Road surface condition detection device, three-dimensional object detection device, vehicle, and three-dimensional object detection method

Abstract

To provide a technology for accurately detecting the condition of a road surface extending forward of a vehicle, a technology for accurately detecting a three-dimensional object existing on the road surface in consideration of the condition of the road surface, and a vehicle equipped with the three-dimensional object detection technology.SOLUTION: A distance sensor attached to a vehicle traveling on a road surface receives the electromagnetic waves reflected in front of the vehicle while scanning the electromagnetic waves emitted from above and diagonally toward the road surface in a direction perpendicular to the traveling direction of the vehicle, thereby acquiring point cloud data indicating the distance from the vehicle to the reflection point and the direction of the reflection point relative to the vehicle. When the point cloud corresponding to the multiple reflection points can draw an arc-shaped or elliptical-arc-shaped road surface identification trajectory based on the point cloud data, it is determined that there is no solid object on the road surface.SELECTED DRAWING: Figure 5C

Description

Technical Field

[0001] This invention relates to a detection technique for detecting the condition of a road surface on which a vehicle travels and three-dimensional objects present on the road surface, and a vehicle equipped with the same.

Background Art

[0002] In order to detect three-dimensional objects present on a road surface, for example, in Patent Document 1, a grid map in which three-dimensional distance data point clouds measured by a lidar are accumulated is generated. Then, the road surface and three-dimensional objects are determined from the grid map.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the above-described conventional technology, detection of three-dimensional objects is performed on the premise that the road surface extending in the traveling direction of the vehicle is at the same height as the current traveling position of the vehicle, that is, the road surface is flat. More specifically, based on the three-dimensional coordinates of points obtained from a lidar fixed to the vehicle, from known road surface coordinates, with a threshold value in a certain height direction, it is determined whether the point is a point of a three-dimensional object. Therefore, for example, when the front road surface is an inclined surface inclined upward, this upward inclined surface exceeds the threshold value. As a result, the upward inclined surface is mistaken for a three-dimensional object. Thus, in the conventional technology, since three-dimensional object detection is performed without accurately grasping the condition of the front road surface, it has been difficult to accurately detect three-dimensional objects.

[0005] This invention has been made in view of the above-mentioned problems, and aims to provide a technology for accurately detecting the condition of the road surface extending in front of a vehicle, a technology for accurately detecting three-dimensional objects present on the road surface while taking the road surface conditions into consideration, and a vehicle equipped with said three-dimensional object detection technology. [Means for solving the problem]

[0006] A first aspect of the present invention is a road surface condition detection device, characterized by comprising: a distance sensor attached to a vehicle traveling on a road surface, which receives electromagnetic waves reflected in front of the vehicle while scanning the road surface in a direction perpendicular to the vehicle's direction of travel with respect to electromagnetic waves irradiated from diagonally above the road surface, thereby acquiring point cloud data indicating the distance from the vehicle to the reflection point and the direction of the reflection point relative to the vehicle for a plurality of reflection points that have reflected electromagnetic waves; a point cloud trajectory determination unit that determines whether the point cloud corresponding to the plurality of reflection points can draw an arc-shaped or elliptical arc-shaped road surface discrimination trajectory based on the point cloud data; and a road surface determination unit that determines that there are no three-dimensional objects on the road surface when the point cloud trajectory determination unit determines that the point cloud can draw a road surface discrimination trajectory.

[0007] Furthermore, a second aspect of the present invention is a three-dimensional object detection device, which is mounted on a vehicle traveling on a road surface and receives electromagnetic waves reflected in front of the vehicle while scanning the road surface in a direction perpendicular to the direction of travel of the vehicle with respect to electromagnetic waves irradiated from diagonally above the road surface, thereby acquiring point cloud data indicating the distance from the vehicle to the reflection point and the direction of the reflection point relative to the vehicle for a plurality of reflection points that have reflected electromagnetic waves; a deviation point identification unit that compares the point cloud trajectory drawn by the point cloud corresponding to the plurality of reflection points based on the point cloud data with an arc-shaped or elliptical arc-shaped road surface discrimination trajectory, and identifies deviation points among the points constituting the point cloud where the point cloud trajectory deviates from the road surface discrimination trajectory; and a three-dimensional object identification unit that identifies the presence of a three-dimensional object on the road surface based on the deviation points identified by the deviation point identification unit.

[0008] Furthermore, a third aspect of the present invention is a vehicle equipped with the above-mentioned three-dimensional object detection device, characterized in that its driving on the road surface is controlled based on the presence of a three-dimensional object identified by the three-dimensional object detection device.

[0009] Furthermore, a fourth aspect of the present invention is a method for detecting three-dimensional objects, characterized by comprising the steps of: acquiring point cloud data indicating the distance from the vehicle to the reflection point and the direction of the reflection point relative to the vehicle for a plurality of reflection points that reflected electromagnetic waves, by scanning the road surface with respect to the road surface in a direction perpendicular to the direction of travel of the vehicle, and receiving the electromagnetic waves reflected in front of the vehicle, which are attached to a vehicle traveling on a road surface and irradiated from diagonally above the road surface; comparing the point cloud trajectory drawn by the point cloud corresponding to the plurality of reflection points based on the point cloud data with an arc-shaped or elliptical arc-shaped road surface discrimination trajectory; identifying deviation points among the points constituting the point cloud where the point cloud trajectory deviates from the road surface discrimination trajectory; and identifying the presence of a three-dimensional object on the road surface based on the deviation points.

