JP2022152404A - Axis deviation detection device - Google Patents

Abstract

To provide an axis deviation detection device which can easily detect a deviation of an optical axis of a lidar mounted on a vehicle.SOLUTION: An axis deviation detection device 101 comprises: a lidar 10 which detects a circumferential state of a vehicle and is provided so that an engine hood as a part of a vehicle body is included in a detection range; and an axis deviation calculation part 51 which calculates, based upon positions of a plurality of points on a surface of the engine hood detected by the lidar 10, a quantity of axis deviation from a reference axis which is a reference of an optical axis of the lidar 10. The axis deviation calculation part 51 calculates the quantity of axis deviation based upon a vector obtained by connecting points of both right and left edge parts of the engine hood or a normal vector to the surface which is calculated based upon a position of a point group.SELECTED DRAWING: Figure 7

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a shaft deviation detection device for detecting a deviation amount of an output shaft of a vehicle-mounted detector such as a lidar.

As a device of this type, conventionally, an optical axis adjusting device is arranged in front of a rider unit mounted on a vehicle, a laser beam is emitted from the lidar unit to the optical axis adjusting device, and light is emitted based on the irradiation result. A device adapted to detect axial misalignment has been described (see, for example, US Pat.

JP 2020-56891 A

However, in the device described in Patent Document 1, a special jig called an optical axis adjustment device must be prepared in order to detect the deviation of the optical axis (output axis), and the deviation of the axis can be easily corrected. cannot be detected.

A shaft deviation detection device that is one aspect of the present invention detects a surrounding situation of a vehicle, and includes a detector that is provided so that an object that is a part of the vehicle body is included in a detection range; an axis deviation calculator for calculating the amount of axis deviation from a reference axis, which is the reference of the output axis of the detector, based on the positions of the plurality of points on the surface of the object.

ADVANTAGE OF THE INVENTION According to this invention, the axis deviation of the detector mounted in the vehicle can be detected easily.

1 is a perspective view of the entire vehicle, which is obliquely seen from the front and has the shaft deviation detection device according to the embodiment of the present invention; FIG. FIG. 2 is a diagram showing a detectable range from a viewpoint from a rider mounted on the vehicle of FIG. 1; FIG. 4 is a diagram schematically showing an example of the positional relationship between the rider and the edge points when viewed forward from inside the vehicle, with the rider attached at the correct position. FIG. 4 is a diagram schematically showing an example of the positional relationship between the rider and the edge points when viewed forward from inside the vehicle with the mounting position of the rider shifted. FIG. 5 is a diagram schematically showing an example of the positional relationship between the rider and the edge points with the rider as a reference when the mounting position of the rider is displaced; FIG. 4 is a diagram schematically showing an example of the positional relationship between the rider and feature points on the hood as viewed from the side of the vehicle, with the rider attached at a normal position; FIG. 5 is a diagram schematically showing an example of the positional relationship between the rider and characteristic points on the hood when viewed forward from inside the vehicle with the mounting position of the rider shifted. FIG. 10 is a diagram schematically showing an example of the positional relationship between the rider and feature points on the hood with the rider as a reference when the mounting position of the rider is displaced; FIG. 2 is a block diagram showing the main configuration of the shaft deviation detection device according to the embodiment; The front view which shows an example of the detection range from the lidar attached to the regular position. FIG. 4 is a front view showing an example of a detection range from a lidar mounted in a displaced state; FIG. 11 is a front view showing another example of a detection range from a lidar mounted in a displaced state; FIG. 8 is a flow chart showing an example of shaft deviation detection processing executed by the controller in FIG. 7 ; FIG.

An embodiment of the present invention will be described below with reference to FIGS. 1 to 10. FIG. A shaft deviation detection device according to an embodiment of the present invention is mounted on a vehicle. FIG. 1 is a perspective view of a vehicle 100 having a shaft deviation detection device according to an embodiment of the present invention, viewed obliquely from the front. In FIG. 1, arrows indicate the front-rear direction (length direction), the left-right direction (width direction), and the up-down direction (height direction) of the vehicle 100 .

