JP7344744B2 - Roadside edge detection method and roadside edge detection device - Google Patents

Description

The present invention relates to a roadside edge detection method and a roadside edge detection device.

In vehicle driving support systems, the position of a roadside object such as a curb on the road surface on which the vehicle is traveling is detected as the roadside edge, and driving assistance is provided so that the vehicle drives inside the detected roadside edge. .

For example, according to the driving support system disclosed in Patent Document 1, position information of a plurality of measurement points along the width direction of the road surface is acquired using a laser radar, and discontinuous portions at a group of measurement points lined up in the vehicle width direction are detected. Detected as roadside end.

Japanese Patent Application Publication No. 2016-45507

In Patent Document 1, a loss may occur in the position information measured at a group of measurement points lined up in the vehicle width direction. In such a case, there is a possibility that the discontinuous portion that occurs due to the lack of positional information does not correspond to the position of the actual roadside object, and the roadside end cannot be detected accurately.

The present invention has been made to solve the above problems, and provides a roadside edge detection method and a roadside edge detection device that improve the detection accuracy of roadside edges.

The roadside edge detection method of the present invention is a roadside edge detection method in which a roadside edge on which a roadside object is provided is detected by a roadside edge detection device on a road surface on which a vehicle travels. The roadside edge detection method includes a road surface height detection step that detects the height of the road surface in a predetermined range around the vehicle, and a road surface height detection step that detects the height of the road surface in the predetermined range detected in the road surface height detection step. A total sum calculation step of calculating the sum along the vehicle running direction according to the position in the width direction, and determining that a position in the vehicle width direction where the sum calculated in the sum calculation step exceeds a threshold value is the roadside edge. , a roadside edge determination step.

According to the roadside edge detection method of the present invention, it is possible to improve the accuracy of roadside edge detection.

FIG. 1 is a schematic configuration diagram of a driving support system that executes the roadside edge detection method according to the first embodiment. FIG. 2 is an explanatory diagram showing acquisition control of ranging data. FIG. 3 is an explanatory diagram of a method for generating a grid map from ranging data. FIG. 4 is a diagram showing an example of an occupancy grid map. FIG. 5 is a flowchart of roadside edge detection control. FIG. 6 is a diagram showing measurement points of distance measurement data from above. FIG. 7 is an explanatory diagram of part of the processing in the roadside edge detection control. FIG. 8 is an explanatory diagram of an example of a detailed example of roadside edge detection control. FIG. 9 is a schematic configuration diagram of a driving support system that executes the roadside edge detection method according to the second embodiment. FIG. 10 is a schematic configuration diagram of a driving support system that executes the roadside edge detection method according to the third embodiment. FIG. 11 is a diagram showing a comparison between a case where the driving lane is straight and a case where the driving lane is curved. FIG. 12 is a diagram showing an example of division of a grid map in a modified example.

DESCRIPTION OF THE PREFERRED EMBODIMENTS A roadside edge detection method and a roadside edge detection device according to an embodiment of the present invention will be described below with reference to the drawings.

(First embodiment)
FIG. 1 is a schematic configuration diagram of a driving support system that executes the roadside edge detection method according to the first embodiment.

The driving support system 100 is a system installed in a vehicle for the purpose of assisting a driver in driving the vehicle. The driving support system 100 includes a moving object detection section 11, a distance measurement sensor 12, a distance measurement data filter section 13, an odometry measurement section 14, a distance measurement data storage section 15, an accumulated distance measurement data 16, and an occupancy grid. It includes a map generation section 17, an occupancy grid map 18, an occupancy grid map addition section 19, a discontinuous boundary detection section 20, a roadside edge determination section 21, and an output section 22.

Among these components, a distance measurement data filter section 13, a distance measurement data storage section 15, an accumulated distance measurement data 16, an occupancy grid map generation section 17, an occupancy grid map 18, an occupancy grid map addition section 19, and a discontinuous boundary detection section. 20, the roadside edge determination section 21, and the output section 22 are integrally configured as the roadside edge detection device 10. In the roadside edge detection device 10, the distance measurement data filter section 13, the distance measurement data storage section 15, and the accumulated distance measurement data 16 to output section 22 may be configured in the same microcomputer, or may be configured in different microcomputers. may be configured. Further, the moving object detection section 11, the distance measurement sensor 12, and the odometry measurement section 14 are provided as sensors separate from the roadside edge detection device 10.

Below, details of each configuration of the driving support system 100 will be explained.

The moving object detection section 11 detects a moving object (moving object) existing around the vehicle, and outputs information indicating the position of the detected moving object to the ranging data filter section 13. For example, the moving object detection unit 11 is an image processing unit including a camera, and detects the position of a moving object by analyzing captured time-series images.

The distance sensor 12 is a sensor that measures the distance and direction from the vehicle of objects around the vehicle, that is, the object position with respect to the vehicle, and sets a plurality of measurement points on the surface of the road surface and structures on the road surface. and measure the distance to those measurement points. In addition, in the following, the road surface indicated as a distance measurement target by the distance measuring sensor 12 includes roadside objects arranged on the road surface. Therefore, the ranging data acquired by the ranging sensor 12 shows undulations caused by structures and the like in the plane direction where the road surface exists.

The distance measuring sensor 12 is not limited to a laser radar (LiDAR), but may be a stereo camera or the like. The type of distance measuring sensor 12 is not limited as long as it can measure relative distances and directions to a plurality of measurement points with respect to the vehicle. Therefore, the position as distance measurement data is indicated by a vehicle coordinate system, which is a relative coordinate system with the vehicle position as a reference.

The distance measurement data obtained by the distance measurement sensor 12 and the position information of the moving object obtained by the moving object detection section 11 are input to the distance measurement data filter section 13 . The distance measurement data filter section 13 erases the distance measurement data in the area where the moving object detected by the moving object detection section 11 exists from the distance measurement data acquired by the distance measurement sensor 12, and uses the erased distance measurement data. , and output to the distance measurement data storage section 15.