[0010] In this invention, a distance sensor attached to a vehicle traveling on a road surface receives electromagnetic waves reflected in front of the vehicle while scanning electromagnetic waves irradiated from diagonally above the road surface in a direction perpendicular to the vehicle's direction of travel. This acquires point cloud data indicating the distance from the vehicle to the reflection point and the direction of the reflection point relative to the vehicle. When the point cloud data, based on the point cloud data, can trace an arc-shaped or elliptical road surface discrimination trajectory corresponding to multiple reflection points, it is determined that no three-dimensional object exists on the road surface. Therefore, regardless of whether the road surface in front of the vehicle is flat or sloped, and even when the vehicle is stationary, it is possible to accurately detect whether or not a three-dimensional object exists on the road surface. In this specification, "three-dimensional object" includes objects that exist above the road surface, and conversely, depressions, ditches, etc., that exist below the road surface.

[0011] In this case, the system may be configured to determine that the road surface is flat when it is determined that the point cloud can trace an arc-shaped road surface discrimination trajectory, and to determine that the road surface is inclined in the direction of travel when it is determined that it can trace an elliptical arc-shaped road surface discrimination trajectory. In this case, not only can it accurately detect whether there are any three-dimensional objects on the road surface, but it can also accurately detect the road surface condition, such as whether the road surface is flat or inclined.

[0012] Furthermore, if the road surface is an inclined surface that slopes in the direction of travel, the inclination angle of the inclined surface may be further determined based on the curvature of the elliptical arc-shaped road surface discrimination trajectory, thereby enabling more detailed detection of the road surface conditions.

[0013] Furthermore, when determining whether a point cloud can trace a road surface discrimination trajectory, it may be further considered whether a certain percentage or more of the points constituting the point cloud are located on the road surface discrimination trajectory. In other words, if a certain percentage or more of the points are located on the road surface discrimination trajectory, it may be determined that the point cloud can trace a road surface discrimination trajectory, while if a certain percentage or more of the points constituting the point cloud are located on the road surface discrimination trajectory, it may be determined that the point cloud cannot trace a road surface discrimination trajectory and that a three-dimensional object exists on the road surface. This makes it possible to stably detect road surface conditions and the presence or absence of three-dimensional objects by suppressing the effects of disturbances and noise on the point cloud data.

[0014] Furthermore, if the point cloud trajectory does not contain any discontinuities, it may be determined that the point cloud can draw a road surface discrimination trajectory. Conversely, if the point cloud trajectory contains discontinuities, it may be determined that the point cloud cannot draw a road surface discrimination trajectory and that a three-dimensional object exists on the road surface. This allows for accurate detection of not only the road surface conditions but also the presence or absence of three-dimensional objects.

[0015] For the distance sensor, either a two-dimensional sensor that scans electromagnetic waves at a single irradiation angle to acquire point cloud data or a three-dimensional sensor configured to scan at multiple irradiation angles may be used. However, by using a three-dimensional sensor, it is possible to determine whether the point cloud can trace a road surface discrimination trajectory for each different irradiation angle, thereby enabling more accurate detection of road surface conditions and the presence or absence of three-dimensional objects. More specifically, the system may be configured so that when the point cloud trajectory determination unit determines that the point cloud can trace a road surface discrimination trajectory for all of the multiple irradiation angles, it is determined that there are no three-dimensional objects on the road surface, while when the point cloud trajectory determination unit determines that the point cloud cannot trace a road surface discrimination trajectory for at least one of the multiple irradiation angles, it is determined that there are three-dimensional objects on the road surface. This makes it possible to detect three-dimensional objects of various sizes in the vertical direction, expanding the detectable range.

[0016] Furthermore, when the object identification unit identifies a three-dimensional object based on deviation points, the object identification unit may be configured to include a clustering unit that clusters the deviation points identified by the deviation point identification unit and classifies them into one or more deviation point groups, and an individual identification unit that identifies whether a three-dimensional object exists on the road surface for each deviation point group. This allows for accurate detection of three-dimensional objects not only when there is one three-dimensional object on the road surface, but also when there are multiple objects. It is desirable that the clustering unit be configured to classify multiple deviation points that are a certain number or more and adjacent to each other within a threshold distance into a deviation point group. In other words, by adopting the above classification method, it is possible to accurately acquire deviation point groups corresponding to three-dimensional objects while suppressing the effects of noise, and further improve the accuracy of three-dimensional object detection. [Effects of the Invention]

[0017] As described above, according to the present invention, the road surface conditions in front of the vehicle and three-dimensional objects on the road surface can be accurately detected. [Brief explanation of the drawing]

