the positional relationship between the guard rail horizontal plate and the vehicle, such as when the beam is present in parallel with the beam direction of the radar wave emitted by the radar apparatus, reflection in the direction in which the vehicle exists may be extremely small. In such a case, the vehicle cannot receive the reflected wave, and as a result, the vehicle may not be able to recognize the guardrail.

  An object of this indication is to improve the detection accuracy of a guardrail.

  One aspect of the present disclosure is a target determination device mounted on a vehicle, which includes a reflection point acquisition unit (S110), a clustering unit (S120), a size calculation unit (S330), and a column candidate extraction unit (S240). ) And a guardrail determination unit (S130). The reflection point acquisition unit detects one or more reflection points reflecting the radar wave as a reflection point group by irradiating the radar wave within a predetermined detection range, and each of the detected reflection point groups is detected. The reflection point data including the position of the reflection point with respect to the vehicle is acquired. The clustering unit generates a cluster group having one or more clusters by clustering the reflection points included in the reflection point group based on the reflection point data acquired by the reflection point acquisition unit. The size calculation unit calculates a cluster size that is a spatial size of the cluster for each of the clusters included in the cluster group generated by the clustering. The column candidate extraction unit sets at least one extraction condition in advance, and determines that the cluster size calculated by the size calculation unit is within a column size range that is a spatial size corresponding to the column of the guardrail. And a cluster satisfying the extraction condition is extracted as a column candidate. The guardrail determination unit sets at least one guardrail condition in advance, and determines that a cluster extracted as a column candidate is a cluster constituting a target representing the guardrail when the column candidate satisfies the guardrail condition.

According to such a configuration, the detection accuracy of the guardrail can be improved.
In other words, since the shape of the guard rail column is cylindrical, unlike the reflection direction of the radar wave reflected by the guard rail horizontal plate, the reflection direction of the radar wave reflected by the side surface of the columnar column is not unidirectional. Easy to reach the vehicle. Since the position of the guardrail is estimated from the radar wave that has reached the vehicle, the detection accuracy of the guardrail can be improved.

  Note that the reference numerals in parentheses described in this column and in the claims indicate the correspondence with the specific means described in the embodiment described later as one aspect, and the technical scope of the present disclosure It is not limited.

It is a block diagram which shows the structure of a target determination system. It is a flowchart showing the target determination process. It is a flowchart showing guardrail determination processing. It is a flowchart showing pillar candidate extraction processing. It is a bird's-eye view showing the road where a self-vehicle runs. It is a figure showing a reflective point group. It is a figure showing a cluster group. It is a figure showing an estimation track and a guardrail range. It is a figure showing a pillar candidate. It is a figure showing the approximated curve showing distribution of a column candidate, and an estimation track. It is a figure showing the guardrail range in a modification. It is a figure showing the guardrail range in a modification.

Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.
[1. Constitution]
As shown in FIG. 1, a target determination system 1 mounted on a vehicle includes a radar sensor 10, a driving information sensor 20, a target determination device 30, and a processing device 40.

  The radar sensor 10 is a well-known sensor that detects a reflection point, which is a point that reflects a radar wave, by radiating a radar wave within a predetermined detection range and receiving the reflected wave. The radar sensor 10 is installed in the own vehicle so that the beam direction coincides with the front direction of the own vehicle, for example. The beam direction here refers to the center direction of the detection range, which is the range in which the radar sensor 10 emits radar waves. In addition, it is assumed that the reflection point data includes at least the azimuth with respect to the front direction of the own vehicle and the distance to the own vehicle. The azimuth with respect to the own vehicle refers to the angle in the horizontal direction where there is a reflection point obtained with reference to the beam direction. That is, the reflection point data includes position data indicating the position of the reflection point with respect to the host vehicle.