When the odometry measurement unit 14 acquires odometry information regarding a running vehicle, the odometry measurement unit 14 outputs the odometry information to the ranging data storage unit 15 . The odometry information includes the vehicle's running speed, running angle, etc., and can be used for creating a running trajectory, etc.

The distance measurement data storage section 15 receives input of the odometry information of the vehicle acquired by the odometry measurement section 14 and the distance measurement data that has been filtered by the distance measurement data filter section 13 . Then, in response to these inputs, the distance measurement data storage unit 15 converts the distance measurement data observed in the vehicle coordinate system to the odometry coordinate system, and uses a plurality of distance measurement data converted to the odometry coordinate system, Accumulated ranging data 16 is created.

The accumulated distance measurement data 16 is obtained by accumulating distance measurement data for a predetermined period, and each distance measurement data is represented by an absolute coordinate system based on the position where the vehicle is present at the start of the predetermined period. Then, the accumulated distance measurement data 16 is outputted by the distance measurement data storage section 15 to the occupancy grid map generation section 17.

The occupancy grid map generation unit 17 generates a grid map of a predetermined two-dimensional area using the accumulated distance measurement data 16, and creates an occupancy grid map 18 by cutting out a part of the grid map. Then, the occupancy grid map generation unit 17 transmits the created occupancy grid map 18 to the occupancy grid map addition unit 19.

The grid map generated by the occupancy grid map generation unit 17 divides a two-dimensional predetermined area around the vehicle into grids, and discretely indicates the presence of objects in these grids. For each grid, the height of objects present in that grid and the corresponding occupancy value are indicated. In other words, a grid map is a set of grids in which the undulations of the road surface, the presence of structures on the road surface, etc. are discretized in a predetermined area of the road surface where curbs and the like exist.

As grid maps, there are known ones that show discrete values in an XY orthogonal coordinate system and discrete values in an r-θ polar coordinate system. In this embodiment, the two-dimensional grid map generated by the occupancy grid map generation unit 17 is expressed in an XY orthogonal coordinate system. Note that details of this grid map will be explained later using FIG. 3 and the like.

The occupancy grid map addition unit 19 adds occupancy values of grids extending in the vehicle running direction to the occupancy grid map 18 generated by the occupancy grid map generation unit 17 according to the position from the vehicle in the vehicle width direction. Calculate the total sum. As a result, the occupancy grid map addition section 19 generates a distribution of the sum of the occupancy values corresponding to the positions in the vehicle width direction in the traveling direction, and outputs the distribution to the discontinuous boundary detection section 20 .

The discontinuous boundary detection unit 20 detects, as a discontinuous boundary, a position in the vehicle width direction where the total sum exceeds a predetermined threshold in the distribution of the sum of occupancy values obtained by the occupancy grid map addition unit 19. . Then, the discontinuous boundary detection section 20 outputs the detected discontinuous boundary to the roadside edge determination section 21. Note that the detection of the position where the sum exceeds a predetermined threshold value is performed from the position where the vehicle is present to both left and right sides in the vehicle width direction.

When receiving input of a plurality of discontinuous boundaries detected by the discontinuous boundary detection unit 20, the roadside edge determination unit 21 determines which of these discontinuous boundaries are located on each of the left and right sides from the vehicle in the vehicle width direction. The two closest discontinuous boundaries are selected, these two discontinuous boundaries are determined to be roadside edges, and the results are output to the output unit 22.

The output unit 22 outputs the position of the roadside edge determined by the roadside edge determination unit 21 in the vehicle width direction to an automatic driving system or the like outside the driving support system 100. For example, an automatic driving system uses the received position information of roadside edges to support the driving of a vehicle so that the vehicle travels between roadside edges.

Here, an outline of the acquisition control of distance measurement data by the distance measurement sensor 12 will be explained.

FIG. 2 is an explanatory diagram showing acquisition control of distance measurement data by the distance measurement sensor 12. This figure shows a situation in which the vehicle 200 travels in a three-dimensional space having x, y, and z axes. Note that the traveling direction corresponds to the x-axis, the vehicle width direction corresponds to the y-axis, and the height direction corresponds to the z-axis.

The distance measurement sensor 12 sets a plurality of measurement points in a predetermined rectangular area on the xy plane at a predetermined period, and measures distances to these measurement points. In this figure, the distance is measured at four timings: time t-1, t, t+1, and t+2. The distance measurement sensor 12 then generates distance measurement data using this measurement result. The distance measurement data indicates the presence or absence of an object at an arbitrary point in a three-dimensional predetermined area in a relative coordinate system with vehicle 200 as a reference. Therefore, by using the distance measurement data, it is possible to know the height of the road surface including roadside objects such as curbstones.

According to this figure, the distance measurement data at time t-1 includes height information of a cylindrical obstacle on the rear left when the traveling direction is set as the front. The distance measurement data at time t includes height information of a rectangular parallelepiped obstacle on the left front. Further, the distance measurement data at time t+1 includes height information of a cylindrical obstacle in the right front. In this way, the distance measurement data indicates the position in the xy plane and the height in the z direction of the structure on the road surface.

Note that the coordinate system of the distance measurement data is converted by the distance measurement data storage section 15. Specifically, the distance measurement data storage section 15 uses odometry information input from the odometry measurement section 14 to determine the position of the vehicle 200 at each measurement timing (t-1 to t+2) using odometry coordinates. . Then, the distance measurement data storage section 15 converts the distance measurement data that has passed through the distance measurement data filter section 13 into an absolute coordinate system based on a predetermined location indicated by odometry coordinates from a relative coordinate system based on the vehicle 200. Perform the conversion to the system.