[0018] [Figure 1]It is a figure showing a golf cart which is an example of a vehicle equipped with a first embodiment of a road surface condition detection device according to the present invention. [Figure 2] It is a block diagram showing an electrical configuration for driving and controlling the golf cart shown in FIG. 1. [Figure 3] It is a functional block diagram showing the configuration of a road surface condition detection device incorporated in the golf cart shown in FIG. 1. [Figure 4] It is a schematic diagram showing a representative example of the driving environment of a golf cart traveling on a road surface. [Figure 5A] It is a figure in which the position coordinates of reflection points calculated from point cloud data acquired by a distance sensor in a driving environment (A) are plotted on a two-dimensional plane for each irradiation angle. [Figure 5B] It is a figure in which the position coordinates of reflection points calculated from point cloud data acquired by a distance sensor in a driving environment (B) are plotted on a two-dimensional plane for each irradiation angle. <00​​​​​​​​​​​​​​​​​​​​​Figure 1 shows a golf cart, which is an example of a vehicle equipped with a first embodiment of the road surface condition detection device according to the present invention. In the following description, front and rear, left and right, and up and down refer to the front and rear, left and right, and up and down relative to the state in which an occupant is seated in the front seat portion 18 of the golf cart 10 facing the steering wheel 22.

[0020] As shown in Figure 1, the golf cart 10 has a frame section 12. On this frame section 12, a pair of left and right front wheels 14 are rotatably supported at the front, while a pair of left and right rear wheels 16 are rotatably supported at the rear. In the golf cart 10, the front seat section 18 and the rear seat section 20 for the occupants are separated front and rear and supported on the frame section 12 via connecting members (not shown). A steering wheel 22 is provided in front of the front seat section 18. A pair of left and right front pillars 24 are provided in front of the steering wheel 22, and their lower ends are supported on the frame section 12. A pair of left and right rear pillars 26 are provided behind the rear seat section 20, and their lower ends are supported on the frame section 12. The roof section 28 is supported by the front pillars 24 and rear pillars 26 so as to cover the front seat section 18, rear seat section 20 and steering wheel 22 configured in this way from above.

[0021] Figure 2 is a block diagram showing the electrical configuration for driving and controlling the golf cart shown in Figure 1. Inside the golf cart 10, a drive motor 30 is provided to drive the wheels (= front wheels 14 + rear wheels 16) and functions as a drive source. Of course, an engine may be equipped as a drive source instead of the drive motor 30, or a hybrid drive source combining a motor and an engine may be used. Also inside the golf cart 10, a control unit 32 is provided to control the drive motor 30.

[0022] The control unit 32 includes an autonomous driving control unit 34 that controls the golf cart 10 to autonomously travel along a guide line (not shown), a road surface condition detection device 36 corresponding to the first embodiment of the present invention, a warning output unit 38 that issues a warning to the occupants and the surroundings in response to the detection of three-dimensional objects by the road surface condition detection device 36, and a driving speed control unit 40 that controls the driving speed of the golf cart 10 according to the road surface conditions (including the presence or absence of three-dimensional objects) detected by the road surface condition detection device 36. Of the elements constituting the control unit 32, all except the road surface condition detection device 36 are the same as those used in the prior art. In contrast, the road surface condition detection device 36 is a characteristic configuration of the present invention and includes an example of the "three-dimensional object detection device" of the present invention, as will be described below. Therefore, in the following, the configuration and operation of the road surface condition detection device 36 will be described in detail, while the description of other configurations will be omitted.

[0023] Figure 3 is a functional block diagram showing the configuration of the road surface condition detection device incorporated into the golf cart shown in Figure 1. The road surface condition detection device 36 has a microprocessor and a memory unit, and detects the road surface conditions and three-dimensional objects present on the road surface in front of the golf cart 10 by implementing various functions, which will be described in detail later, based on point cloud data acquired by the distance sensor 42.

[0024] As shown in Figure 1, the distance sensor 42 is a three-dimensional sensor mounted on the roof 28 of the golf cart 10, and is a LiDAR (=light detection and ranging) sensor capable of acquiring three-dimensional point cloud data. In this embodiment, the distance sensor 42 irradiates the road surface (see Figure 4) with laser light as an example of electromagnetic waves from diagonally above. The distance sensor 42 scans the laser light in the left-right direction, that is, in the direction Y perpendicular to the direction of travel X of the golf cart 10, and receives the laser light reflected in front of the golf cart 10. The irradiation angle of the scanned laser light (symbol θ in Figure 4) can be switched in multiple stages. As a result, for each irradiation angle θ, the distance sensor 42 acquires point cloud data indicating the distance from the golf cart 10 to the reflection point and the direction of the reflection point relative to the golf cart 10 for multiple reflection points (road surface, three-dimensional objects, etc.) that reflected the laser light, and outputs it to the point cloud trajectory determination unit 44 as shown in Figure 3.

[0025] The point cloud trajectory determination unit 44 determines, based on the point cloud data from the distance sensor 42, whether the point cloud corresponding to multiple reflection points can draw an arc-shaped or elliptical arc-shaped road surface discrimination trajectory, and provides the determination result to the road surface determination unit 46. When the road surface determination unit 46 determines that a road surface discrimination trajectory can be drawn, it determines that there are no three-dimensional objects on the road surface and provides the determination result to the output unit 48. The output unit 48 then outputs the determination result to the autonomous driving control unit 34. This is how the road surface conditions are detected, and the operation of the point cloud trajectory determination unit 44 and the road surface determination unit 46 will be described in detail later with reference to Figures 4, 5A to 5D, 6 and 7, along with the operation of the deviation point identification unit 50 and the three-dimensional object identification unit 52, which will be described next.

