Risk estimation system and vehicle control system

Abstract

To estimate a risk degree in a forward direction of a vehicle with high accuracy.SOLUTION: A risk estimation system estimates a risk degree of a vehicle in a forward direction. Three-dimensional object arranging information indicates a three-dimensional arrangement of a three-dimensional object in the forward direction of the vehicle. A cross road is a road that intersects with a traveling road on which the vehicle is traveling at an intersection in the forward direction of the vehicle. The risk estimation system sets a first region that extends by a first length into the cross road from the intersection. The risk estimation system calculates a shielding degree that indicates a degree to which the first region is shielded by the three-dimensional object when viewed from a position of the vehicle based on the three-dimensional arrangement information. Then, the risk estimation system calculates the risk degree to be higher as the shielding degree increases.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present disclosure relates to a technique for estimating the risk level in front of a vehicle. The present disclosure also relates to a technique for controlling a vehicle based on the risk level.

Background Art

[0002] Patent Document 1 discloses a technique for estimating the risk of jumping out from a blind spot. The blind spot is formed by an obstacle in front of the vehicle. The risk of jumping out is calculated to be greater as the "blind spot size angle" representing the size of the blind spot is larger.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Consider an intersecting road that intersects the road on which the vehicle is traveling. An object entering the intersection from the intersecting road can pose a risk to the vehicle. In particular, when there is an obstacle in front of the intersecting road as seen from the vehicle, the visibility of the intersecting road deteriorates, and there is a risk that an object existing on the intersecting road may become difficult to see. However, the structures of obstacles existing in the actual environment are diverse. There are some obstacles that exist continuously, and there are also some that exist intermittently. There are tall obstacles and there are also low obstacles. Therefore, just because an obstacle exists, it does not necessarily mean that an object existing on the intersecting road cannot be seen.

[0005] In the aforementioned Patent Document 1, risk is estimated based on the size and angle of the blind spot. However, the size and angle of the blind spot alone do not adequately consider the structure of obstructions or the visibility of intersecting roads. In other words, the size and angle of the blind spot cannot quantitatively evaluate the poor visibility of intersecting roads. The accuracy of risk estimation based on such a blind spot size and angle is not necessarily high.

[0006] One objective of this disclosure is to provide a technology that can accurately estimate the degree of risk in front of a vehicle. [Means for solving the problem]

[0007] The first aspect concerns a risk estimation system for estimating the degree of risk ahead of a vehicle. The risk estimation system comprises one or more processors and one or more storage devices that store three-dimensional object arrangement information indicating the three-dimensional arrangement of objects in front of the vehicle. An intersecting road is a road that crosses the road a vehicle is currently traveling on at an intersection ahead of it. One or more processors define a first region extending a first length from the intersection into the intersecting road. Based on the placement information of three-dimensional objects, one or more processors calculate the degree of occlusion, which indicates the degree to which the first region is obscured by three-dimensional objects as viewed from the vehicle's position. Then, one or more processors calculate a risk level such that the risk level increases as the degree of occlusion increases.

[0008] The second perspective concerns vehicle control systems that control the vehicle. The vehicle control system includes the risk estimation system described above. One or more processors perform risk avoidance control, which involves steering the vehicle and / or decelerating it to avoid risks ahead of the vehicle, based on the degree of risk. [Effects of the Invention]

[0009] According to this disclosure, the three-dimensional arrangement of objects in front of a vehicle is taken into consideration when estimating the risk level ahead of the vehicle. More specifically, based on the three-dimensional arrangement of objects, a degree of occlusion is calculated, indicating the degree to which a first area within the intersecting road is obscured by the objects. The risk level is then calculated such that it increases as the degree of occlusion increases. By considering the three-dimensional arrangement of objects, it becomes possible to estimate the risk level ahead of a vehicle with greater accuracy. [Brief explanation of the drawing]

[0010] [Figure 1] This is a conceptual diagram illustrating the overview of the risk estimation system. [Figure 2] This is a conceptual diagram illustrating an example of risk estimation processing. [Figure 3] This is a conceptual diagram to explain variations. [Figure 4] This is a block diagram showing an example configuration of a risk estimation system. [Figure 5] This is a flowchart showing the risk estimation process. [Figure 6] This is a block diagram showing an example configuration of a vehicle control system. [Modes for carrying out the invention]

[0011] 1. Risk Estimation System Figure 1 is a conceptual diagram illustrating the overview of the risk estimation system 100 applied to the vehicle 1 according to this embodiment.