Finally, the distance measurement data storage section 15 generates accumulated distance measurement data 16 using the plurality of distance measurement data converted into the absolute coordinate system. Then, the occupancy grid map generation unit 17 generates a grid map using the accumulated distance measurement data 16, and generates an occupancy grid map 18 by extracting a part of the area of the generated grid map.

Below, a method for generating a grid map from distance measurement data will be explained using FIG. 3, and an example of the occupancy grid map 18, which is a part of the generated grid map, will be explained using FIG.

FIG. 3 is an explanatory diagram of a method for generating a grid map from ranging data.

As described above, the distance measurement sensor 12 sets a plurality of measurement points x _j in a predetermined area around the distance measurement sensor 12, obtains distances to these measurement points x _j , and obtains distance measurement data. seek. The distance measurement data has a normal distribution of covariance Σ with the measurement point x _j as the average (center of distribution) at each arbitrary point in a predetermined three-dimensional area, corresponding to one measurement point x _j . shown. Note that the covariance Σ corresponds to the probability of existence of an object. That is, for one measurement point x _j , a normal distribution (Gaussian distribution) of the covariance Σ in the three-dimensional region is obtained, which has a peak at the measurement point x _j . This means that each of the measurement points x _j defines a normal distribution with a peak at itself, and influences the distribution value of the covariance Σ at any point in the three-dimensional space.

Therefore, using the normal distribution of covariance Σ corresponding to all measurement points x _j , for each grid point x _i provided on the road surface which is a two-dimensional plane (xy plane), the z direction (height direction) Calculate the sum of covariance Σ. Thereby, it is possible to obtain an occupancy value corresponding to the height in the z direction of the structure existing at the grid point x _i in the xy plane. In addition, in calculating the occupancy value, the distance r _j from the ranging sensor 12 to the measurement point x _j and the height h _j from the xy plane at the measurement point x _j are weighted to the distribution value of the covariance Σ. Riding.

The farther the distance r _j from the distance measurement sensor 12 to the measurement point x _j becomes, the sparser the distribution of the measurement points x _j becomes. Therefore, by weighting the distribution value of the covariance Σ by multiplying it by the distance r _j to the measurement point x _j , it is possible to increase the influence of the normal distribution of the measurement point x _j on the obtained occupancy value. Thereby, it is possible to compensate for the sparse distribution of measurement points x _j .

A laser reflected from the road surface is used to measure the measurement point x _j . However, when measuring a measurement point x _j at a certain height, the distribution value may be smaller than the correct distribution value depending on the reflection characteristics of the road surface. Therefore, by weighting the distribution value of covariance Σ by multiplying it by the height h _j of measurement point x _j , the inaccuracy of the normal distribution due to the height of measurement point x _j can be compensated for. I can do it.

Therefore, the occupancy value C _i at any point x _i can be expressed as follows. _{However ,} N (.; Moreover, x _i indicates the position of the grid in relative coordinates with respect to the vehicle on the xy plane. i is an index indicating the relative coordinate position.

Note that the accumulated distance measurement data 16 finally obtained by superimposing the distance measurement data is expressed as a continuous value of the occupancy value C _i . Further, for the occupancy value C _i in each grid, a binarization process may be performed to determine whether or not it is larger than a certain threshold value.

The representation format of the distance measurement data is not limited to the point cloud format in which measurement points that measure the distance from the distance measurement sensor 12 as described above are distributed in a three-dimensional space. The distance distribution from the distance measurement sensor 12 to the measurement point may be expressed as a distance image projected onto a two-dimensional plane, such as an output result from a stereo camera.

FIG. 4 is a diagram showing an example of the occupancy grid map 18 generated by the occupancy grid map generation unit 17. For example, a two-dimensional rectangular area with a predetermined length in the front and rear in the driving direction and left and right in the vehicle width direction is formed with the vehicle 200 at the center, and this rectangular area is divided into a grid to form a grid.

According to this figure, the rectangular area is a range of 10 m (+10 m to -10 m) front and back and 20 m (-20 m to +20 m) left and right with the vehicle 200 at the center. Further, the grid is assumed to be set in a square area of, for example, 10 to 20 cm in length and width. In this way, a grid divided by a predetermined grid width is configured in the rectangular area. In such a two-dimensional occupancy grid map 18, an occupancy value C _i indicating the height of an object existing in each grid is shown.

Below, roadside edge detection control using the occupancy value C _i will be explained.

FIG. 5 is a flowchart of roadside edge detection control by the driving support system 100. The roadside edge detection device 10 of the driving support system 100 is programmed to repeatedly perform a series of processes for roadside detection control shown in this figure at predetermined timing. In the following, for clarity of explanation, the blocks that implement predetermined functions in the roadside edge detection device 10 in FIG. 1 will be described as performing respective processes. Note that the roadside edge detection device 10 may perform a series of processes.

The road edge detection control is mainly shown collectively as step S10 and includes the generation control of the accumulated distance measurement data 16 consisting of steps S11 to S15, and the road edge detection control is collectively illustrated as step S16 and consists of steps S17 to S20. This includes two processes: detection control and detection control.

First, the generation process of the accumulated distance measurement data 16 in step S10 will be explained.

In step S11, the ranging data filter section 13 receives ranging data generated by the ranging sensor 12. As described above, the distance measurement data indicates the occupancy value C _i for the grid point x _i in a predetermined area around the vehicle 200 in the vehicle coordinate system.

In step S12, the ranging data filter section 13 correlates the ranging data acquired from the ranging sensor 12 with the position of the moving object detected by the moving object detecting section 11. Although moving objects exist on the road surface, they are different from roadside objects that define the roadside edge. Therefore, the distance measurement data filter section 13 deletes the distance measurement data of the position where the moving object detected by the moving object detection section 11 exists, based on the result of the association.