[0026] Point cloud data from the distance sensor 42 is provided to the point cloud trajectory determination unit 44 and the deviation point identification unit 50 simultaneously. In addition to the point cloud data, the deviation point identification unit 50 is also provided with the determination result from the point cloud trajectory determination unit 44 (whether or not a road surface discrimination trajectory can be drawn). Based on this information, the deviation point identification unit 50 has the function of identifying deviation points among the points that make up the point cloud whose point cloud trajectory deviates from the road surface discrimination trajectory, and provides information about the deviation points (data from the point cloud data corresponding to the deviation points) to the three-dimensional object identification unit 52. The three-dimensional object identification unit 52 clusters the deviation points and classifies them into one or more deviation point groups. More specifically, as shown in Figures 5C and 5D, multiple deviation points DP that are adjacent to each other by a certain number or more and within a threshold distance from each other are made into one deviation point group DG. Then, for each deviation point group DG, the three-dimensional object identification unit 52 identifies three-dimensional object-related information such as the position of three-dimensional object OB on the road surface RS, and provides the identification result to the output unit 48. The output unit 48 then outputs the identification result to the autonomous driving control unit 34, the warning output unit 38, and the driving speed control unit 40. In this way, the three-dimensional object identification unit 52 functions as a clustering unit 521 and an individual identification unit 522.

[0027] Next, we will describe typical driving environments for the golf cart 10 when it is traveling on a road surface, and the point clouds acquired by the distance sensor 42 in each driving environment, with reference to Figures 4, 5A, 5B, 5C, and 5D. Following this explanation, we will describe the operation of the control unit 32 in detecting road surface conditions and three-dimensional objects.

[0028] Figure 4 is a schematic diagram showing typical driving environments for golf carts traveling on roads, illustrating four representative driving environments (A) to (D). (A) Flat road surface RS and no three-dimensional objects OB on the road surface RS. (B) Uphill road surface RS, and no three-dimensional object OB exists on the road surface RS. (C) Flat road surface RS, and a three-dimensional object OB exists on the road surface RS. (D) On an uphill road surface RS, and there is a three-dimensional object OB on the road surface RS. This is illustrated.

[0029] Furthermore, Figures 5A to 5D are plots on a two-dimensional plane showing the position coordinates of reflection points calculated from point cloud data acquired by the distance sensor in driving environments (A) to (D), respectively, for each irradiation angle θ. In addition, in Figures 5A to 5D, the golf cart 10, the distance sensor 42, and the (+Y) side edge EG+ and (-Y) side edge EG- of the road surface RS are shown with dotted lines for clarity.

[0030] In driving environment (A), the road surface RS extending in front of the golf cart 10 is flat, and there are no three-dimensional objects OB on the road surface RS. Therefore, for example, when the laser beam is scanned at irradiation angle θ0, the distance from the distance sensor 42 to the road surface RS is almost constant, and as shown in Figure 5A, the trajectory of the point cloud PG of the reflection point P has an arc shape. The same is true for other irradiation angles θ1, θ2, ..., θm.

[0031] In driving environment (B), the absence of three-dimensional objects OB on the road surface RS is consistent with driving environment (A), but the road surface RS is an uphill slope. In this case, for example, when the laser beam is scanned at irradiation angle θ0, as shown in Figure 5B, the trajectory of the portion PG1 corresponding to the road surface RS in the point cloud PG of the reflection point P has an elliptical arc shape, while the trajectory of the other portion PG2 has a circular arc shape, similar to driving environment (A), and the elliptical arc portion and the circular arc portion are continuous at each edge EG+, EG- of the road surface RS. The same applies to other irradiation angles θ1, θ2, ..., θm. Note that when the scanning range of the laser beam is within the width of the road surface RS, PG2 is not included, and the trajectory of the point cloud PG becomes an elliptical arc shape. Furthermore, while this explanation focuses on road surfaces RS with an uphill slope, for road surfaces RS with a downhill slope, the trajectory of the point cloud PG will be either a shape where elliptical and circular arc portions are connected (hereinafter referred to as "composite shape") or an elliptical arc shape, similar to the uphill slope described above. However, the curvature of the elliptical arc portion will differ.

[0032] When a three-dimensional object OB does not exist on the road surface RS, the trajectory of the point cloud PG of the reflection point P will be an arc shape, an elliptical arc shape, or a composite shape combining both. Moreover, the trajectory shape of the point cloud PG differs depending on whether the road surface RS is a flat surface or an inclined surface. Therefore, by determining whether the point clouds corresponding to multiple reflection points P can draw arc-shaped or elliptical arc-shaped trajectories (hereinafter referred to as "road surface discrimination trajectories T") based on the point cloud data, it becomes possible to accurately determine the road surface conditions.

[0033] Furthermore, if the road surface RS is an inclined surface, as described above, the trajectory of the portion PG1 corresponding to the road surface RS will be an elliptical arc, and its curvature will correspond to the inclination angle of the inclined surface. Therefore, the inclination angle of the inclined surface can be further determined based on the curvature of the elliptical road surface discrimination trajectory T.