The vehicle 100 is configured, for example, as an automatically driving vehicle having an automatic driving function or a vehicle having a driving support function. Therefore, the vehicle 100 is provided with detectors that detect dynamic and static objects (objects) around the vehicle 100, such as other vehicles, pedestrians, bicycles, guardrails, and median strips. The detector consists of a lidar (LiDAR) that measures the distance and direction to an object by irradiating it with a pulsed laser beam and detecting the reflected light from the object, A radar that detects the distance and direction to an object by detecting , and a camera that has an imaging device such as a CCD or CMOS and captures an image of the surroundings of the vehicle. An axis deviation detection device is used to detect the amount of axis deviation of these detectors. In the following description, it is assumed that the detector to which the axis deviation detection device is applied is the lidar.

As shown in FIG. 1, a rider 10 is attached to the front end and the central portion in the left-right direction of a ceiling portion 1 inside the vehicle so as to face an object in front of the vehicle through the windshield 2 by the rider 10 . detected. When the vehicle 100 is shipped from the factory, the rider 10 is installed at a regular position after adjusting the optical axis. That is, the optical axis adjustment (for example, adjustment of the mounting position of the rider 10) is performed so that the optical axis of the rider 10 extends from the front surface of the rider 10 in a predetermined direction (reference direction). It is possible to accurately detect the position of an object contained within the range. Note that the reference direction is, for example, forward or forward and slightly obliquely downward.

However, as the vehicle 100 has been used for many years, or as some kind of shock acts on the vehicle 100, the rider 10 may be displaced from the mounting position or the like. As a result, the position of the optical axis of the lidar 10 may be shifted from the reference direction, and the detection accuracy may deteriorate. In such a case, it is necessary to calibrate the detection value of the lidar 10. To do so, it is necessary to accurately obtain the amount of deviation of the optical axis of the lidar 10 from the reference axis. If the amount of deviation is detected by, for example, installing a special jig in front of the vehicle 100, preparation of the jig is costly and time-consuming. In addition, when the vehicle 100 is in an inclined posture, the relative positional relationship with a jig placed away from the vehicle 100 is shifted, making it difficult to accurately detect the amount of deviation. There are many restrictions during work.

Therefore, in the present embodiment, the displacement amount of the optical axis of the lidar 10 is detected as follows without using a jig. Specifically, as a detection image is shown in FIG. 1, the rider 10 is provided so as to include not only an object in front of the vehicle but also a part of the vehicle body, that is, the bonnet 3, in the detection range.

FIG. 2 is a diagram showing the detectable range from the viewpoint from the rider 10, and the area within the rectangular frame in FIG. 2 is the detectable range AR1. The detectable range AR1 corresponds to a part of or the entire area of the windshield 2 . As shown in FIG. 2 , the detectable range AR1 includes part of the upper surface of the bonnet 3 . More specifically, it includes the entire left and right front portion of the upper surface of the bonnet 3 . In addition, in FIG. 2, illustration of an object in front of the bonnet 3 is omitted.

The rider 10 detects the positions of a plurality of uneven points P (characteristic points) on the upper surface of the bonnet 3 . That is, the distance and direction from the rider 10 to each feature point P are detected. The feature points P include edge points P1 and P2 at both left and right corners of the bonnet 3 .

FIG. 3 is a diagram schematically showing an example of the positional relationship between the rider 10 and the edge points P1 and P2 when looking forward from inside the vehicle. FIG. 3 especially shows the positional relationship at the time of shipment from the factory. Therefore, the lidar 10 is mounted at a regular position, and the direction of the optical axis of the lidar 10 matches the reference direction. At this time, a reference line indicating the posture of the rider 10 (for example, a reference line La extending in the horizontal direction) is parallel to the line segment L0 connecting the edge points P1 and P2.

FIG. 4A is a diagram showing a state in which the mounting position of the rider 10 is deviated due to aged deterioration or the like (for example, a state in which the mounting position is tilted in the left-right direction). In FIG. 4A, the reference line La is not parallel to the line segment L1 connecting the edge points P1 and P2, but is inclined. Therefore, the direction of the optical axis of the lidar 10 is deviated from the reference direction.