In step S13, the distance measurement data storage section 15 acquires the distance measurement data from the distance measurement data filter section 13 and the odometry information from the odometry measurement section 14. Then, the distance measurement data storage unit 15 converts the distance measurement data from the vehicle coordinate system to the absolute coordinate system using the odometry information.

As described above, since the distance measurement data is indicated with the position of the vehicle 200 as the zero point, it becomes a relative coordinate system in which the zero point changes depending on the acquisition timing of the distance measurement data. The odometry coordinates are, for example, an absolute coordinate system whose zero point is the position of the vehicle 200 at a specific timing such as when the vehicle starts traveling. Therefore, by using odometry information, it is possible to convert the acquired ranging data from a relative coordinate system to an absolute coordinate system.

In step S14, the distance measurement data storage section 15 further records the distance measurement data converted into odometry coordinates into the time buffer.

Here, the time buffer is a memory area that records a plurality of elements of a predetermined length, and is generally referred to as an array. In this example, the time buffer is configured to be able to record a predetermined number of ranging data. Further, in this time buffer, distance measurement data acquired at a certain timing is recorded as one element constituting the array of the time buffer, and distance measurement data for a predetermined period is recorded in the entire time buffer.

More specifically, when the ranging data storage unit 15 receives new ranging data from the ranging data filter unit 13, it adds the new ranging data as the final element of the time buffer, and also adds the new ranging data as the final element of the time buffer. The oldest ranging data, which is the first element of , is deleted. By doing so, the time buffer stores the distance measurement data of a predetermined length of time from the present to the past since the number of stored distance measurement data is held constant. For example, the time buffer can store 100 pieces (100 steps) of distance measurement data. A memory structure in which new elements are added and the oldest elements are deleted is called a ring buffer.

In step S15, the distance measurement data storage unit 15 generates accumulated distance measurement data 16 by superimposing the distance measurement data in the absolute coordinate system recorded in each element of the array in the time buffer.

Here, since the ranging data stored in the time buffer has been converted into odometry coordinates in step S13, their zero points are equal. Therefore, by superimposing the distance measurement data recorded in the time buffer, it is possible to generate the accumulated distance measurement data 16 indicating the occupancy value in the three-dimensional space of the absolute coordinate system based on the same zero point. In this way, high-density accumulated distance measurement data 16 generated from a plurality of distance measurement data can be generated.

Next, the roadside edge detection control in step S16 will be explained.

In step S17, the occupancy grid map generation section 17 receives input of accumulated distance measurement data 16 from the distance measurement data storage section 15. The occupancy grid map generation unit 17 generates a variance value C _i by superimposing the distribution values in the height direction on the accumulated distance measurement data 16 indicating a normal distribution of the covariance Σ in a three-dimensional space so as to project it onto the road surface. By calculating, a grid map is generated.

In this grid map, the occupancy value C _i at each grid point x _i formed by dividing the road surface into a grid is shown. Since the occupancy value C _i shown in this grid map is the sum of the normal distribution of the covariance Σ in the height direction, it corresponds to the height of an object that can exist at the grid point x _i . Therefore, the grid map shows the height of the road surface according to the undulations of the surface of the road surface and structures on the road surface. Then, the occupancy grid map generation unit 17 cuts out a portion as shown in FIG. 4 from the generated grid map to generate an occupancy grid map 18.

In step S18, the occupancy grid map adding unit 19 calculates the sum of the occupancy values C _i in each of the grid rows in the traveling direction in the occupancy grid map 18 as the road surface height sum v. Thereby, the distribution of the road surface height sum v corresponding to the position from the vehicle 200 in the vehicle width direction can be determined.

In step S19, the discontinuous boundary detection unit 20 receives an input of the distribution of the total road surface height v generated in step S18. The discontinuous boundary detection unit 20 determines whether the total road surface height v is larger than the threshold value v _th from the vehicle 200 toward both left and right sides in the vehicle width direction, and determines whether the total road surface height v is larger than the threshold value v _th . The position in the vehicle width direction where it is determined that the distance is also large is determined to be a discontinuous boundary. Note that a plurality of discontinuous boundaries may be detected.

In step S20, the roadside edge determination unit 21 selects one by one the closest discontinuous boundaries from the vehicle 200 toward both left and right sides in the vehicle width direction, from among the plurality of discontinuous boundaries detected in step S19. , it is determined that the location of those discontinuous boundaries is the roadside edge. The roadside edge determination unit 21 then outputs the determined position of the roadside edge in the vehicle width direction to the output unit 22 .

Through such processing, roadside edge detection control is performed. The roadside edge detection control will be explained below using a specific example with reference to FIGS. 6 to 8.

FIG. 6 is a diagram showing measurement points by the ranging sensor 12 from above.

This figure shows a road surface 61 seen from above, and a vehicle 200 traveling toward the top of the drawing is shown in the center. On the road surface 61, curb stones 62 are provided on the left and right sides of the vehicle 200, respectively. Since the vehicle 200 runs between the two curbs 62, the direction in which the curbs 62 extend and the direction in which the vehicle 200 runs substantially coincide. Note that the positions of the side surfaces of these curbstones 62 on the vehicle 200 side are detected as roadside edges.

Further, in this figure, the distance measurement points by the distance measurement sensor 12 are indicated by circles. Since the radar etc. output from the distance measurement sensor 12 are blocked at the curb 62, distance measurement data is acquired in an area 63 extending along the curb 62 with a predetermined width on the opposite side of the vehicle 200 to the curb 62. Can not. On the other hand, in a region 64 near the wall surface of the curb 62 on the side of the vehicle 200, distance measurement data is acquired at a plurality of measurement points. Further, on the road surface 61 near the vehicle 200, there is an area 65 that becomes a blind spot of the vehicle 200 and cannot acquire distance measurement data.