  The reflection point data is generated by performing known signal processing such as fast Fourier transform processing on the received reflected wave. The radar wave here may be a so-called millimeter wave radar using an electromagnetic wave in the millimeter wave band, a laser radar using a laser beam as a radar wave, or a sonar using a sound wave as a radar wave. In any case, the antenna unit that transmits and receives the radar wave is configured to detect the arrival direction of the reflected wave with respect to the horizontal plane. Note that the antenna unit may be an array antenna including a transmission element and a reception element, for example.

  The driving information sensor 20 is a sensor for detecting driving information that is information related to the behavior of the own vehicle and the driving operation that affects the behavior of the own vehicle. The detection information of the driving information sensor 20 includes, for example, an operation amount of an accelerator pedal and a brake pedal, a steering angle of a steering wheel, a vehicle speed, a vehicle acceleration, a yaw rate that is a turning angular velocity, a tire rotation number, and the like. That is, the driving information sensor 20 includes various sensors such as a vehicle speed sensor, a yaw rate sensor, and a steering angle sensor.

  The processing device 40 is a device that performs a predetermined process such as notifying that there is a stationary obstacle that may cause a collision of the vehicle according to the result output from the target determination device 30. .

  The target determination apparatus 30 is mainly configured by a known microcomputer having a CPU 31 and a semiconductor memory (hereinafter, memory 32) such as a RAM, a ROM, and a flash memory. Various functions of the target determination device 30 are realized by the CPU 31 executing a program stored in a non-transitional tangible recording medium. In this example, the memory 32 corresponds to a non-transitional tangible recording medium that stores a program. Also, by executing this program, a method corresponding to the program is executed. The number of microcomputers constituting the target determination device 30 may be one or more.

  The target determination device 30 is realized by the CPU 31 executing a program, but this method is not limited to software, and some or all of the elements are realized using one or a plurality of hardware. May be. For example, when the above function is realized by an electronic circuit that is hardware, the electronic circuit may be realized by a digital circuit including a large number of logic circuits, an analog circuit, or a combination thereof.

[2. processing]
<Target determination processing>
Next, the target determination process executed by the target determination device 30 will be described with reference to the flowchart of FIG. The target determination process is periodically performed while, for example, the ignition of the own vehicle is on.

  In S110, the target determination device 30 performs a reflection point acquisition process that is a process of acquiring the reflection point data of the reflection points included in the reflection point group for each acquisition cycle. The reflection point group here refers to a reflection point acquired from the radar sensor 10. The acquisition cycle refers to an arbitrary period set in advance for acquiring the reflection point.

  In S120, the target determination device 30 performs a clustering process that is a process of generating a cluster having one or more reflection points by dividing the reflection point group based on the position of the reflection point represented by the reflection point data. . Here, as the reflection point data used in the clustering process, the reflection point data (hereinafter, current data) acquired in S110 in the current target determination process is used. However, the reflection point data used in the clustering process uses, in addition to the current data, the reflection point data (hereinafter referred to as past data) stored in S140, which will be described later, in the target determination process of the preset number of times before the previous time. May be. When past data is used, the position of the reflection point represented by the past data may be data corrected based on the movement of the vehicle from the time when the past data is acquired until the current data is acquired. Further, the preset number of times before the previous time may be, for example, five times.

  Here, the correction of the past data is performed by moving the position of the reflection point represented by the past data by the movement amount of the own vehicle from the time when the past data is acquired to the time when the current data is acquired. For example, the movement amount of the host vehicle may be calculated based on the driving information acquired by the driving information sensor 20.

  As a specific method of clustering, for example, the shortest distance method, the group average method, the Ward method, or the longest distance method is used. Here, when the shortest distance method is used, for example, clustering may be performed with other reflection points that are closest to a certain reflection point and within a range of 2 m as the same cluster. Note that one or more clusters obtained as a result of the clustering process performed on the reflection point group are hereinafter referred to as a cluster group.

  In S130, the target determination device 30 performs guardrail determination processing. The guardrail processing here refers to whether a cluster included in the cluster group represents a guardrail target or a non-guardrail target by determining whether or not the cluster generated in S120 satisfies the guardrail condition. This is a process for determining whether or not it is a thing. The guardrail target here refers to a target representing the guardrail, and the non-guardrail target refers to a target other than the guardrail target. Details of the guardrail determination process will be described later.