[0034] The above considered the case where there is no three-dimensional object OB on the road surface RS. Next, we will consider the case where there is a three-dimensional object OB on the road surface RS. Driving environments (C) and (D) are the same as driving environments (A) and (B) in that the road surface RS is flat and inclined, respectively, but they differ from driving environments (A) and (B) in that there is a three-dimensional object OB on the road surface RS. In both cases, for example, when the laser beam is scanned at irradiation angle θ0, as shown in columns (C) and (D) of Figure 4, a portion of the laser beam scanned in the Y direction is reflected by the three-dimensional object OB. These reflection points P are on the golf cart 10 side, i.e., on the (-X) side, than the reflection point P on the road surface RS. As a result, as shown in Figures 5C and 5D, the portion PG3 of the point cloud PG corresponding to the three-dimensional object OB is located further to the (-X) side than the rest. In other words, the trajectory of the point cloud PG includes discontinuities and is neither a circular arc shape nor an elliptical arc shape. This is also true when the three-dimensional object OB is a depression or a ditch. Therefore, by determining whether the point clouds corresponding to multiple reflection points P can trace a circular or elliptical road surface discrimination trajectory T based on the point cloud data, it becomes possible to accurately determine the presence or absence of three-dimensional objects OB on the road surface RS. Furthermore, if it is determined that the trajectory of the point cloud PG contains discontinuities, it is also possible to determine that three-dimensional objects exist on the road surface RS.

[0035] Figure 6 is a flowchart illustrating the operation of the golf cart shown in Figure 1. The control unit 32 controls each part of the golf cart 10 according to a program pre-stored in a memory unit (not shown), guided by electromagnetic waves emitted from guide wires embedded in the road surface RS, thereby enabling the golf cart 10 to move autonomously. During or while the cart is moving, the road surface condition detection device 36 performs road surface condition detection and three-dimensional object detection as shown in Figures 6 and 7, according to the program. An example of the operation of the golf cart 10 will be described below with reference to Figures 5A to 5D, 6 and 7.

[0036] When the golf cart 10 is autonomously driving or temporarily stopped, the road surface condition detection device 36 uses the distance sensor 42 to repeatedly perform three-dimensional spatial imaging of the area in front of the golf cart 10 at regular intervals. In each spatial imaging, the laser beam is scanned in the Y direction from diagonally above the road surface RS at an irradiation angle θ, and the irradiation angle θ is switched in multiple stages. In this embodiment, the point cloud data acquired while the laser beam is scanned in the Y direction at an irradiation angle θm is analyzed, and the position coordinates of the reflection points P are plotted on a two-dimensional plane, which is called a "layer," and a spatial image is constructed by multiple layers 0, 1, ..., m. However, in this embodiment, since a three-dimensional LiDAR is used as the distance sensor 42, "m" is a natural number.

[0037] In this embodiment, the variable k is set to an initial value of 0 (step S1). The road surface condition detection device 36 then analyzes the point cloud data acquired while scanning the laser beam in the Y direction at an irradiation angle θk, and obtains a layer k (a plot of the position coordinates of the reflection point P on a two-dimensional plane). That is, for example, a point cloud PG as shown in Figures 5A to 5D is acquired (step S2). The layer k thus acquired is temporarily stored in a storage unit (not shown) for use in the three-dimensional object detection process which will be explained later.

[0038] The road surface condition detection device 36 calculates the trajectory of the point cloud PG, i.e., the road surface discrimination trajectory T, based on the position coordinates of the point cloud PG included in layer k (step S3). For this calculation, conventionally known approximation methods such as the least squares method can be used.

[0039] In the next step S4, the road surface condition detection device 36 calculates the percentage of points in the point cloud PG that match the road surface discrimination trajectory T obtained in step S3 (hereinafter referred to as the "matching percentage"). Here, for example, as shown in Figures 5A and 5B, if the three-dimensional object OB does not exist on the road surface RS, the matching percentage will be high. On the other hand, for example, as shown in Figures 5C and 5D, if the three-dimensional object OB exists on the road surface RS, discontinuities are included in the road surface discrimination trajectory T, and the matching percentage decreases.

[0040] Subsequently, the road surface condition detection device 36 determines whether the analysis (calculation of road surface discrimination trajectory T and matching ratio) for these layers k has been performed for all layers, that is, whether k = m (step S5). As long as there are layers that have not been analyzed (NO in step S5), the road surface condition detection device 36 increments the variable k by 1 (step S6) and then returns to step S2.

[0041] Once the matching ratio is calculated for all layers, the road surface condition detection device 36 determines whether all of the matching ratios obtained in step S4 are equal to or greater than a preset threshold (corresponding to an example of the "constant ratio" in the present invention) (step S7). That is, when a three-dimensional object OB is present on the road surface RS, the laser light is reflected by the three-dimensional object OB at least some irradiation angles θ. At these irradiation angles θ, the point indicating the position coordinates of the reflection point P reflected by the three-dimensional object OB deviates from the road surface discrimination trajectory T, reducing the matching ratio. Therefore, if the device determines "YES" in step S7, that is, if it determines that the three-dimensional object OB does not exist, the road surface condition detection device 36 proceeds directly to step S9. On the other hand, if the device determines "NO" in step S7, the road surface condition detection device 36 performs three-dimensional object detection processing (step S8) and then proceeds to step S9.