FIG. 7 is an explanatory diagram of the processing of steps S18 to S20 in the roadside edge detection control shown in FIG.

FIG. 7(a) shows an occupancy grid map 18 corresponding to the area shown in FIG. 6. In this occupancy grid map 18, regions where the occupancy value C _i is large and the height of objects existing in the grid are high are marked with dark hatching. On the other hand, areas where the occupancy value C _i is small and the height of objects existing in the grid are low are lightly hatched.

A road surface 61 serving as a reference in the height direction is hatched with medium shading. On the other hand, areas where tall curbstones 62 exist are marked with dark hatching. Note that the area 63 that is in the shadow of the curb 62 and the area 65 that is a blind spot of the vehicle 200 are shown in white because distance measurement data cannot be obtained.

FIG. 7(b) shows the distribution of the total road surface height v depending on the position in the vehicle width direction.

The road surface height sum v is calculated in step S18, and is the sum of the occupancy values C _i of the grid rows in the traveling direction of the vehicle 200 in the occupancy grid map 18. Therefore, in the region where the curbstone 62 exists in the vehicle width direction, the total road surface height v is larger than in the region where only the road surface 61 exists. On the other hand, in the area 63 where distance measurement data cannot be acquired, the total road surface height v becomes zero. Further, at a position in the vehicle width direction where a region 65 serving as a blind spot of the vehicle 200 exists, the total road surface height v becomes relatively low.

In step S19, the discontinuous boundary detection unit 20 sequentially checks whether the total road surface height v is larger than the threshold value v _th from the center of the vehicle 200 toward both sides (left and right) in the vehicle width direction. Then, the position in the vehicle width direction where the total road surface height v is determined to be larger than the threshold value v _th is detected as a discontinuous boundary 71 . In this example, two discontinuous boundaries 71 are detected, but three or more may be detected.

Note that in the ignored region 72 where the vehicle 200 is present, it is not necessary to compare the total road surface height v with the threshold value v _th . Due to the shape of the vehicle 200 and the location of the distance measurement sensor 12, the ignored region 72 becomes a blind spot for the distance measurement sensor 12, so there is no need to detect the discontinuous boundary 71 in this region.

Note that the threshold value v _th is determined as follows.

First, the average value of the distribution of the total road surface height v at all positions in the vehicle width direction can be calculated as v _mean as shown in the following equation using the occupancy value C _i obtained by equation (1). However, i (1≦i≦N) in the following equation is an index indicating the position of the grid point x _i in the vehicle width direction, and is the same as in equation (1).

The threshold value v _th can be calculated using the average value v _mean , the maximum value v _max in the total road surface height v, and the coefficient α (0≦α≦1) as shown in the following equation.

In this way, in order to determine the threshold value _{v th} used for detecting the discontinuous boundary 71 based on the average value v _mean of the road surface height total sum v, a threshold value v _th that corresponds to a change in the distribution of the road surface height total v is used. Specifically, if the total road surface height v is large overall, the threshold value v _th becomes large, and if the total road surface height v is small overall, the threshold value v _th becomes small. In this way, by dynamically changing the threshold value v _th , the detection sensitivity of the discontinuous boundary 71 can be changed, so that changes in the detection accuracy of the discontinuous boundary 71 due to changes in the acquisition status of ranging data can be prevented. This makes it possible to accurately determine the roadside edge.

Further, the average value v _mean of the total road surface height v is used as a reference, and a threshold value v _th is determined by a value of a certain ratio according to the coefficient α for the difference between the average value v _mean and the maximum value v _max . . Therefore, when the difference between the average value v _mean and the maximum value v _max is small, the threshold value v _th becomes closer to the average value v _mean , and when the difference between the average value v _mean and the maximum value v _max is large, the threshold value v _th becomes a value closer to the maximum value v _max . This allows the detection sensitivity of the discontinuous boundary 71 to be dynamically changed according to the maximum value v _max in addition to changes in the distribution of the total road surface height v, so that the roadside edge can be determined with high accuracy. It becomes possible to do so.

Note that equation (3) is a weighted average of the average value v _mean and the maximum value v _max , and can be expressed as the following equation.

FIG. 8 is an explanatory diagram of a more detailed example of roadside edge detection control.

FIG. 8(a) shows a grid map obtained by superimposing accumulated distance measurement data 16 in the height direction, as seen from above and behind the vehicle 200, and an occupancy grid map 18 that is a part of the grid map. ing. In addition to the curb 62, there are white lines 81 and the like on the road surface 61, and these objects are shown on the grid map.

FIG. 8(b) shows the total road surface height v calculated using the occupancy grid map 18 of FIG. 8(a). In this figure, an ignored region 72 is shown as in FIG. 7(b), and locations where the total road surface height v is greater than the threshold value v _th are detected as discontinuous boundaries 71A to 71D. Then, among these discontinuous boundaries 71A to 71D, it is determined that the curb 62 exists at the discontinuous boundaries 71B and 71C, which are closer to the vehicle 200, and are detected as the roadside edge positions.

In this way, the occupancy value C _i indicating the height of the roadside edge such as the curb 62 or the road surface 61 in a predetermined range around the vehicle 200 is calculated using the distance measurement data, and the occupancy value C _i is calculated along the traveling direction. By calculating the sum, the road surface height sum v is calculated. Note that the total road surface height v is determined as a distribution according to the position of the vehicle 200 in the vehicle width direction. Then, by comparing the total road surface height v and the threshold value v _th , discontinuity in the distribution of the total road surface height v is evaluated, and a roadside edge is detected. Because of this configuration, even if some of the measurement points of the distance measurement data cannot be detected, the roadside edge can be detected using information from the measurement points before and after it, improving detection accuracy. can be achieved.

According to the first embodiment, the following effects can be obtained.