The solid lines in FIG. 4B show the positions of the edge points P1 and P2 as viewed from the rider 10 (FIG. 4A) with the reference line La tilted, that is, with the rider 10 as a reference. Note that the dotted line in FIG. 4B indicates the initial state (FIG. 3) before the reference line La is inclined. As shown in FIG. 4B, when the mounting position of the rider 10 shifts, the edge points P1 and P2 detected by the rider 10, that is, the positions of the edge points P1 and P2 based on the rider shift. Therefore, the line segment L1 connecting the edge points P1 and P2 is inclined with respect to the line segment L0. The amount of deviation of the line segment L1 from the line segment L0 in this case corresponds to the amount of deviation of the optical axis of the rider 10 . This amount of deviation can be calculated, for example, as follows, by processing in the vehicle-mounted controller.

First, in an inspection process when the vehicle 100 is shipped from the factory, the controller identifies the edge points P1 and P2 from among the feature points detected by the rider 10, and acquires the three-dimensional position coordinates of the edge points P1 and P2. (Fig. 3). For the edge points P1 and P2, for example, among a plurality of feature points P within a predetermined distance Lx from the rider 10, those having a large change in position coordinates between adjacent feature points are extracted. can be identified by The predetermined distance Lx corresponds to the distance L3 from the rider 10 to the bonnet 3 plus a predetermined length ΔL (for example, several centimeters to several tens of centimeters). That is, the predetermined distance Lx is a distance that includes at least the edge points P1 and P2. The controller then calculates a vector v0 (x0, y0, z0) along the line segment L0 connecting the edge points P1 and P2. A vector v0 along the line segment L0 is called a reference vector. The calculated reference vector v0 is stored in the memory of the onboard controller.

Thereafter, at the present time after a predetermined period of time has passed since the shipment from the factory, the controller uses the position coordinates of the edge points P1 and P2 detected by the rider 10 to determine the vector v1 (x1 , y1, z1) are calculated (FIG. 4A). Then, the controller uses the stored reference vector v0 (x0, y0, z0) and the calculated vector v1 (x1, y1, z1) to calculate the axis deviation amount α1 (φ, θ, ψ). . For example, the axis deviation amount α1 is calculated by applying the vectors v0 and v1 to the following determinant (I) and solving the three equations for the above three variables.

As described above, the axis deviation amount α1 can be calculated using the vectors v0 and v1 connecting the left and right edge points P1 and P2. Note that this is called calculation of the axis deviation amount α1 based on the edge point vector. The controller can also calculate the amount of misalignment of the rider 10 by other methods. For example, the normal vector of the bonnet 3 can be used, that is, the axis deviation amount α2 can be calculated based on the normal vector.

Calculation of the axis deviation amount α2 based on the normal vector will be described with reference to FIGS. 5 to 6B. FIG. 5 is a diagram schematically showing an example of the positional relationship between the rider 10 and the feature point P on the hood 3 as viewed from the side of the vehicle. In FIG. 5, the detectable range AR2 by the rider 10 is schematically indicated by a triangle (a chain double-dashed line). Similar to FIG. 3, FIG. 5 shows the positional relationship of the vehicle 100 when it is shipped from the factory. The posture of the rider 10 in this case is represented, for example, by a reference line Lb extending in the front-rear direction. The direction of the reference axis of the optical axis of the rider 10 matches, for example, the direction of the reference line Lb.

As shown in FIG. 5, a plurality of feature points P are included in the detectable range AR2. Although illustration is omitted, the detectable range AR2 includes a plurality of characteristic points P not only in the longitudinal direction of the bonnet 3 but also in the lateral direction. A detection plane (referred to as a reference plane) along the upper surface of the bonnet 3 can be generated from the point cloud data of these plurality of characteristic points P. FIG. FIG. 5 schematically shows a normal vector (referred to as a reference normal vector) n0 to this reference plane.