In S140, the target determination apparatus 30 performs a data storage process that is a process of storing the reflection point data acquired in S110 in the memory 32 for each acquisition cycle.
The process at S110 corresponds to a reflection point acquisition unit, the process at S120 corresponds to a clustering unit, the process at S130 corresponds to a guardrail determination unit, and the process at S140 corresponds to a storage unit.
<Guardrail judgment processing>
The target determination device 30 will be described in detail with reference to FIG.

In S210, the target determination device 30 acquires the cluster group generated in S120.
In S220, the target determination device 30 determines whether the number of clusters constituting the cluster group acquired in S210 satisfies the number condition. Here, the number condition is a condition that defines whether or not the number of clusters necessary for deriving an approximate curve representing a guardrail exists in the cluster group in S250 described later. Here, the number required to configure the guard rail is, for example, two or more.

When the target determination device 30 determines that the cluster group does not satisfy the number condition, the processing proceeds to S230.
In S230, the target determination device 30 determines that the cluster included in the cluster group represents a non-guardrail target, and outputs a determination result. That is, the cluster included in the cluster group does not include the guardrail target, and the determination result that all the clusters represent the non-guardrail target is output to the processing device 40, and the guardrail determination process ends.

On the other hand, if the target determination device 30 determines that the cluster group satisfies the number condition, the process proceeds to S240.
In S240, the target determination device 30 performs a column candidate extraction process, which is a process of extracting a cluster that is estimated to be a guard rail column from the clusters included in the cluster group as a column candidate. Details of the column candidate extraction process will be described later.

  In S250, the target determination device 30 derives an approximate curve representing the distribution of the column candidates extracted in S240. Here, the approximate curve may be derived by fitting an arc shape or a clothoid curve shape to the distribution of the column candidates. In addition, the fitting may be performed by a well-known fitting method such as a least square method or a Levenberg Marquett method. At this time, for example, the error between the arc shape or clothoid curve shape used for fitting and the position of each column candidate may be calculated, and the one with the smallest calculated error may be derived as the approximate curve.

  In S260, the target determination device 30 determines whether or not the approximate curve derived in S250 meets the curve condition. The curve condition referred to here is a condition for determining whether or not the vehicle shape matches the road shape on which the vehicle can travel. Here, the road shape on which the vehicle can travel is a general public road shape. That is, the road shape on which the host vehicle can travel is an arc shape or a clothoid curve, and its curvature is a size within a range defined by law or the like. Further, the curve condition may be such that, for example, when the fitting is performed in S250, the calculated error is within a predetermined range.

If the target determination device 30 determines in S260 that the approximate curve derived in S250 does not match the curve condition, the process proceeds to S230.
On the other hand, if the target determination device 30 determines in S260 that the approximate curve derived in S250 matches the curve condition, the process proceeds to S270.

  In S270, the target determination device 30 assumes that the clusters extracted as the column candidates in S240 among the clusters included in the cluster group obtained in S210 represent guardrail targets, and other clusters are non-guardrail objects. Reflection point data is output to the processing device 40 as representing the target, and the guardrail determination processing is terminated.

Note that the processing in S240 corresponds to a column candidate extraction unit.
<Pole candidate extraction process>
The target determination apparatus 30 will be described in detail with reference to FIG.

  In S310, the target determination device 30 performs a trajectory estimation process for estimating the estimated trajectory of the host vehicle. The estimated trajectory is a virtual curve representing a trajectory estimated to travel. The estimated trajectory is estimated based on driving information such as a vehicle speed, a yaw rate, and a steering wheel steering angle acquired from the driving information sensor 20, for example.

In S320, the target determination device 30 sets a guardrail range.
Here, the guardrail range refers to a range that is widened by a distance corresponding to the center to the end of the lane in the vehicle width direction from the estimated track estimated in S310, for example.