[0012] Vehicle 1 may be a vehicle driven by an operator, or it may be an autonomous vehicle equipped with an autonomous driving system. The operator may be a driver on board Vehicle 1, or it may be a remote operator who remotely drives Vehicle 1. Vehicle 1 is traveling on road 2. Intersecting road 3 is a road that intersects with road 2 at intersection 4 ahead of Vehicle 1. The X direction is the direction of extension of road 2, and the Y direction is the direction of extension of intersecting road 3. The X and Y directions intersect.

[0013] A typical example of a risk ahead of vehicle 1 is an object 5 present on the road 2. Examples of objects 5 include pedestrians, bicycles, motorcycles, other vehicles, etc. Objects 5 present on the road 2 can be said to be a "manifest risk" for vehicle 1. However, the risks ahead of vehicle 1 are not limited to such manifest risks, but can also include "potential risks" that could become manifest risks in the future. For example, an object 5 that may enter the road 2 in the future is a "potential risk" for vehicle 1.

[0014] The risk estimation system 100 not only recognizes the apparent risks in front of vehicle 1, but also predicts the potential risks in front of vehicle 1. The risk estimation system 100 then estimates (calculates) a "risk degree R" that represents the magnitude of the risks in front of vehicle 1. The risk degree R obtained by the risk estimation system 100 is used, for example, for risk avoidance control of vehicle 1 or for risk notification to the operator of vehicle 1.

[0015] The following explanation will focus specifically on the potential risks in front of Vehicle 1 and their risk level R. The methods for addressing the apparent risks are optional and not particularly limited.

[0016] As described above, intersecting road 3 intersects with road 2 at intersection 4 ahead of vehicle 1. Object 5, located on intersecting road 3, may enter intersection 4 and therefore poses a potential risk to vehicle 1. As shown in Figure 1, there are often three-dimensional objects (obstructions) OBJ located in front of intersecting road 3 from the perspective of vehicle 1. Examples of three-dimensional objects OBJ include buildings, walls, fences, guardrails, signs, utility poles, and other stationary objects. If intersecting road 3 is obstructed by such three-dimensional objects OBJ, the visibility of intersecting road 3 is reduced, and object 5 located on intersecting road 3 may become difficult for the driver or the autonomous driving system to see. Therefore, the risk level R when three-dimensional objects OBJ are present should generally be calculated to be higher than the risk level R when three-dimensional objects OBJ are not present.

[0017] However, the structures of the three-dimensional objects OBJ existing in the actual environment are diverse. There are three-dimensional objects OBJ that exist continuously (see (A) in FIG. 1), and there are also three-dimensional objects OBJ that exist intermittently (see (B) in FIG. 1). There are high three-dimensional objects OBJ, and there are also low three-dimensional objects OBJ (see (C) in FIG. 1). The intermittently existing three-dimensional objects OBJ and the low three-dimensional objects OBJ do not completely hide the intersection road 3 from the driver or the automatic driving system. That is, even if there are intermittently existing three-dimensional objects OBJ or low three-dimensional objects OBJ, the visibility of the intersection road 3 does not necessarily deteriorate extremely.

[0018] Thus, just because there is a three-dimensional object OBJ in front of the intersection road 3 as seen from the vehicle 1, it does not necessarily mean that the object 5 existing in the intersection road 3 cannot be seen. If the risk level R is estimated only based on the presence or absence of the three-dimensional object OBJ, the estimation accuracy cannot necessarily be said to be high. Therefore, this embodiment proposes a technique that can estimate the risk level R in front of the vehicle 1 with higher accuracy by considering the structure (three-dimensional arrangement) of the three-dimensional object OBJ.

[0019] FIG. 2 is a conceptual diagram for explaining an example of the risk estimation process by the risk estimation system 100 according to this embodiment.