FIG. 6A is a diagram showing a state in which the mounting position of the rider 10 is deviated due to aged deterioration or the like (for example, a state in which the mounting position is tilted downward). In FIG. 6A, the slope of the reference line Lb is different from that in FIG. Therefore, the direction of the optical axis of the lidar 10 deviates from the reference direction, and a detection plane different from the reference plane in FIG. 5 is generated by the characteristic point P within the detectable range AR2. Therefore, the normal vector n1 to the detection surface is different from the reference normal vector n0 in FIG.

A solid line in FIG. 6B indicates the position of the feature point P as seen from the rider 10 (FIG. 6A) with the reference line Lb tilted. Note that the dotted line in FIG. 6B indicates the initial state (FIG. 5) before the reference line Lb is tilted. As shown in FIG. 6B, when the mounting position of the rider 10 shifts, the position of the feature point P detected by the rider 10 shifts. For this reason, the normal vector n1 obtained from the plurality of feature points P deviates from the reference normal vector n0. The amount of deviation of the normal vector n1 from the reference normal vector n0 in this case corresponds to the amount of deviation of the optical axis of the rider 10. FIG. This amount of deviation can be calculated, for example, as follows, by processing in the vehicle-mounted controller.

First, in an inspection process when the vehicle 100 is shipped from the factory, the controller selects a plurality of feature points P located within a predetermined distance Lx from among the feature points detected by the rider 10, that is, a plurality of feature points P on the upper surface of the bonnet 3. Select point P (FIG. 5). Then, the controller generates a reference plane along the upper surface of the bonnet 3 from the plurality of selected feature points P, calculates a reference normal vector n0 (xn0, yn0, zno) for the reference plane, and calculates the reference normal vector n0 (xn0, yn0, zno) Vector n0 is stored in the memory of the on-board controller.

Thereafter, at present, after a predetermined period of time has passed since the shipment from the factory, the controller generates a detection plane using a plurality of feature points P on the upper surface of the bonnet 3 detected by the rider 10, and normal vector n1 to the detection plane. Calculate (xn1, yn1, zn1) (FIG. 6A). Then, the controller uses the stored reference normal vector n0 (xn0, yn0, zno) and the calculated normal vector n1 (xn1, yn1, zn1) to Calculate For example, by replacing (xo, y0, z0) and (x1, y1, z1) in the determinant of the above formula (I) with (xn0, yn0, zn0) and (xn1, yn1, zn1), respectively, the axis deviation Calculate the quantity α2.

FIG. 7 is a block diagram showing the main configuration of the shaft deviation detection device 101 according to this embodiment. As shown in FIG. 7 , the axis deviation detection device 101 has a lidar 10 and a controller 50 . The controller 50 includes a computer having a CPU, ROM, RAM, and other peripheral circuits, and has an axis deviation calculator 51, a calibrator 52, and a memory 53 as functional components.

The storage unit 53 stores a reference value calculated based on detection values detected by the rider 10 in the inspection process at the time of shipment from the factory, that is, a reference vector v0 (FIG. 3) obtained by connecting the edge points P1 and P2. , and the reference normal vector n0 (FIG. 5) of the reference plane generated by the plurality of feature points P are stored. Various thresholds such as the predetermined distance Lx are also stored in the storage unit 53 .

The axis deviation calculator 51 calculates the axis deviation amounts α1 and α2 of the optical axis of the rider 10 based on the edge point vector reference (FIG. 4B) and the normal vector reference (FIG. 6B) described above. That is, the axis deviation amount α1 is calculated based on the reference vector v0 stored in advance in the storage unit 53 and the vector v1 obtained by connecting the edge points P1 and P2 detected by the rider 10 at this time. Also, based on the reference normal vector n0 stored in advance in the storage unit 53 and the normal vector n1 obtained from the feature point P detected by the rider 10 at this time, the axis deviation amount α2 is calculated. Then, the average value of these two axis deviation amounts α1 and α2 is calculated and stored in the storage unit 53 as the current axis deviation amount α3.