According to the roadside edge detection method of the first embodiment, the heights of the road surface 61 and the curbstone 62 are detected in a predetermined range around the vehicle 200, and the occupancy grid map 18 shows the occupancy value C _i in each grid. By calculating the sum of the road surface height detection step (S17) that generates the occupancy value C _i shown in the occupancy grid map 18 along the traveling direction of the vehicle 200, the road surface height corresponding to the position in the vehicle width direction is calculated. a summation calculation step (S18) for calculating the distribution of the summation v; a roadside end determination step (S20) for determining that the position in the vehicle width direction corresponding to the road surface height summation v exceeding a threshold value _vth is the roadside end; Equipped with

With this configuration, in a predetermined range around the vehicle 200, discontinuities in the total road surface height v are evaluated in correspondence with positions in the vehicle width direction, and the position of the roadside edge is determined. Therefore, even if some of the distance measurement data acquired in a predetermined range is missing or erroneous and cannot be detected, the roadside edge can be detected using the distance measurement data before and after in the traveling direction. In this way, it is possible to compensate for the undetected ranging data, so it is possible to improve the detection accuracy of the roadside edge.

According to the roadside edge detection method of the first embodiment, in the roadside edge determination step (S20), the total road surface height v and the threshold value v _th are compared to the outside toward the left and right in the vehicle width direction starting from the position of the vehicle 200. The position in the vehicle width direction where the total road surface height v is determined to exceed the threshold value v _th is determined to be the roadside edge.

With this configuration, the detected position of the roadside edge can be determined to be the inner wall of the roadside object such as the curb 62 on the vehicle 200 side. Can be detected well.

According to the roadside edge detection method of the first embodiment, in the roadside edge determination step (S20), if there are a plurality of discontinuous boundaries 71 such that the total road surface height v exceeds the threshold value _vth , the vehicle The position of the discontinuous boundary 71 close to 200 is determined to be the roadside end.

The roadside edge that is the side edge of the driving lane of the vehicle 200 is the discontinuous boundary 71 that is observed as the innermost discontinuous boundary among the plurality of discontinuous boundaries 71 when viewed from the vehicle 200 . Therefore, when a plurality of discontinuous boundaries 71 are detected, the discontinuous boundary 71 closest to the vehicle 200 in the vehicle width direction is determined to be the roadside edge. Thereby, the position of the roadside edge that defines the travelable range of vehicle 200 can be detected with high accuracy.

According to the roadside edge detection method of the first embodiment, in the roadside edge determination step (S20), the larger the average value v _mean of the road surface height summation v according to the position in the vehicle width direction, the more the road surface height summation v is compared. The threshold value v _th is set large.

With such a configuration, the threshold value v _th changes in accordance with a change in the distribution of the total road surface height v. That is, when the distribution has a large road surface height sum v as a whole, the threshold value v _th becomes large, and when the road surface height sum v has a small distribution, the threshold value v _th becomes small. Therefore, the sensitivity for detecting the discontinuous boundary 71 that is the roadside edge can be dynamically changed depending on the distribution of the road surface height sum v, so that when the measurement accuracy of distance measurement data changes due to weather etc. However, the influence caused by changes in the distribution of distance measurement data is suppressed, and the roadside edge can be detected with high accuracy.

According to the roadside edge detection method of the first embodiment, the threshold value v _th is a weighted average of the average value v _mean and the maximum value v _max at a predetermined ratio.

With such a configuration, the threshold value v _th can be further appropriately set. For example, when the difference between the average value v _mean and the maximum value v _max is small, the threshold value v _th is set to a value close to the average value v _mean , thereby making it easier to detect the roadside edge. On the other hand, when the difference between the average value v _mean and the maximum value v _max is large, false detection of the roadside edge can be suppressed by setting the threshold value v _th to a value closer to the maximum value v _max . In this way, by using the maximum value v _max in addition to the average value v _mean , it is possible to dynamically change the detection sensitivity of the discontinuous boundary 71, thereby suppressing the influence of changes in the distribution of distance measurement data. This makes it possible to detect the roadside edge with high accuracy.

(Second embodiment)
In the first embodiment, when a plurality of discontinuous boundaries 71 are detected by the discontinuous boundary detection unit 20, the roadside edge determination unit 21 selects a discontinuous boundary 71 near the vehicle from among these discontinuous boundaries 71. It was detected that the roadside edge was located at the position in the vehicle width direction. In the second embodiment, another roadside edge detection method will be described when a plurality of discontinuous boundaries 71 are detected.

FIG. 9 is a schematic configuration diagram of a driving support system 100 that executes the roadside edge detection method according to the second embodiment.

According to this figure, compared to the driving support system 100 of the first embodiment shown in FIG. In addition, input from the self-position estimating unit 30 and map data 31 is accepted.

The self-position estimating unit 30 estimates the current position of the vehicle using GPS or the like, and outputs the estimated position of the vehicle. Furthermore, the map data 31 is a high-definition map (HD map) that shows position information of the curb 62 and the like. The roadside edge determining unit 21 estimates the approximate position of the curb by associating the current position of the vehicle input from the self-position estimating unit 30 with the high-definition map of the map data 31, and estimates the approximate position of the curb closest to the estimated position of the curb. A discontinuous boundary is detected as a roadside edge.

In this way, if the extending direction of the curb can be obtained from map information, etc., the sum of the occupancy values C _i may be calculated along the extending direction of the curb to calculate the total road surface height v. . That is, the calculation of the total road surface height v is not limited to the sum of the occupancy values C _i along the traveling direction in the relative coordinate system as seen from the vehicle as in the first embodiment, but can be determined from the position information of the curb in the absolute coordinate system. It may be the sum total of occupancy values C _i along the traveling direction indicated by the possible driving lanes.