[0020] The risk estimation system 100 recognizes the intersection 4 and the intersection road 3 in front of the vehicle 1. Further, the risk estimation system 100 sets a "first area D1" for estimating the risk level R in the intersection road 3 in front of the vehicle 1. The first area D1 is an area where there may be an object 5 entering the intersection 4 in the near future. More specifically, the first area D1 is set to extend by a first length L1 from the vicinity of the intersection 4 into the intersection road 3 (that is, in the Y direction). For example, the first area D1 is set to extend by a first length L1 from the edge of the intersection 4 into the intersection road 3 (that is, in the Y direction). The width of the pedestrian lane may also be included in the first length L1.

[0021] The first length L1 is determined by, for example, the following criteria. Suppose that an object 5 in the intersecting road 3 can be recognized from the position of vehicle 1 before intersection 4, and that vehicle 1 will stop in 2 seconds due to emergency braking control (AEBS) which starts almost simultaneously with recognition. The first length L1 is set to the distance that object 5 is expected to move during those 2 seconds. For example, if it is assumed that object 5 (e.g., pedestrian, bicycle) moves at 4 m / s, the first length L1 is set to 8 m. If the first region D1 of such first length L1 is clearly visible without being obstructed by the three-dimensional object OBJ, then the risk level R can be said to be low. The time it takes for vehicle 1 to stop due to emergency braking control also depends on the speed of vehicle 1. Therefore, the risk estimation system 100 may set the first length L1 to be larger as the speed of vehicle 1 increases. In other words, the risk estimation system 100 may set the first length L1 to be smaller as the speed of vehicle 1 decreases. Alternatively, the first length L1 may be set to a predetermined value.

[0022] The width of the first region D1 in the X direction is arbitrary. The width of the first region D1 in the X direction may be a predetermined width, or it may be the width of the intersecting road 3. Furthermore, the first region D1 may be a space having a predetermined height from the road surface of the intersecting road 3.

[0023] Furthermore, the risk estimation system 100 acquires "object placement information" that shows the three-dimensional arrangement of the three-dimensional object OBJ in front of vehicle 1. The three-dimensional arrangement includes the horizontal position in the XY plane and the height in the direction perpendicular to the XY plane. Typically, the horizontal position is expressed in absolute coordinates (latitude, longitude). The height may be the relative height as seen from vehicle 1, the height from the road surface, or the altitude in the absolute coordinates.

[0024] The risk estimation system 100 calculates the "degree of occlusion S" of the first region D1 based on the three-dimensional arrangement of the three-dimensional object OBJ indicated by the three-dimensional object arrangement information. The degree of occlusion S indicates the degree to which the first region D1 is occluded by the three-dimensional object OBJ when viewed from the reference position of vehicle 1. The reference position of vehicle 1 may be the installation position of the recognition sensor mounted on vehicle 1, the eye position of a standard driver, or a predetermined position on vehicle 1. The reference position of vehicle 1 in the absolute coordinate system is obtained based on the position of vehicle 1 acquired by a GNSS sensor or the like. The positions of intersection 4 and intersecting road 3 in the absolute coordinate system are obtained from map information. The risk estimation system 100 can calculate the degree of occlusion S of the first region D1 based on the reference position of vehicle 1, map information, and three-dimensional object arrangement information. The degree of occlusion S can also be said to be the reciprocal of the visibility of the first region D1 when viewed from the reference position of vehicle 1.

[0025] As an example, a method for calculating the degree of occlusion S in two dimensions will be explained with reference to Figure 2. In this example, the height of the three-dimensional object OBJ is the height from the road surface. Three-dimensional objects OBJ are classified into low-level three-dimensional objects OBJ-A, which are less than a predetermined height h_th, and high-level three-dimensional objects OBJ-B, which are greater than or equal to a predetermined height h_th. For example, the predetermined height h_th is 0.8m. When viewed from the reference position of vehicle 1, it is assumed that the first region D1 is occluded by the high-level three-dimensional object OBJ-B but not by the low-level three-dimensional object OBJ-A. The risk estimation system 100 classifies the three-dimensional objects OBJ in front of vehicle 1 into low-level three-dimensional objects OBJ-A and high-level three-dimensional objects OBJ-B based on the three-dimensional object placement information. The risk estimation system 100 then calculates the degree of occlusion S by determining that the first region D1 is not occluded by the low-level three-dimensional object OBJ-A, but is occluded by the high-level three-dimensional object OBJ-B.