The calibrating unit 52 calibrates the detected value of the rider 10 using the shaft misalignment amount α3 stored in the storage unit 53 . That is, the coordinates of the feature points detected by the rider 10 are coordinate-transformed based on the axis deviation amount α3.

8, 9A, and 9B are front views showing how the front of the vehicle is viewed from the rider 10. FIG. 8 is a front view of vehicle 100 when it is shipped from the factory, like FIGS. 3 and 5. FIG. 9A is a front view when the mounting position of the rider 10 is displaced counterclockwise, and FIG. 9B is a front view when the mounting position of the rider 10 is shifted. is a front view in the case where is deviated downward;

As shown in FIG. 8 , an actual detection range AR3 is set inside the detectable range AR1 of the rider 10 . As shown in FIGS. 9A and 9B, when the mounting position of the rider 10 shifts, the detection range AR3 also shifts. The calibration unit 52 corrects the detection range AR3 by performing calibration according to this deviation. That is, in the example of FIG. 9A, the detection range AR3 is corrected clockwise, and in the example of FIG. 9B, the detection range AR3 is corrected upward.

FIG. 10 is a flowchart showing an example of processing (axis deviation detection processing) mainly executed by the axis deviation calculator 51 of the controller 50 of FIG. The processing shown in this flow chart is started when, for example, a user inputs a shaft deviation detection command. The shaft deviation detection process may be performed at a predetermined cycle without depending on the instruction from the user. It is preferable to perform the shaft deviation detection process after the power switch of the vehicle 100 is turned on and before the vehicle 100 starts running.

First, in step S1, data of feature points detected by the rider 10 are obtained. Next, in step S2, feature points P whose distance from the rider 10 is within a predetermined distance Lx are extracted from the feature points acquired in step S1. Next, in step S3, edge points P1 and P2 are specified from among the feature points P extracted in step S2, and a vector v1 connecting the edge points P1 and P2 is calculated. Next, in step S4, the calculated vector v1 and the pre-stored reference vector v0 are used to calculate the axis deviation amount α1 based on the edge point vector.

Next, in step S5, a reference plane corresponding to the upper surface of the bonnet 3 is generated using the characteristic points P extracted in step S2, and a normal vector n1 perpendicular to the reference plane is calculated. Next, in step S6, the calculated normal vector n1 and the pre-stored reference normal vector n0 are used to calculate the axis deviation amount α2 based on the normal vector. Next, in step S7, an axis deviation amount α3, which is an average value of the axis deviation amounts α1 and α2, is calculated. Finally, the shaft misalignment amount α3 is stored in the memory, and the shaft misalignment detection process ends. Thereafter, the detected value of the rider 10 is calibrated using the stored axis deviation amount α3.

According to this embodiment, the following effects can be obtained.
(1) The axis deviation detection device 101 according to the present embodiment detects the surrounding conditions of the vehicle 100, and the rider 10 is provided so that the bonnet 3, which is a part of the vehicle body, is included in the detection range, and the rider 10 and an axis deviation calculation unit 51 for calculating the amount of deviation from the reference axis, which is the reference of the optical axis of the rider 10, based on the positions of the plurality of feature points P on the surface of the hood 3 detected by (Fig. 7 ).

As a result, when the mounting position of the rider 10 deviates due to aged deterioration or the like, the amount of axis deviation can be easily calculated without separately using a jig or the like for detecting axis deviation. Therefore, the axial deviation amount can be detected at low cost. Further, since the amount of axis deviation is detected based on the detection data of the upper surface of the part of the vehicle body (the bonnet 3) detected by the rider 10, even if the attitude of the vehicle 100 changes, the influence of the attitude change can be avoided. The amount of axis misalignment can be detected, and the amount of axis misalignment can be easily detected.

(2) The axis deviation detection device 101 detects the amount of axis deviation with reference to characteristic points on the bonnet 3 extending in the longitudinal direction and the lateral direction of the vehicle 100 . Since the bonnet 3, which has a relatively large area, is used as a reference, errors in detecting the amount of shaft misalignment are small.