According to the roadside edge detection method of the second embodiment, in the position information acquisition step, the road information including the position information of the roadside edge and the vehicle position information are used to detect the outer side of the vehicle in the vehicle width direction. Get the location of the roadside edge in . If multiple discontinuous boundaries are detected in the roadside edge determination step, a discontinuous boundary close to the roadside edge position acquired in the position information acquisition step is selected from among those discontinuous boundaries and detected as the roadside edge. do. Through such processing, it is possible to improve the detection accuracy of the roadside edge.

(Third embodiment)
In the first and second embodiments, the roadside edge detection control is performed when the travel lane of the vehicle 200 is a straight line, but the present invention is not limited to this. In the third embodiment, an example will be described in which roadside edge detection control is performed when the travel lane is curved.

FIG. 10 is a schematic configuration diagram of a driving support system 100 that executes the roadside edge detection method according to the third embodiment.

According to this figure, compared to the driving support system 100 of the first embodiment shown in FIG. Then, input from the road curvature detection section 40 is further accepted.

The road curvature detection unit 40 detects the radius of curvature r of the lane in which the vehicle 200 is traveling using input from the self-position estimation unit 30, map data 31, etc. (not shown in FIG. 10) shown in the second embodiment. do. Further, the road curvature detection unit 40 may obtain the radius of curvature r using information such as the white line recognition result, the yaw rate, and the current steering angle.

The occupancy grid map generation unit 17 uses the grid map generated from the accumulated distance measurement data 16 to generate an occupancy grid map 18 whose length in the traveling direction is l. Note that the length l of the occupancy grid map 18 in the traveling direction of the vehicle is determined according to the following equation.

As shown in equation (5), the length l of the occupancy grid map 18 in the traveling direction is determined by multiplying the radius of curvature r of the traveling lane by a predetermined proportionality constant ε. Therefore, when the curvature of the driving lane is large and the radius of curvature r is small, the length l of the occupancy grid map 18 in the driving direction becomes short, and when the curvature is small and the radius of curvature r is large, the length l of the driving lane becomes long. Note that l _max and l _min respectively indicate the maximum length and minimum length of the occupancy grid map 18 necessary for the discontinuous boundary detection unit 20 to detect the roadside edge.

FIG. 11 is a diagram showing a comparison between a case where the driving lane is straight and a case where the driving lane is curved. FIG. 11(a) shows the case where the running path is straight, and FIG. 11(b) shows the case where the running path is curved. In both cases, the vehicle 200 and the curb 62 serving as the roadside edge of the driving lane are shown together with the occupancy grid map 18. Note that in FIG. 11(b), it is assumed that the radius of curvature of the driving lane is r.

When the driving lane is curved as shown in FIG. 11(b), the curbstone 62 has a wider area in the vehicle width direction than when the driving lane is straight as shown in FIG. 11(a). It will be distributed in Therefore, the peak of the total road surface height v, which is the total sum of the occupancy values C _i along the traveling direction, becomes low, and there is a possibility that the accuracy of detecting the roadside edge by comparison with the threshold value v _th decreases. Further, as the curvature of the driving lane increases and the radius of curvature r decreases, the curb stones 62 become wider in the vehicle width direction, further reducing the accuracy of detecting the roadside edge.

Therefore, as shown in Equation (4), as the curvature increases, that is, as the radius of curvature r decreases, the length l of the vehicle 200 in the traveling direction of the occupancy grid map 18 is decreased, so that the length l of the curbstone 62 in the vehicle width direction is decreased. The distribution of can be made narrower. As a result, it is possible to suppress a decrease in the peak of the total road surface height v, and it is possible to detect the roadside edge with high accuracy.

As described above, according to the roadside edge detection method of the third embodiment, the radius of curvature r of the travel lane of the vehicle 200 is acquired in the curvature acquisition step, and the smaller the radius of curvature r is, the more the occupancy grid map is determined in the road surface height detection step. The running direction length l of the 18 vehicles 200 is set short.

Here, when the driving lane is curved, the portions of the occupancy grid map 18 where the curbstones 62 where the occupancy value C _i becomes large are arranged are distributed over a wide area in the vehicle width direction. Therefore, the peak of the total road surface height v becomes small, and there is a possibility that the total road surface height v exceeding the threshold value v _th cannot be appropriately detected.

However, by shortening the length l of the occupancy grid map 18 in the traveling direction of the vehicle 200, the curb stones 62 are distributed narrowly in the vehicle width direction. As a result, the peak of the total road surface height v is suppressed from becoming small, and the total road surface height v exceeding the threshold value v _th can be appropriately detected, so that the detection accuracy of the roadside edge is improved. Furthermore, the larger the curvature, that is, the smaller the radius of curvature r, the smaller the length l of the occupancy grid map 18 in the traveling direction of the vehicle 200, thereby further improving the roadside edge detection accuracy.

(Modified example)
In the above-described embodiment, an example is shown in which a roadside edge is detected in the occupancy grid map 18 having a length l in the traveling direction according to a radius of curvature r, but the present invention is not limited to this. It is also possible to divide the occupancy grid map 18 into a plurality of regions, detect the roadside edges in each divided region, and compare these roadside edges.

FIG. 12 is a diagram showing an example of division of the grid map in this modification.

In this example, the occupancy grid map generation unit 17 further divides the occupancy grid map 18 generated in the first embodiment into D divided areas 18-i (0≦i≦D), and divides them into D divided areas 18-i (0≦i≦D). The roadside edge is detected in each of the divided regions 18-i. The divided areas 18-i are arranged side by side from the front to the rear in the running direction within a range that does not exceed the occupancy grid map 18, and the divided area 18-i is the i-th one arranged from the front. shall be shown.