[0026] More specifically, the risk estimation system 100 sets a virtual straight line connecting the reference position of vehicle 1 and the determination point within the first region D1. If a tall object OBJ-B exists on the virtual straight line, the determination point is determined to be obscured by the tall object OBJ-B. On the other hand, if a tall object OBJ-B does not exist on the virtual straight line, the determination point is determined to be unobstructed and visible. The visible region DA is the set of visible determination points within the first region D1. Conversely, the occluded region DB is the set of determination points within the first region D1 that are obscured by the tall object OBJ-B.

[0027] As shown in Figure 2, only a hypothetical straight line connecting the reference position of vehicle 1 and the edge of the high-dimensional object OBJ-B may be considered. As shown in Figure 2, the region enclosed by such a hypothetical straight line becomes the visible region DA or the occluded region DB. The visible region DA and the occluded region DB appear alternately. By considering a hypothetical straight line connecting the reference position of vehicle 1 and the edge of the high-dimensional object OBJ-B, the first region D1 can be easily divided into the visible region DA and the occluded region DB. This method is preferable from the viewpoint of reducing computational load.

[0028] In this way, the risk estimation system 100 divides the first region D1 into a visible region DA and a shielded region DB. The risk estimation system 100 then calculates the degree of shielding S as the ratio of the total area of ​​the shielded region DB to the area of ​​the first region D1. When the degree of shielding S is expressed as a percentage from 0 to 100%, the degree of shielding S can also be referred to as the "shielding rate".

[0029] Alternatively, the length LA of the visible area DA and the length LB of the shielding area DB along the Y direction may be considered. For example, a specific X-direction position belonging to the first region D1 is specified. At the specific X-direction position, the length LA of the visible area DA and the length LB of the shielding area DB along the Y direction are obtained. The risk estimation system 100 calculates the degree of shielding S (shielding rate) as the ratio of the sum of the lengths LB of the shielding area DB to the first length L1 of the first region D1. For example, if the first length L1 is 8.0m and the sum of the lengths LB of the shielding area DB is 4.0m, the degree of shielding S (shielding rate) is 50%.

[0030] Furthermore, the first region D1 may be a space having a predetermined height from the road surface of the intersecting road 3. In this case as well, the above method can be extended and applied. Specifically, the visible region DA is the region (space) that is not occluded by the three-dimensional object OBJ, and the occluded region DB is the region (space) that is occluded by the three-dimensional object OBJ. The risk estimation system 100 divides the first region D1 into the visible region DA and the occluded region DB. The risk estimation system 100 then calculates the degree of occlusion S (occluding rate) as the ratio of the total volume of the occluded region DB to the volume of the first region D1.

[0031] After calculating the degree of occlusion S, the risk estimation system 100 calculates a risk degree R, which represents the magnitude of the risk in front of vehicle 1. The risk degree R includes a first risk degree R1, which represents the magnitude of the “potential risk” related to the intersecting road 3 in front of vehicle 1. The first risk degree R1 is expressed as a function of the degree of occlusion S of the first region D1. More specifically, the higher the degree of occlusion S of the first region D1, the worse the visibility in the first region D1, and therefore the higher the first risk degree R1. Conversely, the lower the degree of occlusion S of the first region D1, the better the visibility in the first region D1, and therefore the lower the first risk degree R1. The risk estimation system 100 calculates the first risk degree R1 (i.e., risk degree R) such that it becomes higher as the degree of occlusion S of the first region D1 increases.

[0032] As described above, according to this embodiment, when estimating the risk degree R in front of vehicle 1, the three-dimensional arrangement of three-dimensional objects OBJ in front of vehicle 1 is taken into consideration. More specifically, based on the three-dimensional arrangement of three-dimensional objects OBJ, a degree of occlusion S is calculated, which indicates the degree to which the first region D1 within the intersecting road 3 is occluded by the three-dimensional objects OBJ. Then, the risk degree R is calculated such that it increases as the degree of occlusion S increases. By taking into consideration the three-dimensional arrangement of three-dimensional objects OBJ, it becomes possible to estimate the risk degree R in front of vehicle 1 with higher accuracy.