  In S330, the target determination device 30 performs a size calculation process that is a process of calculating a cluster size that is a spatial size of each cluster included in the cluster group. Here, for example, in each cluster, the cluster size may be a distance from the center of gravity representing the center of the distribution of reflection points in the cluster to the farthest reflection point among the reflection points included in the cluster.

  In S340, the target determination device 30 selects one cluster among the clusters included in the cluster group as the selected cluster. The selected cluster selected here is selected from unselected clusters that have not been subjected to the processing of S350 to S380 described later.

  In S350, the target determination device 30 determines whether or not the cluster size of the selected cluster satisfies the size condition. The size condition here means that the cluster size is within a preset support size range. The column size range is a range set based on the column size. The column size represents the spatial size of the column of the guardrail, and specifically refers to a size corresponding to the outer diameter of a general column of the guardrail.

The selected cluster that is determined not to satisfy the size condition in S350 is set as a non-column candidate in S360, and the target determination device 30 moves the process to S390.
On the other hand, the selected cluster determined to satisfy the size condition in S350 is determined whether or not the range condition is satisfied in S370. Here, the range condition is that the position of the selected cluster is included in the guardrail range set in S320.

The selected cluster that is determined not to satisfy the range condition in S370 is set as a non-column candidate in S360, and the target determination device 30 moves the process to S390.
The selected cluster determined to satisfy the range condition in S370 is set as a column candidate in S380, and the target determination device 30 moves the process to S390.

In S390, the target determination device 30 determines whether an unselected cluster exists in the cluster group.
If an unselected cluster exists in the cluster group, the process returns to S340 and the subsequent processing is performed.

If there is no unselected cluster in the cluster group, the column candidate extraction process is terminated.
That is, by the column candidate extraction process, a cluster that satisfies the size condition and the range condition among a plurality of clusters included in the cluster group is extracted as a column candidate.

The process at S310 corresponds to the trajectory estimation unit, and the process at S330 corresponds to the size calculation unit.
That is, when the number of clusters included in the cluster group satisfies the number condition, each cluster satisfies the size condition and the range condition, and the cluster position distribution satisfies the curve condition, it is output as a guardrail target. .

<Example of results of each process>
As an example of the result of each process, the result of each process in the case where the guard rail column P, the other vehicle Q, and the other target T are present is shown.

  Here, as an example, on a curved road as shown in FIG. 5, there are a plurality of guardrail posts P and other targets T on the front left side of the vehicle, and the road center line L is sandwiched on the front right side of the vehicle. Consider the case where there is another vehicle Q that is an oncoming vehicle.

  In the situation shown in FIG. 5, the reflection point group as shown in FIG. 6 is obtained by performing the reflection point acquisition process of S110. The reflection point group shown in FIG. 6 includes a plurality of reflection points detected corresponding to the positions of the guardrail column P, the other vehicle Q, and the other target T, respectively. Note that the reflection point group shown in FIG. 6 may further include past data stored in S140.

  Clustering processing is performed in S120 on the reflection point group as shown in FIG. 6 to obtain a cluster group including the cluster as shown in FIG. Here, the cluster group includes a column cluster Cp representing the guard rail column P, another vehicle cluster Cq representing the other vehicle Q, and another target cluster Ct representing the other target T. The position and size of each cluster correspond to the position and size of each target.

  When the trajectory estimation process is performed in S310, an estimated trajectory E representing the trajectory that the host vehicle is estimated to travel as shown in FIG. 8 is set as the traveling direction of the host vehicle. Further, when the guardrail range is set in S320, a range in which the guardrail range Z1 is widened from the estimated track in the vehicle width direction by a distance corresponding to the center to the end as shown in FIG.

  Here, when the column candidate extraction processing from S340 to S390 is performed for each cluster included in the cluster group in FIG. 8, as shown in FIG. 9, the column candidate Sp corresponding to the column cluster Cp is extracted. The That is, for example, the other vehicle cluster Cq is not extracted as a column candidate because the cluster size does not satisfy the size condition in S350. Further, the other target cluster Ct is not extracted as a column candidate because the position of the cluster does not satisfy the range condition in S370.

