JP5163087B2 - Object detection device - Google Patents

Description

  The present invention relates to an object detection apparatus that detects an object using a detection point by a radar.

In recent years, driving support devices such as collision prevention devices and inter-vehicle control devices have been developed. In these driving assistance devices, it is important to detect a vehicle traveling in front of the host vehicle. The object detection apparatus groups detection points that satisfy a predetermined condition among a large number of detection points detected by a radar such as a laser radar, and obtains object information from the group of detection points of each group (Patent Document) 1).
JP 2003-57339 A

  Detected when two or more objects in the detection area of the laser radar approach (for example, two pedestrians who are far apart approach each other and overlap when viewed from the laser radar side) When the grouping is performed only with the detected points, the detection point groups for the plurality of objects may be grouped into one group. In such a case, a plurality of objects are detected as one object.

  Therefore, an object of the present invention is to provide an object detection apparatus that detects an object with high accuracy in object detection using a radar.

The object detection apparatus according to the present invention includes a radar detection means for detecting the position of an object, and the two detection points when the distance between the two detection points detected this time by the radar detection means is less than a threshold value. Grouping means for grouping each detection point detected by the radar detection means by making the same group, and dividing means for dividing the group currently grouped by the grouping means into a plurality of groups based on a plurality of previous groups If, for each final divided was the group in the grouping means or dividing means, and object detecting means for detecting an object by using the detection points included in the group, of the previous groups based on the information of the previous group Position estimation means for estimating the current position, the position estimation means for each previous group detected by the object detection means as an object. When the object detecting unit previously detected two or more objects, the dividing unit determines whether the estimated position estimated by the position estimating unit for each detection point included in the currently grouped group and the radar detecting unit The distance from the detected detection point is calculated, and when the distance is less than a preset threshold, the detection point is divided according to the closest estimated position, and the group is divided.

  In this object detection device, an electromagnetic wave is transmitted by the radar detection means and the reflected electromagnetic wave is received, and a detection point for the object is acquired based on transmission / reception information of the electromagnetic wave. In the object detection device, the detection points acquired this time by the grouping means are grouped according to a predetermined condition. At this time, since two or more objects approach each other, the detection point group for each object may be grouped as one group. However, since the two or more objects were separated before approaching, the detection point group for each object should be grouped as a separate group. Therefore, using the previous grouping result, the object detection apparatus divides the group grouped this time into a plurality of groups based on a plurality of previous groups (corresponding to a plurality of objects detected last time) by the dividing unit. And in an object detection apparatus, an object is detected using the detection point of each group grouped by the object detection means. As described above, in this object detection apparatus, the group can be subdivided in consideration of the previous grouping result (corresponding to the previous object detection result), so that the object can be detected with high accuracy. As the previous grouping result (previous group), not only the grouping result at the latest previous time but also the past grouping result before the previous time such as the grouping result at the previous time may be used.

  In this object detection apparatus, the position estimation unit estimates the position where the previous group has moved up to the current detection based on the information (position, speed, movement direction, etc.) of the previous group. In the object detection device, whether or not it is a detection point for the previously detected object based on the positional relationship between the estimated position of the previous group and the position of each detection point of the group grouped this time by the dividing unit. The group grouped this time is divided into a plurality of groups. As described above, in this object detection apparatus, it is possible to perform subdivision with high accuracy by subdividing a group based on the current estimated position of the previous group.

  The present invention can detect an object with high accuracy by re-dividing a group in consideration of the previous result.

  Hereinafter, an embodiment of an object detection device according to the present invention will be described with reference to the drawings.

  In the present embodiment, the object detection device according to the present invention is applied to an obstacle detection device mounted on a vehicle. The obstacle detection device according to the present embodiment detects obstacles around the vehicle (for example, other vehicles, bicycles, and pedestrians), and uses the detected obstacle information as a driving support device such as an inter-vehicle control device or a collision prevention device. To provide. The detection direction may be the entire circumferential direction of the vehicle, or only the direction according to the purpose such as front, side, and rear.

  With reference to FIGS. 1-3, the obstruction detection apparatus 1 which concerns on this Embodiment is demonstrated. FIG. 1 is a configuration diagram of an obstacle detection apparatus according to the present embodiment. FIG. 2 is an explanatory diagram of a group subdivision technique in the ECU of FIG. FIG. 3 is an example of group subdivision.

  The obstacle detection apparatus 1 acquires detection points by radar using laser light, groups the detection points, and obtains obstacle information from the detection points of each group. In particular, the obstacle detection apparatus 1 re-divides the group grouped at the current time in consideration of the grouping result (obstacle detection result) at the previous time in order to improve the grouping accuracy. For this purpose, the obstacle detection device 1 includes a laser radar 2 and an ECU [Electronic Control Unit] 3.

  In the present embodiment, the laser radar 2 corresponds to the laser detection means described in the claims, and each processing in the ECU 3 includes the grouping means, the dividing means, the object detection means, and the position estimation described in the claims. Corresponds to means.