[0033] Furthermore, according to this embodiment, the degree of shielding S of the first region D1 is calculated quantitatively. Since a quantitative index called the degree of shielding S is calculated, it becomes possible to quantitatively estimate the risk degree R based on the degree of shielding S. The quantitative risk degree R has a physical meaning, rather than being subjective. Such a quantitative risk degree R is easy to use in risk avoidance control, which will be described later.

[0034] Furthermore, according to this embodiment, the quantitative degree of shielding S is calculated geometrically using a simple method. This is preferable from the viewpoint of reducing the computational load.

[0035] Furthermore, the risk level R may include other risk levels from a different perspective than the first risk level R1 (R = R1 + R2...). For example, the risk level R may include a second risk level R2 that depends on the environment of intersection 4. For example, if traffic lights are installed at intersection 4, the second risk level R2 will be lower compared to when there are no traffic lights. As another example, the wider the road 2 on which vehicle 1 is traveling, the lower the second risk level R2 may be.

[0036] Figure 3 is a conceptual diagram illustrating a modified example. In this modified example, we consider the case where intersection 4 is a crossroads. Left intersecting road 3L is the intersecting road 3 on the left side as seen from vehicle 1. Right intersecting road 3R is the intersecting road on the right side as seen from vehicle 1. In front of vehicle 1, the road 2 intersects with both left intersecting road 3L and right intersecting road 3R. At a crossroads, object 5 is likely to move from one of the left intersecting roads 3L or right intersecting road 3R to the other. Therefore, the speed at which object 5 emerges into intersection 4 tends to be higher. To reflect this tendency specific to crossroads in the first risk level R1, the first risk level R1 may be calculated to be higher in the case of a crossroads.

[0037] Specifically, the risk estimation system 100 calculates the degree of occlusion S for both the left intersecting road 3L and the right intersecting road 3R. The left occlusion degree SL is the degree of occlusion S with respect to the left intersecting road 3L. The right occlusion degree SR is the degree of occlusion S with respect to the right intersecting road 3R. If both the left occlusion degree SL and the right occlusion degree SR are not zero, the risk estimation system 100 calculates a "total occlusion degree SA" such that it is greater than the simple sum of the left occlusion degree SL and the right occlusion degree SR. Then, the risk estimation system 100 calculates a first risk degree R1 such that it increases as the total occlusion degree SA increases. Figure 3 also shows an example of a map of the first risk degree R1 for a combination of left occlusion degree SL and right occlusion degree SR. In the modified version, the accuracy of the risk degree R is further improved because the first risk degree R1 is calculated taking into account the tendencies specific to intersections.

[0038] 2. Example of a Risk Estimation System Configuration Figure 4 is a block diagram showing an example configuration of the risk estimation system 100 according to this embodiment. Typically, the risk estimation system 100 is mounted on the vehicle 1. Alternatively, a part of the risk estimation system 100 may be located in an external device outside the vehicle 1.

[0039] The risk estimation system 100 includes a communication device 110, a recognition sensor 120, a vehicle state sensor 130, a position sensor 140, and a control device 150. The recognition sensor 120 is mounted on the vehicle 1 and recognizes (detects) the surrounding conditions of the vehicle 1. Examples of the recognition sensor 120 include a camera, LIDAR (Laser Imaging Detection and Ranging), radar, etc. The vehicle state sensor 130 detects the state of the vehicle 1. The vehicle state sensor 130 includes a speed sensor, acceleration sensor, yaw rate sensor, steering angle sensor, etc. The position sensor 140 detects the position and orientation of the vehicle 1. For example, the position sensor 140 includes a GNSS (Global Navigation Satellite System).

[0040] The control device 150 is a computer that controls the risk estimation system 100. The control device 150 includes one or more processors 151 (hereinafter simply referred to as processor 151) and one or more storage devices 152 (hereinafter simply referred to as storage devices 152). The processors 151 perform various processes. For example, the processor 151 includes a CPU (Central Processing Unit). The storage devices 152 store various information necessary for processing by the processors 151. Examples of storage devices 152 include volatile memory, non-volatile memory, HDD (Hard Disk Drive), SSD (Solid State Drive), etc.

[0041] The risk estimation program 153 is a computer program executed by the processor 151. The processor 151 executes the risk estimation program 153, thereby realizing the functions of the control device 150. The risk estimation program 153 is stored in the storage device 152. Alternatively, the risk estimation program 153 may be recorded on a computer-readable recording medium.