Furthermore, as shown in FIG. 10, an approximate curve G representing the distribution of the extracted column candidates Sp is derived in S250.
If it is determined in S260 that the shape of the derived approximate curve G matches the curve condition, the cluster representing the column candidate Sp is used as a guardrail target, and the clusters that are not extracted as other column candidates are used as non-guardrail targets. , S270.

  On the other hand, if it is determined in S260 that the shape of the approximate curve does not conform to the curve condition, all detected clusters are output to the processing device 40 in S230 as representing non-guardrail targets.

[3. effect]
According to the first embodiment described in detail above, the following effects are obtained.
(1) According to the above-described embodiment, since the guard rail support column is detected and it is determined whether the guard rail is based on the detected support column, it is possible to improve the detection accuracy for detecting the guard rail. In other words, since the shape of the guard rail column is cylindrical, unlike the reflection direction of the radar wave reflected by the guard rail horizontal plate, the reflection direction of the radar wave reflected by the side surface of the columnar column is not unidirectional. Easy to reach the vehicle. Since the position of the guardrail is estimated from the radar wave that has reached the vehicle, the detection accuracy of the guardrail can be improved.

  As a result, it is possible to improve the detection accuracy of stationary obstacles. That is, out of the targets detected by the radar device, non-guardrail targets existing on the road on which the vehicle travels are more accurately extracted as stationary obstacles as the detection accuracy of the guardrail targets improves. Can do.

[4. Other Embodiments]
As mentioned above, although embodiment of this indication was described, this indication is not limited to the above-mentioned embodiment, and can carry out various modifications.

  (1) In the above embodiment, the guardrail range is a range that is widened from the estimated track in the vehicle width direction by a distance corresponding to the center to the end of the lane. However, the guardrail range is limited to a range that is widened around the estimated track. It is not something. For example, as shown in FIG. 11, the guard rail range may be set to a range Z2 centered on the roadway center line L existing in the center of the roadway on which the host vehicle travels. In such a case, the position of the roadway center line L of the road on which the vehicle travels is assumed to be, for example, that the estimated path passes through the center of the lane divided by the centerline, and the vehicle width direction from the estimated path May be set to exist at a position translated by half the width of the lane.

Moreover, it is not limited to a thing centering on a center line, You may extend only the distance from the center of the road where the own vehicle drive | works to the road edge only to the road side where the own vehicle drive | works from the center line.
(2) Further, for example, as shown in FIG. 12, the guardrail range is widened by a predetermined distance in the vehicle width direction from a position that is separated from the estimated track E by the lateral width of the lane in the vehicle width direction. A range Z3 may be set.

(3) In the guardrail determination processing in the above embodiment, the arc shape or the clothoid curve shape is used for fitting, but the shape of the estimated trajectory may be used.
(4) A plurality of functions of one constituent element in the above embodiment may be realized by a plurality of constituent elements, or a single function of one constituent element may be realized by a plurality of constituent elements. . Further, a plurality of functions possessed by a plurality of constituent elements may be realized by one constituent element, or one function realized by a plurality of constituent elements may be realized by one constituent element. Moreover, you may abbreviate | omit a part of structure of the said embodiment. In addition, at least a part of the configuration of the above embodiment may be added to or replaced with the configuration of the other embodiment. In addition, all the aspects included in the technical idea specified from the wording described in the claims are embodiments of the present disclosure.

  (5) In addition to the target determination device 30 described above, a system including the target determination device 30 as a constituent element, a program for causing a computer to function as the target determination device 30, a semiconductor memory storing the program, and the like The present disclosure can also be realized in various forms such as a non-transition actual recording medium and a target determination method.