[0042] Figure 7 is a flowchart showing the 3D object detection process performed while the golf cart is in operation. In this 3D object detection process, the road surface condition detection device 36 executes steps S81 to S86 to acquire a group of deviation points DG corresponding to the 3D object OB for each layer 0, 1, ..., m. That is, after the variable k is set to an initial value of zero (step S81), the road surface condition detection device 36 reads the point cloud PG of layer k acquired in step S2 from the storage unit (step S82). Then, the road surface condition detection device 36 identifies points that deviate from the road surface discrimination trajectory T (hereinafter referred to as "deviation points DP") from the read point cloud PG (step S83). For example, in Figures 5C and 5D, points that are more than a certain distance away from the road surface discrimination trajectory T, i.e., deviation points DP, constitute the group of deviation points DG. For example, in the above driving environments (C) and (D), there is one 3D object OB. Therefore, point DP corresponds to the reflection point P of the laser beam at the object OB, and there is only one deviation point group DG in layer k. If there are multiple objects OB, there will be many deviation points DP, and multiple deviation point groups DG will also exist. Therefore, the road surface condition detection device 36 clusters the deviation points DP and classifies them into one or more deviation point groups (step S84).

[0043] Once the clustering process for layer k is complete, the road surface condition detection device 36 determines whether the clustering process for all layers is complete, that is, whether k = m (step S85). If it determines that it is not complete, the road surface condition detection device 36 increments the variable k by 1 (step S86), returns to step S82, and performs a series of processes (steps S82 to S84) for the next layer.

[0044] On the other hand, once it is determined that clustering processing has been completed for all layers ("YES" in step S85), the road surface condition detection device 36 determines whether or not at least one deviation point group DG exists (step S87). If the existence of a deviation point group DG is confirmed ("YES" in step S87), the road surface condition detection device 36 identifies the location of the object OB corresponding to the deviation point group DG based on the position coordinates of the deviation point DP included in the deviation point group DG (step S88). Subsequently, the road surface condition detection device 36 informs the warning output unit 38 that the object OB is present on the road surface RS in front of the golf cart 10, and the warning output unit 38 warns the occupants of the golf cart 10 of the presence of the object OB (step S89), before proceeding to step S9 in Figure 6. In contrast, if no deviation point groups DG exist at all ("NO" in step S87), the road surface condition detection device 36 proceeds to step S9 in Figure 6 without issuing the above warning.

[0045] Returning to Figure 6, the explanation continues. In step S9, the road surface condition detection device 36 determines whether the shape of the road surface discrimination trajectory T, particularly the shape within the range enclosed by both edges EG+ and EG- of the road surface RS, is an arc shape or an elliptical arc shape. This takes into account that the shape of the road surface discrimination trajectory T corresponds to the condition of the road surface RS, as described above. In other words, when the road surface condition detection device 36 determines that the shape is an arc shape, it determines that the road surface RS is flat (step S10), while when it determines that the shape is an elliptical arc shape, it determines that the road surface RS is an inclined surface (step S11).

[0046] As described above, in the first embodiment, point cloud data is acquired by a distance sensor 42 consisting of a three-dimensional LiDAR attached to a golf cart 10 traveling on the road surface RS. Based on the point cloud data, when the point cloud PG corresponding to multiple reflection points P can draw an arc-shaped or elliptical arc-shaped road surface discrimination trajectory T, it is determined that no three-dimensional object exists on the road surface RS. Therefore, regardless of whether the road surface RS in front of the golf cart 10 is a flat surface or an inclined surface, and even when the golf cart 10 is stationary, it is possible to accurately detect whether or not a three-dimensional object OB exists on the road surface RS.

[0047] Furthermore, in the first embodiment, the shape of the road surface discrimination trajectory T allows for high-precision, real-time detection of whether the road surface RS is a flat surface or an inclined surface extending in the direction of travel X. In particular, when the road surface RS is an inclined surface, its inclination angle is related to the curvature of the road surface discrimination trajectory T. Therefore, the road surface condition detection device 36 can further determine the inclination angle of the inclined surface (road surface RS) from the curvature. This function may be incorporated into the road surface condition detection device 36, thereby enabling more detailed detection of the road surface condition.

[0048] Furthermore, in the first embodiment, the deviation point group DG is determined as a discontinuity in the road surface discrimination trajectory T based on point cloud data, and three-dimensional objects OB are detected. Therefore, as shown in Figures 5C and 5D, not only can objects extending upward from the road surface RS be detected with high accuracy, but sinkholes and ditches extending downward from the road surface RS can also be detected with high accuracy. In other words, the road surface condition detection device 36 according to the first embodiment has excellent versatility.

[0049] Furthermore, in the first embodiment, since one or more deviation point groups DG are acquired to detect three-dimensional objects OB, all three-dimensional objects OB present on the road surface RS can be accurately detected, including their positional information.

[0050] Thus, in the first embodiment, the road surface condition detection device 36 has the function of detecting road surface conditions as described above, as well as the function of detecting three-dimensional objects OB as a three-dimensional object detection device. In other words, as shown by the dashed line in Figure 3, an example of the "three-dimensional object detection device" of the present invention is formed by some elements of the road surface condition detection device 36 (distance sensor 42, deviation point identification unit 50, and three-dimensional object identification unit 52).