[0042] The control device 150 acquires driving environment information 200 that indicates the driving environment of vehicle 1. The driving environment information 200 is stored in the storage device 152. The driving environment information 200 includes map information 210, surrounding situation information 220, vehicle status information 230, location information 240, three-dimensional object arrangement information 250, risk level information 260, etc.

[0043] Map information 210 shows road configuration, lane arrangement, etc. Map information 210 may also be high-precision map information in which the 3D arrangement of three-dimensional objects (OBJ) is registered. Examples of three-dimensional objects (OBJ) include buildings, walls, fences, guardrails, signs, utility poles, and other stationary objects. For example, the control device 150 communicates with the map management system via the communication device 110 and obtains map information 210 from the map management system.

[0044] The surrounding environment information 220 is information indicating the conditions around vehicle 1 and is obtained from the recognition results of the recognition sensor 120. For example, the surrounding environment information 220 includes images captured by a camera. Another example is that the surrounding environment information 220 includes point cloud information measured by LIDAR. Furthermore, the surrounding environment information 220 includes object information regarding objects around vehicle 1. Examples of objects around vehicle 1 include pedestrians, bicycles, other vehicles, three-dimensional objects (OBJ), white lines, crosswalks, traffic lights, signs, etc. The object information indicates the relative position and relative velocity of the object with respect to vehicle 1. For example, an object can be identified based on a LIDAR point cloud, and its relative position and relative velocity can be obtained. Another example is that an object can be identified and its relative position can be calculated by analyzing images captured by a camera. The object information may also include the direction and speed of movement of the object.

[0045] The vehicle status information 230 indicates the vehicle status detected by the vehicle status sensor 130. The vehicle status includes the vehicle's speed, acceleration, yaw rate, steering angle, etc.

[0046] Location information 240 indicates the current position and orientation of vehicle 1. Location information 240 is obtained by position sensor 140. High-precision location information 240 may be obtained by self-localization processing (Localization) using map information 210 and surrounding situation information 220 (object information).

[0047] The 3D object placement information 250 indicates the three-dimensional placement of the 3D object OBJ in front of the vehicle 1. The 3D placement includes the horizontal position in the XY plane and the height in the direction perpendicular to the XY plane. Typically, the horizontal position is expressed in absolute coordinates (latitude, longitude). The height may be the relative height as seen from the vehicle 1, the height from the road surface, or the altitude in the absolute coordinates. For example, the control device 150 obtains information on the relative position of the 3D object OBJ to the vehicle 1 from the surrounding situation information 220 (object information). The position of the vehicle 1 in the absolute coordinates is obtained from the position information 240. By combining the position of the vehicle 1 in the absolute coordinates and the relative position of the 3D object OBJ, the 3D placement of the 3D object OBJ in the absolute coordinates can be obtained. As another example, if map information 210 containing the 3D placement of the 3D object OBJ is available, the control device 150 may obtain the 3D object placement information 250 from that map information 210. If the horizontal plane in front of vehicle 1 is divided into a grid of a predetermined size (e.g., 0.2m square), the three-dimensional object placement information 250 may indicate the presence or absence of three-dimensional objects (OBJs) and height information for each grid. If the space in front of vehicle 1 is divided into voxels of a predetermined size (e.g., 0.2m cube), the three-dimensional object placement information 250 may indicate the presence or absence of three-dimensional objects (OBJs) for each voxel.

[0048] The risk level information 260 indicates the risk level R, which represents the magnitude of the risk in front of vehicle 1. The control device 150 estimates (calculates) the risk level R by performing the risk estimation process described in Section 1 above.

[0049] Figure 5 is a flowchart that summarizes the risk estimation process.

[0050] In step S100, the control device 150 determines whether or not an intersection 4 exists within a predetermined distance in front of vehicle 1. The positions of the intersecting road 3 and intersection 4 in absolute coordinate systems are obtained from map information 210. The position of vehicle 1 in absolute coordinate systems is obtained from position information 240. Based on the map information 210 and position information 240, the control device 150 determines whether or not an intersection 4 exists in front of vehicle 1. Alternatively, the control device 150 may determine whether or not an intersection 4 exists in front of vehicle 1 based on surrounding situation information 220 (object information). If an intersection 4 exists within a predetermined distance in front of vehicle 1 (step S100; Yes), the process proceeds to step S110. Otherwise (step S100; No), the process returns to step S100.