In this figure, l _max indicates the maximum possible length of the occupancy grid map 18 in the traveling direction. l is a value determined using equation (4), and indicates the length of one divided region 18-i in the traveling direction of the vehicle 200. In such a case, the number of divisions D of the occupancy grid map 18 can be expressed by the following equation.

However, [x] is a Gauss symbol and indicates the largest integer not exceeding x.

By dividing the occupied grid map 18 with the maximum length l _max by the number of divisions D, D divided regions 18-i are provided. The lengths of these divided regions 18-i in the front-back direction are shown as l _i , and l _i is all l regardless of i.

The distance x _i from the center of the vehicle 200 to the front end of the divided region 18-i can be expressed as follows.

The occupancy grid map generation unit 17 generates the occupancy grid map 18 by associating the position x _i of the divided area 18-i in the traveling direction with the position of the roadside edge in the vehicle width direction detected in the divided area 18-i. The location of the roadside edge can be detected.

Note that the occupancy grid map 18 may always be divided by a constant number of divisions D, regardless of the radius of curvature r of the travel route, and the plurality of divided regions 18-i may be evaluated individually. By doing so, conversely, the curvature of the driving lane can be estimated from the detected position of the roadside edge in the vehicle width direction.

As described above, according to the roadside edge detection method of the modified example, the roadside edge in the vehicle width direction is determined for each of the divided areas 18-i that are formed by dividing the occupancy grid map 18 into D pieces along the traveling direction. Detect the position of. Then, the position of the roadside end in the divided area 18-i in the vehicle width direction is detected in association with the position of the divided area 18-i in the traveling direction. By doing so, even if the driving lane is curved, the position of the roadside edge in the occupancy grid map 18 can be appropriately determined in the driving direction.

Although the embodiments of the present invention have been described above, the above embodiments merely show a part of the application examples of the present invention, and are not intended to limit the technical scope of the present invention to the specific configurations of the above embodiments. do not have. Furthermore, the above embodiments can be combined as appropriate.

10 roadside edge detection device 12 distance sensor 14 odometry measurement section 15 distance measurement data storage section 16 accumulated distance measurement data 17 occupancy grid map generation section 18 occupancy grid map 18-i: divided area 19 occupancy grid map addition section 20 discontinuous boundary Detection unit 21 Roadside edge determination unit 30 Self-position estimation unit 31 Map data 40 Road curvature detection unit 61 Road surface 62 Curb 71 Discontinuous boundary 72 Ignored area 100 Driving support system 200 Vehicle

Claims (9)

A roadside edge detection method for detecting a roadside edge on which a roadside object is provided on a road surface on which a vehicle runs using a roadside edge detection device , the method comprising:
a road surface height detection step of detecting the height of the road surface in a predetermined range around the vehicle;
a total sum calculation step of calculating the sum of the heights of the road surface in the predetermined range detected in the road surface height detection step along the traveling direction of the vehicle according to the position in the vehicle width direction with respect to the vehicle; and,
A roadside edge detection method, comprising: a roadside edge determination step of determining that a position in the vehicle width direction where the summation calculated in the summation calculation step exceeds a threshold is the roadside end. The roadside edge detection method according to claim 1,
In the roadside edge determination step, the positions where the sum exceeds the threshold are sequentially detected in the vehicle width direction from the position where the vehicle is present toward the side, and when the sum exceeds the threshold, the position is detected first. A roadside edge detection method, wherein the detected position in the vehicle width direction is determined to be the roadside edge. The roadside edge detection method according to claim 1 or 2,
In the roadside edge determining step, if a plurality of positions in the vehicle width direction are detected where the sum total exceeds a threshold value, the position closest to the vehicle in the vehicle width direction is determined to be the roadside edge. , Roadside edge detection method. The roadside edge detection method according to claim 1 or 2,
Furthermore, a position information acquisition step of acquiring position information of the roadside end based on road information including the position information of the vehicle and the position information of the roadside end,
In the roadside edge determination step, if a plurality of positions in the vehicle width direction are detected where the sum exceeds the threshold, the position closest to the location indicated by the roadside edge position information acquired in the position information acquisition step is detected. A roadside edge detection method that determines that the position is the roadside edge. The roadside edge detection method according to any one of claims 1 to 4,
A roadside edge detection method, wherein in the roadside edge determination step, the larger the average value of the sum is, the larger the threshold value is. The roadside edge detection method according to claim 5,
In the roadside edge determination step, the threshold value is a weighted average of the average value and the maximum value of the total sum at a predetermined ratio. The roadside edge detection method according to any one of claims 1 to 6,
Furthermore, a curvature acquisition step of acquiring the curvature of the driving lane of the vehicle,
In the road surface height detection step, the greater the curvature of the travel lane, the shorter the length of the predetermined range in the direction in which the vehicle travels. The roadside edge detection method according to any one of claims 1 to 6,
In the roadside edge determination step,
The predetermined range is divided into a plurality of regions along the direction in which the vehicle travels, and for each of the divided regions arranged in parallel in the direction in which the vehicle travels, the divided region is divided into a plurality of regions along the direction in which the vehicle travels. determining the position of the roadside end in the vehicle width direction in association with the position up to;
A roadside edge detection method, for each of the divided regions, determining that the roadside edge is located at the determined position of the roadside edge in the vehicle width direction. A roadside edge detection device that detects a roadside edge on which a roadside object is provided on a road surface on which a vehicle runs,
a road surface height detection unit that detects the height of the road surface in a predetermined range around the vehicle;
a total sum calculation unit that calculates the sum of the heights of the road surface in the predetermined range detected by the road surface height detection unit along the traveling direction of the vehicle according to the position in the vehicle width direction with respect to the vehicle; and,
A roadside edge detection device, comprising: a roadside edge determination unit that determines that a position in the vehicle width direction where the summation calculated by the summation calculating unit exceeds a threshold is the roadside edge.

Images