  The laser radar 2 is a radar that detects an object using laser light. One or a plurality of laser radars 2 are provided according to the detection direction (detection region) of the obstacle, and are attached to a predetermined location (for example, the center of the front end) of the vehicle according to the detection direction of the obstacle. The laser radar 2 emits laser light at a predetermined angle in the left-right direction at regular intervals, and receives the reflected laser light. Each time the laser beam is received, the laser radar 2 calculates a relative distance from the detection point (reflection point) based on at least the time from emission to reception, and detects the detection point (reflection point) from the emission angle of the laser beam. The relative direction is calculated. Then, the laser radar 2 transmits information (relative distance, relative direction, etc.) about all the detection points that can be detected by one scan at regular intervals to the ECU 3 as a radar signal.

  The ECU 3 is an electronic control unit including a CPU [Central Processing Unit], a ROM [Read Only Memory], a RAM [Random Access Memory], and the like, and comprehensively controls the obstacle detection apparatus 1. The ECU 3 receives a radar signal from the laser radar 2 at regular intervals. Then, the ECU 3 performs obstacle detection processing using information on detection points included in the radar signal at regular intervals, and when the obstacle is detected, the ECU 3 is operated using the obstacle information as an obstacle information signal. Send to support device.

  The obstacle detection process will be described. When the ECU 3 acquires the detection point information, it performs normal grouping. As this grouping method, a conventional method is applied. For example, when the distance between two detection points is less than a threshold, the two detection points are set to the same group. In the case of the example shown in FIG. 2, at the current time, a group G1 composed of detection points P1,..., P12 and a group G2 composed of detection points P13,.

  The ECU 3 reads an obstacle detection result (corresponding to a grouping result) at a previous time from a predetermined area of the RAM. If there is information on two or more obstacles in the obstacle detection result at the previous time, the ECU 3 determines the position of the obstacle at the previous time for each obstacle at the previous time (corresponding to the group at the previous time), Based on the speed and direction of movement, the position moved by the current time is estimated. In the example shown in FIG. 2, there are three obstacles O1, O2, and O3 as the obstacle detection results at the previous time. The obstacle O1 is estimated to have moved from the position Op1 at the previous time to Op1 ′ at the current time, the obstacle O2 is estimated to have moved from the position Op2 at the previous time to Op2 ′ at the current time, and the obstacle O3 has been moved to the previous time. It is estimated that the current position has moved from position Op3 to Op3 ′ at the current time. At this time, the estimated current position Op1 'and the estimated current position Op2' are quite close, and the obstacle O1 and the obstacle O2 that existed apart at the previous time are close at the current time.

  Subsequently, the ECU 3 sequentially extracts the groups grouped by the above normal grouping, and performs the following processing for all the groups. The ECU 3 calculates the distance between the detection point and the estimated current position of each obstacle at the previous time for each detection point included in the group. The ECU 3 divides all detection points included in the group according to the estimated current position of the group at the nearest previous time. At this time, if the distance between the detection point and the estimated current position is greater than or equal to the threshold value, the ECU 3 determines that there is no relationship with the obstacle at the previous time, and does not divide according to the estimated current position. This threshold value is a threshold value for determining whether or not the detection point detected at the current time is a detection point for the obstacle detected at the previous time, and is set in advance by an experiment or the like. In the example shown in FIG. 2, in the case of the group G1 at the current time, the detection points P1, ..., P7 are closest to the estimated current position Op1 ', and the detection points P8, ..., P12 are the estimated current position Op2'. , P7 and detection points P8,..., P12. In the case of the group G2 at the current time, the detection points P13,..., P19 are closest to the estimated current position Op3 ′, but the distance between each detection point 13,..., P19 and the estimated current position Op3 ′. Is more than the threshold value, and is not related to the obstacle O3 and cannot be divided.

  When the detection points of the group at the current time are divided, the ECU 3 generates each divided detection point group as a new group and replaces the original group. In the example shown in FIG. 2, the group G1 at the current time is divided into two detection point groups, detection points P1,..., P7 and detection points P8,. .., P7 generates a group G11, detection points P8,..., P12 generate a group G12, and the group G1 is deleted. Further, the group G2 at the current time remains as it is. Therefore, the final grouping results at the current time include group G11 (detection points P1,..., P7), group G12 (detection points P8,..., P12), group G2 (detection points P13,. .., P19).

  When the processing for all the groups at the current time is completed, the ECU 3 determines, for each finally divided group, a group (that is, an obstacle) from the relative positions and relative directions of all the detection points included in the group. The position, size, speed, moving direction, and the like of this are calculated, and these obstacle detection results are temporarily stored in a predetermined area of the RAM. The speed and the moving direction are calculated based on the change from the position at the previous time using the obstacle information at the previous time (may also use past obstacle information such as the previous time). It should be noted that a group in which the number of detection points of the group is less than a predetermined number is not recognized as an obstacle and information on the obstacle is not calculated.

  If no obstacle is detected at the previous time, or if only one obstacle is detected, the group re-division process is not performed based on the obstacle detection result at the previous time. The result is used as it is.