[0051] In step S110, the control device 150 sets a first region D1 (see Figure 2) within the intersecting road 3 in front of the vehicle 1. The first region D1 is set to extend by a first length L1 from the vicinity of the intersection 4 into the intersecting road 3 (i.e., in the Y direction). For example, the first region D1 is set to extend by a first length L1 from the edge of the intersection 4 into the intersecting road 3 (i.e., in the Y direction). The control device 150 may set the first length L1 to be larger as the speed of the vehicle 1 increases. The speed of the vehicle 1 is obtained from the vehicle status information 230.

[0052] In step S120, the control device 150 acquires the three-dimensional object arrangement information 250 as described above.

[0053] In step S130, the control device 150 calculates the degree of occlusion S of the first region D1. The degree of occlusion S indicates the degree to which the first region D1 is occluded by the three-dimensional object OBJ when viewed from the reference position of the vehicle 1. The reference position of the vehicle 1 may be the installation position of the recognition sensor 120 mounted on the vehicle 1, the position of a typical driver's eyes, or a predetermined position on the vehicle 1. The reference position of the vehicle 1 in the absolute coordinate system is obtained based on the position information 240. The positions of the intersection 4 and the intersecting road 3 in the absolute coordinate system are obtained from the map information 210. Based on the reference position of the vehicle 1, the map information 210, and the three-dimensional object placement information 250, the control device 150 calculates the degree of occlusion S of the first region D1.

[0054] In step S140, the control device 150 calculates a risk degree R representing the magnitude of the risk ahead of vehicle 1. The risk degree R includes a first risk degree R1 representing the magnitude of the “potential risk” related to the intersecting road 3 ahead of vehicle 1. The first risk degree R1 is expressed as a function of the shielding degree S of the first region D1. The control device 150 calculates the first risk degree R1 (i.e., risk degree R) such that it increases as the shielding degree S of the first region D1 increases.

[0055] Steps S110 to S140 are repeated until vehicle 1 passes through the intersection. If the first length L1 depends on the speed of vehicle 1, the first region D1 also changes in accordance with the change in the speed of vehicle 1. When vehicle 1 passes through the intersection (step S150; Yes), the process returns to step S100.

[0056] 3. Vehicle control system The risk level R in front of vehicle 1 may be used to control vehicle 1. The vehicle control system 100X that controls vehicle 1 will be described below. Note that the vehicle control system 100X includes the risk estimation system 100 described above.

[0057] Figure 6 is a block diagram showing an example configuration of the vehicle control system 100X according to this embodiment. In addition to the configuration shown in Figure 4, the vehicle control system 100X further includes a driving device 160 and a notification device 170.

[0058] The running gear 160 includes a steering gear, a drive gear, and a braking gear. The steering gear steers the wheels. For example, the steering gear includes an electric power steering (EPS) system. The drive gear is a power source that generates driving force. Examples of drive gears include an engine, an electric motor, an in-wheel motor, etc. The braking gear generates braking force. The control device 150 controls the running gear 160 (steering gear, drive gear, and braking gear) to perform vehicle driving control (steering control, drive control, and braking control).

[0059] The control device 150 may perform automatic driving control based on the driving environment information 200. More specifically, the control device 150 generates a driving plan for vehicle 1 based on the driving environment information 200. Furthermore, the control device 150 generates a target trajectory necessary for vehicle 1 to drive according to the driving plan based on the driving environment information 200. The target trajectory includes a target position and a target speed. The control device 150 then performs vehicle driving control so that vehicle 1 follows the target trajectory.

[0060] The control device 150 may perform "risk avoidance control" to proactively avoid risks in front of the vehicle 1. Typically, the risk for the vehicle 1 is an object 5 (see Figure 1) in front of the vehicle 1. Risk avoidance control includes at least one of steering control and deceleration control. For example, the control device 150 steers the vehicle 1 away from an apparent or potential risk in front of the vehicle 1. As another example, the control device 150 decelerates before reaching an apparent or potential risk in front of the vehicle 1.