(3) A plurality of feature points P on the hood detected by the rider 10 include edge points P1 and P2 on both left and right edges of the hood 3 (FIG. 2). The axis deviation calculator 51 calculates the axis deviation amount α1 based on the vectors v0 and v1 obtained by connecting the left and right edge points P1 and P2 (FIG. 10). This makes it possible to easily and accurately calculate the amount of misalignment.

(4) A plurality of feature points P on the hood detected by the rider 10 are a point group that generates a plane (reference plane, detection plane) corresponding to the surface shape of the bonnet 3 (Figs. 6A). The axis deviation calculator 51 calculates the axis deviation amount α2 based on the normal vectors n0 and n1 to the surface calculated based on the position of the point group (FIG. 10). This makes it possible to easily and accurately calculate the amount of misalignment.

The above embodiment can be modified in various forms. Some modifications will be described below. In the above embodiment, the lidar 10 is used as a detector for detecting the surrounding conditions of the vehicle 100, and the position of the bonnet 3, which is a part of the vehicle body, is detected by the lidar 10. may be a cover member other than the bonnet. A part of the vehicle body other than the cover member may be used as the target, and the rider may detect the position of the target.

In the above embodiment, the rider 10 is provided at the front part inside the vehicle, but the rider 10 can be provided at other positions such as the rear part inside the vehicle or outside the vehicle. In other words, the mounting position of the lidar is not limited to that described above as long as it is provided so that an object that is a part of the vehicle body is included in the detection range. In the above embodiment, the amount of deviation of the optical axis (output axis) of the lidar 10 is detected, but the amount of deviation of the output axis of another detector (eg radar) may be detected. Therefore, detectors to which the present invention is applied are not limited to lidars.

In the above embodiment, the axis deviation calculator 51 is based on the vectors v0 and v1 obtained by connecting the corner points (edge points P1 and P2) of the left and right ends of the bonnet 3 at symmetrical positions in the left and right direction, or , based on the normal vectors n0 and n1 to the upper surface of the bonnet 3 obtained from the point cloud data of the bonnet 3, the shaft deviation amounts α1 and α2 are calculated, but the configuration of the shaft deviation calculation unit is not limited to this. . That is, if the amount of axis deviation from the reference axis, which is the reference of the output axis of the detector, is calculated based on the positions of a plurality of points on the surface of the object detected by the detector, the axis deviation calculation unit Any configuration is possible. Instead of calculating the average value of the shaft misalignment amounts α1 and α2, the shaft misalignment amount α1 or α2 may be set as it is as the shaft misalignment amount α3.

The above description is merely an example, and the present invention is not limited by the above-described embodiments and modifications as long as the features of the present invention are not impaired. It is also possible to arbitrarily combine one or more of the above embodiments and modifications, and it is also possible to combine modifications with each other.

3 bonnet 10 rider 50 controller 51 axis deviation calculator 100 vehicle 101 axis deviation detector P characteristic point P1, P2 edge point v0 reference vector v1 vector n0 reference normal vector n1 normal Vector, α1, α2, α3 Amount of misalignment

Claims (4)

a detector that detects the surrounding conditions of the vehicle and is provided so that an object that is a part of the vehicle body is included in the detection range;
an axis deviation calculator that calculates an axis deviation amount from a reference axis, which is a reference of the output axis of the detector, based on the positions of the plurality of points on the surface of the object detected by the detector. An axis deviation detection device characterized by: In the axis deviation detection device according to claim 1,
The shaft deviation detection device, wherein the object is a cover member extending in the front-rear direction and the left-right direction of the vehicle. In the axis deviation detection device according to claim 2,
The plurality of points include points on both left and right edges of the cover member,
The axis deviation detection device, wherein the axis deviation calculator calculates the amount of axis deviation based on a vector obtained by connecting points of the left and right edges. In the axis deviation detection device according to claim 2,
the plurality of points is a point group that generates a surface corresponding to the surface shape of the cover member;
The axis deviation detection device, wherein the axis deviation calculation unit calculates the axis deviation amount based on a normal vector to the surface calculated based on the position of the point group.

Images