  In the example shown in FIG. 3, two pedestrians W1 and W2 that have been detected as separate obstacles in the obstacle detection at the previous time approached at the current time, so that in the normal grouping at the current time, 2 This is a case where detection points P1,..., P15 for human pedestrians W1, W2 are grouped as one group G1. It is estimated that the pedestrian W1 has moved from the position Wp1 at the previous time to the position Wp1 'at the current time. The pedestrian W2 is estimated to have moved from the position Wp2 at the previous time to the position Wp2 'at the current time. By the group subdivision process, the detection points P1,..., P7 are closest to the estimated current position Wp1 ′, and the detection points P8,..., P15 are closest to the estimated current position Wp2 ′. ... It is divided into a group G11 composed of P7 and a group G12 composed of detection points P8,. As a result, even at the current time, the group G11 and the group G12 can be detected as two obstacles (pedestrians W1 and W2).

  With reference to FIG. 1, the operation | movement in the obstruction detection apparatus 1 is demonstrated. In particular, the processing in the ECU 3 will be described along the flowchart of FIG. FIG. 4 is a flowchart showing the flow of processing in the ECU of FIG.

  The laser radar 2 emits a laser beam at a predetermined angle in the left-right direction and receives the reflected light at regular time intervals. The laser radar 2 calculates information about each detection point and transmits a radar signal to the ECU 3.

  Every time a radar signal is received (every fixed time), the ECU 3 acquires information on detection points (S1). Then, the ECU 3 groups detection points by normal grouping (S2).

  The ECU 3 estimates the position at the current time from the position, speed, and moving direction at the previous time for all obstacles (groups) at the previous time (S3).

  The ECU 3 sequentially extracts the groups at the current time grouped in S2. Then, the ECU 3 divides all detection points included in the extracted group according to the estimated current position of the obstacle at the nearest previous time (S4). At this time, when there is no close estimated current position, it is not divided. When the detection points included in the extracted group are divided into several groups, the ECU 3 generates a new group for each of the divided detection points, and the group based on the plurality of generated groups (extracted Replace with (Group) (S5).

  The ECU 3 determines whether or not the processing for all groups at the current time grouped in S2 is completed (S6). If it is determined in S6 that the process for all groups has not been completed, the ECU 3 returns to S4, extracts the next group, and performs the process for this group.

  If it is determined in S6 that the processing for all the groups has been completed, the ECU 3 calculates obstacle information for each finally divided group based on the detection point information included in each group ( S7). Then, the ECU 3 transmits the obstacle information as an obstacle information signal to the driving support device.

  According to this obstacle detection device 1, after grouping the normal grouping at the current time, the grouped group is subdivided in consideration of the obstacle detection result at the previous time (final grouping result at the previous time), Even when a plurality of obstacles are grouped as one group, they can be subdivided into groups corresponding to the obstacles. As a result, appropriate grouping can be performed and obstacles can be detected with high accuracy. In particular, since the obstacle detection apparatus 1 calculates the estimated current position of the obstacle at the previous time and divides the detection points detected at the current time according to the estimated current position, it is possible to perform subdivision with high accuracy. it can.

  As mentioned above, although embodiment which concerns on this invention was described, this invention is implemented in various forms, without being limited to the said embodiment.

  For example, in the present embodiment, the present invention is applied to an obstacle detection device mounted on a vehicle. In addition to obstacles around the vehicle, various objects can be detected as detection targets according to the usage of the object detection device. In addition to the vehicle, the vehicle can be mounted on other objects such as a robot.

  In this embodiment, a laser radar is used as the radar detection means, but other radars such as a millimeter wave radar may be used.

  In the present embodiment, only the obstacle detection result at the previous time is used. However, the obstacle detection result at the previous time and the obstacle detection result before that may be used.

It is a block diagram of the obstruction detection apparatus which concerns on this Embodiment. It is explanatory drawing of the group subdivision method in ECU of FIG. It is an example of group subdivision. It is a flowchart which shows the flow of the process in ECU of FIG.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Obstacle detection apparatus, 2 ... Laser radar, 3 ... ECU

Claims (1)

Radar detecting means for detecting the position of the object;
When the distance between the two detection points detected this time by the radar detection means is less than the threshold value, each detection point detected by the radar detection means is made by grouping the two detection points into the same group. Grouping means for grouping,
A dividing unit that divides the group grouped this time by the grouping unit based on a plurality of previous groups into a plurality of groups;
For each final divided was the group with the grouping unit or the dividing means, and object detecting means for detecting an object by using the detection points included in the group,
A position estimation means for estimating the current position of the previous group based on the information of the previous group;
The position estimation means estimates the current position for each previous group detected by the object detection means as an object,
When the object detection unit has previously detected two or more objects, the dividing unit determines whether the estimated position estimated by the position estimation unit for each detection point included in the currently grouped group and the radar detection unit An object detection apparatus that calculates a distance from each detected detection point and divides the group by dividing the detection point according to the closest estimated position when the distance is less than a preset threshold.

Images