[0061] The risk level R is used in this risk avoidance control. For example, if the risk level R exceeds the activation threshold, the control device 150 activates the risk avoidance control. As another example, the control device 150 may increase the steering amount or deceleration in the risk avoidance control as the risk level R increases. According to this embodiment, since the risk level R is estimated with high accuracy, the accuracy of the risk avoidance control based on that risk level R is also improved. For example, unnecessary activation of the risk avoidance control is suppressed.

[0062] The notification device 170 notifies the operator of vehicle 1 of various information. For example, the notification device 170 includes a display device. Examples of display devices include a display, a head-up display (HUD), etc. The notification device 170 may also include a speaker. The control device 150 can notify the operator of vehicle 1 of various information via the notification device 170.

[0063] The control device 150 may perform a "risk notification process" to notify the operator of the presence of a risk in front of the vehicle 1. The risk level R is used in this risk notification process. For example, if the risk level R exceeds an activation threshold, the control device 150 performs the risk notification process. As another example, the control device 150 may increase the risk notification intensity as the risk level R increases. The risk notification intensity is, for example, the size or volume of the display. According to this embodiment, since the risk level R is estimated with high accuracy, the accuracy of the risk notification process based on that risk level R is also improved. For example, unnecessary operation of the risk notification process is suppressed. [Explanation of Symbols]

[0064] 1…Vehicle, 2…Road, 3…Intersecting road, 4…Intersection, 100…Risk estimation system, 100X…Vehicle control system, D1…First domain, OBJ…Three-dimensional object, S…Shading degree, R…Risk degree

Claims

1. A risk estimation system for estimating the degree of risk ahead of a vehicle, One or more processors, One or more storage devices that store three-dimensional object arrangement information showing the three-dimensional arrangement of three-dimensional objects in front of the vehicle, Equipped with, The intersecting road is the road that intersects with the road the vehicle is traveling on at an intersection ahead of the vehicle. The one or more processors described above are: A first region is established that extends from the aforementioned intersection into the aforementioned intersecting road by a first length, As the speed of the vehicle increases, the first length is set to be larger. Based on the three-dimensional object arrangement information, a degree of occlusion is calculated, which indicates the degree to which the first region is obscured by the three-dimensional object when viewed from the position of the vehicle. The risk level is calculated such that it increases as the degree of shielding increases. Risk estimation system.

2. A risk estimation system according to claim 1, The one or more processors calculate the degree of occlusion by determining that the first region is not occluded by the three-dimensional object of less than a predetermined height, and that the first region is occluded by the three-dimensional object of greater than or equal to the predetermined height. Risk estimation system.

3. A risk estimation system for estimating the degree of risk in front of a vehicle, One or more processors, One or more storage devices that store three-dimensional object arrangement information showing the three-dimensional arrangement of three-dimensional objects in front of the vehicle, Equipped with, The intersecting road is the road that intersects with the road the vehicle is traveling on at an intersection ahead of the vehicle. The one or more processors described above are: A first region is established that extends from the aforementioned intersection into the aforementioned intersecting road by a first length, Based on the three-dimensional object arrangement information, a degree of occlusion is calculated, which indicates the degree to which the first region is obscured by the three-dimensional object when viewed from the position of the vehicle. The risk level is calculated such that it increases as the degree of shielding increases, At the aforementioned intersection, the aforementioned road intersects with the left-hand and right-hand intersecting roads. The aforementioned left-hand intersecting road is the intersecting road on the left side as viewed from the vehicle. The aforementioned right-hand intersecting road is the intersecting road on the right side as viewed from the vehicle. The left occlusion degree is the occlusion degree with respect to the left intersecting road, The right occlusion degree is the occlusion degree with respect to the right intersecting road, The one or more processors described above are: If both the left shading degree and the right shading degree are not zero, the total shading degree is calculated to be greater than the sum of the left shading degree and the right shading degree. The risk level is calculated such that it increases as the total shielding level increases. Risk estimation system.

4. A vehicle control system for controlling a vehicle, A risk estimation system according to any one of claims 1 to 3, The one or more processors perform risk avoidance control, which involves steering and decelerating the vehicle to avoid risks in front of the vehicle, based on the risk level. Vehicle control system.

Images