DESCRIPTION OF SYMBOLS 1 ... Target determination system, 10 ... Radar sensor, 20 ... Driving information sensor, 30 ... Target determination device, 31 ... CPU, 32 ... Memory, 40 ... Processing device, Cp ... Column cluster, Cq ... Other vehicle cluster, Ct ... Other target cluster, E ... Estimated trajectory, G ... Approximate curve, L ... Roadway center line, P ... Stand, Q ... Other car, Sp ... Pole candidate, T ... Other targets, Z1, Z2, Z3 ... Guardrail range.

Claims (8)

A target determination device (30) mounted on a vehicle,
By irradiating a radar wave within a predetermined detection range, one or more reflection points reflecting the radar wave are detected as a reflection point group, and each of the detected reflection point groups is detected with respect to the vehicle. A reflection point acquisition unit (S110) configured to acquire reflection point data including the position of the reflection point;
Clustering configured to generate a cluster group having one or more clusters by clustering the reflection points included in the reflection point group based on the reflection point data acquired by the reflection point acquisition unit. Part (S120),
A size calculation unit (S330) configured to calculate a cluster size that is a spatial size of the cluster for each of the clusters included in the cluster group generated by the clustering;
The cluster size calculated by the size calculation unit is set as one of at least one extraction condition that the cluster size is in a support column size range corresponding to a guard rail support column, and the extraction is performed. A column candidate extraction unit (S240) configured to extract the cluster satisfying the condition as a column candidate;
When at least one guardrail condition is set in advance and the column candidate satisfies the guardrail condition, the cluster extracted as the column candidate is determined to be the cluster constituting the target representing the guardrail. A guardrail determination unit configured in (S130),
A target judgment device. The target determination apparatus according to claim 1,
The guardrail determination unit determines in advance a difference between the shape of the approximate curve (G) calculated from the distribution of positions where the plurality of column candidates existing within the guardrail range and the shape of the arc or clothoid curve. A target determination device that uses the guardrail condition as being within a certain range. The target judgment device according to claim 1 or 2,
A trajectory estimation unit (S310) configured to estimate an estimated trajectory (E) that is a trajectory estimated to travel the vehicle;
The guardrail determination unit is configured so that a difference between the shape of the approximate curve calculated from the distribution of positions where the plurality of column candidates existing in the guardrail range and the shape of the estimated trajectory are determined in advance is determined. A target determination apparatus that uses the guardrail condition as being present. The target determination apparatus according to claim 1,
A trajectory estimation unit (S310) configured to estimate an estimated trajectory that is a trajectory that the vehicle is estimated to travel;
The column candidate extraction unit is set so as to include, as one of the extraction conditions, being within a guardrail range that is a range set based on the estimated trajectory, and the clusters satisfying the extraction condition are set as the column A target determination device configured to extract as a candidate. The target judgment device according to claim 3 or claim 4,
The target rail judging device, wherein the guard rail range is set to a range widened in the vehicle width direction of the vehicle so as to include a distance from the center of the lane in which the vehicle is traveling to the road edge from the estimated track. The target judgment device according to claim 3 or claim 4,
The target rail determination device, in which the guardrail range is set to a range that is widened by a predetermined range around a position that is a distance from the center of the road on which the vehicle is traveling to the end of the road. The target determination apparatus according to any one of claims 3 to 6, wherein
The trajectory estimator is
The turning angular velocity detected by the driving information sensor (20) for obtaining driving information including at least one of the behavior of the vehicle mounted on the vehicle and the driving operation affecting the behavior of the vehicle, and steering of the steering wheel A target determination device configured to estimate the estimated trajectory based on at least one of the corners and a vehicle speed. The target determination apparatus according to any one of claims 1 to 7, wherein
The reflection point acquisition unit is configured to acquire the reflection point data for each acquisition cycle,
A storage unit (S140) configured to store the reflection point data;
The clustering unit
Based on the reflection point data of the cluster group acquired in the current acquisition cycle and the reflection point data of the cluster group acquired in the previously set number of acquisition cycles before the previous time stored in the storage unit A target determination apparatus configured to perform the clustering.

Images