[0051] It should be noted that the present invention is not limited to the embodiments described above, and various modifications can be made in addition to those described above, without departing from the spirit of the invention. For example, in the first embodiment, the present invention is applied to a golf cart 10, which corresponds to an example of a "vehicle" of the present invention. However, vehicles to which the present invention can be applied include automated guided vehicles (AGVs) and robotic vacuum cleaners used in semiconductor factories and logistics warehouses. For example, an AGV used in a semiconductor factory travels on a road surface RS installed within the semiconductor factory while constantly communicating with a host computer that controls the entire semiconductor factory. The present invention can be applied to such an AGV (second embodiment). In this second embodiment, if a three-dimensional object OB exists on the road surface RS, it can be accurately detected in real time. Moreover, since the distance sensor 42 is a three-dimensional sensor and performs three-dimensional spatial imaging in front of the AGV, the three-dimensional object detection process in the second embodiment may be configured as follows.

[0052] Figure 8 is a flowchart showing the three-dimensional object detection process performed in the second embodiment of the road surface condition detection device according to the present invention. The main difference between this second embodiment and the three-dimensional object detection process (step S8) of the first embodiment is the processing when the deviation point group DG is acquired; other configurations and operations are basically the same as in the first embodiment. Therefore, the following explanation will focus on the differences, and identical configurations and operations will be denoted by the same reference numerals and their explanations will be omitted.

[0053] In the second embodiment, similar to the first embodiment, a deviation point group DG corresponding to the three-dimensional object OB is acquired for each layer 0, 1, ..., m (steps S81 to S86). After that, the road surface condition detection device 36 determines whether or not at least one deviation point group DG exists (step S87). If the existence of a deviation point group DG is confirmed (YES in step S87), the road surface condition detection device 36 identifies the position of the three-dimensional object OB corresponding to the deviation point group DG based on the position coordinates of the deviation point DP included in the deviation point group DG, and further identifies the size and shape of the three-dimensional object OB by taking into account the layer information of the deviation point DP (laser beam irradiation angle θ) (step S88A). In other words, in step S88A, the road surface condition detection device 36 identifies relevant information such as the position, size and shape of the three-dimensional object OB corresponding to the deviation point group DG. Subsequently, the road surface condition detection device 36 outputs relevant information about the three-dimensional object OB identified in step S88A to the host computer (step S89A).

[0054] Thus, in the second embodiment, instead of simply detecting the presence of a three-dimensional object OB on the road surface RS, detailed related information about the three-dimensional object OB present on the road surface RS, that is, a spatial image of the three-dimensional object OB, is provided to the semiconductor factory. Upon receiving this, the host computer analyzes the three-dimensional object OB and takes action, such as changing the movement path of the automated guided vehicle to avoid the three-dimensional object OB. In this way, by applying the present invention to automated guided vehicles that are responsible for the automatic transport of semiconductor wafers and the like in a semiconductor factory, it becomes possible to perform the above-mentioned automatic transport stably.

[0055] Furthermore, although the distance sensor 42 is configured as a three-dimensional LiDAR in the above embodiment, a two-dimensional LiDAR may be used instead of a three-dimensional LiDAR for the purpose of detecting road surface conditions and detecting three-dimensional objects. In this case, the detection process will be performed using only layer 0 in the above embodiment. In addition, the distance sensor 42 is not limited to LiDAR, and any sensor having the following functions can be used as the distance sensor 42. That is, any sensor that can receive electromagnetic waves reflected in front of the vehicle while scanning electromagnetic waves irradiated from diagonally above the road surface RS in a direction Y perpendicular to the vehicle's travel direction X with respect to the road surface RS, and thereby acquire point cloud data indicating the distance from the vehicle to the reflection point P and the direction of the reflection point P relative to the vehicle for multiple reflection points P that reflected electromagnetic waves, can be used. [Industrial applicability]

[0056] This invention can be applied to detection technologies in general for detecting the condition of the road surface on which a vehicle is traveling and three-dimensional objects present on the road surface, as well as to vehicles equipped with such technologies. [Explanation of Symbols]

[0057] 10…Golf cart (vehicle) 32…Control Unit 36...Road surface condition detection device 42... Distance sensor 44...Point cloud trajectory determination section 46...Road surface judgment section 50... Deviation point identification unit 52…3D object identification part 521...Clustering section 522... Individual Identification Department DG... Deviance Group DP…Deviation OB…3D object P…Reflection point PG…Point cloud RS…road surface T…Road surface discrimination trajectory X…Travel direction Y… (direction perpendicular to the direction of travel) θ…Irradiation angle

Claims

1. a distance sensor attached to a vehicle traveling on a road surface, which scans the road surface in a direction perpendicular to the traveling direction of the vehicle with electromagnetic waves irradiated from diagonally above the road surface while receiving electromagnetic waves reflected in front of the vehicle, thereby obtaining point cloud data indicating the distance from the vehicle to a plurality of reflection points that reflect the electromagnetic waves and the direction of the reflection points relative to the vehicle; a point cloud trajectory determination unit that determines whether a point cloud corresponding to the plurality of reflection points can draw a circular arc-shaped or elliptical arc-shaped road surface discrimination trajectory based on the point cloud data; a road surface determination unit that determines that no three-dimensional object exists on the road surface when the point cloud trajectory determination unit determines that the point cloud can draw the road surface discrimination trajectory; and A road surface condition detection device comprising:

2. The road surface condition detection device according to claim 1, The road surface determination unit When the road surface determination unit determines that the point cloud can draw the arc-shaped road surface determination locus, the road surface is determined to be flat; a road surface condition detection device that determines, when the road surface determination unit determines that the point cloud can draw the road surface discrimination locus in the shape of an elliptical arc, that the road surface is an inclined surface that is inclined in the traveling direction;

3. The road surface condition detection device according to claim 2, When the road surface determination unit determines that the road surface is an inclined surface inclined in the traveling direction, the road surface determination unit further calculates an inclination angle of the inclined surface based on the curvature of the elliptical arc-shaped road surface determination locus.

4. The road surface condition detection device according to any one of claims 1 to 3, The point cloud trajectory determination unit determining that the point cloud can trace the road surface discrimination trajectory when a certain proportion or more of the points constituting the point cloud are located on the road surface discrimination trajectory; determining that the point cloud cannot depict the road surface discrimination trajectory when the number of points located on the road surface discrimination trajectory among the plurality of points constituting the point cloud is less than a certain percentage; The road surface determination unit determines that a three-dimensional object exists on the road surface when the point cloud trajectory determination unit determines that the point cloud cannot draw the road surface discrimination trajectory.

5. The road surface condition detection device according to any one of claims 1 to 3, The point cloud trajectory determination unit determining that the point cloud can draw the road surface discrimination trajectory when the point cloud trajectory drawn by the point cloud does not include any discontinuous portions; determining that the point cloud cannot draw the road surface discrimination trajectory when a discontinuous portion is included in the point cloud trajectory drawn by the point cloud; The road surface determination unit determines that a three-dimensional object exists on the road surface when the point cloud trajectory determination unit determines that the discontinuous portion is included.

6. The road surface condition detection device according to any one of claims 1 to 5, the distance sensor is a three-dimensional sensor configured to be able to scan the irradiation angle of the electromagnetic wave in multiple stages, The point cloud trajectory determination unit determines whether the point cloud can draw the road surface discrimination trajectory for each of different illumination angles.

7. The road surface condition detection device according to claim 6, The road surface determination unit when the point cloud trajectory determination unit determines that the point cloud can trace the road surface discrimination trajectory for all of the plurality of illumination angles, it is determined that no three-dimensional object exists on the road surface; a road surface condition detection device that determines that a three-dimensional object exists on the road surface when the point cloud trajectory determination unit determines that the point cloud cannot trace the road surface discrimination trajectory for at least one of the plurality of irradiation angles.

8. a distance sensor attached to a vehicle traveling on a road surface, which scans the road surface in a direction perpendicular to the traveling direction of the vehicle with electromagnetic waves irradiated from diagonally above the road surface while receiving electromagnetic waves reflected in front of the vehicle, thereby obtaining point cloud data indicating the distance from the vehicle to a plurality of reflection points that reflect the electromagnetic waves and the direction of the reflection points relative to the vehicle; a deviation point identification unit that compares a point cloud locus drawn by a point cloud corresponding to the plurality of reflection points with a circular arc or elliptical arc road surface discrimination locus based on the point cloud data, and identifies deviation points among the points that constitute the point cloud where the point cloud locus deviates from the road surface discrimination locus; a three-dimensional object identification unit that identifies the presence of a three-dimensional object on the road surface based on data corresponding to the deviation point among the point cloud data acquired by the distance sensor; A three-dimensional object detection device comprising:

9. The three-dimensional object detection device according to claim 8, The three-dimensional object identification unit a clustering unit that clusters the deviation points identified by the deviation point identification unit and classifies them into one or more deviation point groups; an individual identification unit that identifies, for each of the deviation point groups, the presence of a three-dimensional object on the road surface; A three-dimensional object detection device having the above.

10. The three-dimensional object detection device according to claim 9, The clustering unit classifies a certain number or more of the deviation points that are adjacent to each other at a distance within a threshold into the deviation point groups.

11. A three-dimensional object detection device according to any one of claims 8 to 10, A vehicle characterized in that its running on the road surface is controlled based on the presence of the three-dimensional object identified by the three-dimensional object detection device.

12. a step of acquiring point cloud data indicating the distance from the vehicle to a plurality of reflection points that reflect the electromagnetic waves and the direction of the reflection points relative to the vehicle by scanning the road surface in a direction perpendicular to the traveling direction of the vehicle with electromagnetic waves that are irradiated from obliquely above the road surface and receiving the electromagnetic waves reflected in front of the vehicle; a step of comparing a point cloud locus drawn by a point cloud corresponding to the plurality of reflection points based on the point cloud data with a circular arc-shaped or elliptical arc-shaped road surface discrimination locus; identifying deviation points, among the points constituting the point cloud, at which the point cloud trajectory deviates from the road surface discrimination trajectory; identifying the presence of a three-dimensional object on the road surface based on data corresponding to the deviation point in the point cloud data; A three-dimensional object detection method comprising:

Images