Automatic driving device

Abstract

To provide an automatic driving device capable of suppressing required performance of a necessary device by reducing communication and arithmetic processing.SOLUTION: The three dimensional point group of one's own vehicle is projected on a ground two-dimension plane (300), and point group coarseness (the number of point groups / unit volume) is calculated for each grid (302). Then, an angle at which a region where the point group density is equal to or less than a predetermined threshold value exists is calculated (304), and only the three dimensional point group data of the other vehicle 26 existing at the angle is acquired (306 to 316).SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] The present invention relates to an automatic driving device. [Background technology]

[0002] Patent Document 1 proposes a vehicle driving assistance system in which vehicle-to-vehicle communication is performed between the vehicle and another vehicle, and the vehicle receives blind spot images captured by a camera in the other vehicle, thereby enabling the vehicle to reliably detect the presence of another vehicle located in a blind spot in the far lane of a main road. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-018648 Summary of the Invention [Problem to be solved by the invention]

[0004] As with the technology in Patent Document 1, even if blind spot images captured by a camera in another vehicle are input into the autonomous driving system of an autonomous vehicle, the presence or absence of obstacles such as pedestrians, bicycles, or other vehicles can be determined by image recognition, but the positional relationship between the obstacle and the vehicle, or the direction or speed of the other vehicle's movement, cannot be determined, and therefore cannot be used as supplementary data for autonomous driving. It is necessary to obtain information to be used as supplementary data for autonomous driving, which increases communication and calculation processing, leaving room for improvement.

[0005] The present invention has been made in consideration of the above-mentioned circumstances, and aims to provide an automatic driving device that can reduce communication and calculation processing and suppress the performance requirements of necessary devices. [Means for solving the problem]

[0006] The autonomous driving device according to the first aspect includes a receiving unit that receives position information of other vehicles and three-dimensional point cloud data detected by the other vehicles, and a control unit that projects the three-dimensional point cloud data detected by the vehicle onto a two-dimensional plane to calculate a point cloud density, and controls the receiving unit to receive three-dimensional point cloud data detected by the other vehicles that are present in an area where the point cloud density is below a predetermined threshold, based on the position information received by the receiving unit.

[0007] According to the first aspect, the receiving unit receives position information of other vehicles and three-dimensional point cloud data detected by the other vehicles. The control unit then projects the three-dimensional point cloud data detected by the vehicle onto a two-dimensional plane to calculate a point cloud density, and controls the receiving unit to receive three-dimensional point cloud data detected by other vehicles located in an area where the point cloud density is equal to or less than a predetermined threshold, based on the position information received by the receiving unit. This reduces communication and processing loads, making it possible to lower the performance requirements of necessary devices. Furthermore, since obstacle information recognized and identified by each autonomous vehicle is not shared with other vehicles, but rather three-dimensional point cloud-level data is shared, other vehicles' information can be utilized within the scope of responsibility for the recognition performance of the autonomous driving system of the vehicle, without being affected by the obstacle recognition level of the other vehicles. [Effects of the Invention]

[0008] As described above, according to the present invention, it is possible to provide an automatic driving device that can reduce communication and calculation processing and suppress the performance requirements of necessary devices. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a block diagram showing a schematic configuration of an automatic driving device. [Figure 2] 10 is a flowchart showing an example of the flow of a reception selection process performed in the automatic driving device according to the basic embodiment. [Figure 3] 10 is a flowchart illustrating an example of the flow of a point cloud data merge process performed in the automatic driving device according to the basic embodiment. [Figure 4]FIG. 2 is a diagram for explaining an overview of a reception selection process in the autonomous driving device according to the present embodiment. [Figure 5] 10 is a flowchart illustrating an example of the flow of a reception selection process of the autonomous driving device according to the present embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] An example of an embodiment of the present invention will be described in detail below with reference to the drawings. In this embodiment, an automatic driving device that is mounted on an automatic driving vehicle and has basic functions (Localization Detection, Decision Planning Actuation) implemented for the automatic driving of the automatic driving vehicle and supports the automatic driving will be described as an example.

[0011] (basic form) Before describing the automatic driving device according to this embodiment, an automatic driving device according to a basic form will be described. Fig. 1 is a block diagram showing a schematic configuration of the automatic driving device according to the basic form.

[0012] The automatic driving device 10 according to the basic embodiment utilizes the detection results of the ADS sensor of other vehicles in order to reduce recognition-impairing blind spots that cannot be captured by the ADS sensor of the own vehicle in an automatic driving vehicle. Examples of recognition-impairing blind spots include blind spots on the other side of recognition-impairing obstacles caused by oncoming vehicles, blind spots on the other side of recognition-impairing obstacles caused by buildings, plants, walls, etc. along intersections, blind spots on the other side of obstacles caused by vehicles ahead, blind spots where the detection range in all directions is narrowed due to fog or rain, and blind spots where the forward detection range is narrowed in backlighting.

[0013] The automatic driving device 10 in the basic form retains the basic functions implemented in an automatic driving vehicle, but the detection function utilizes not only the sensing data of the own vehicle but also the sensing data of other vehicles in the three-dimensional point cloud data.

[0014] As shown in Figure 1, the automatic driving device 10 in the basic form includes a vehicle-to-vehicle communication unit 12, a reception selection processing unit 14, a point cloud data merge processing unit 16, a noise removal unit 20, a static object removal unit 22, and a map database (DB) 24.

[0015] The vehicle-to-vehicle communication unit 12 transmits and receives information to and from other vehicles 26 present in the vicinity of the vehicle. The vehicle-to-vehicle communication unit 12 performs so-called V2V (Vehicle to Vehicle) communication. In this embodiment, the vehicle-to-vehicle communication unit 12 transmits and receives position information of the other vehicles 26 and detection results of an ADS (Autonomous Driving System) sensor mounted on the other vehicles 26. Specifically, the vehicle-to-vehicle communication unit 12 transmits and receives three-dimensional point cloud data detected by a LiDAR 18 as an example of an ADS sensor. Note that hereinafter, the three-dimensional point cloud data may be simply referred to as point cloud data.

[0016] When the vehicle-to-vehicle communication unit 12 acquires the detection results of the ADS sensor from nearby other vehicles 26, the reception selection processing unit 14 selects the other vehicles 26 to acquire the detection results from, and controls the reception of the detection results of the LiDAR 18 detected by the other vehicles 26 via vehicle-to-vehicle communication from the selected other vehicles 26. In a basic form, the other vehicles 26 are selected from other vehicles 26 within a predetermined threshold distance from the position of the host vehicle. Note that the processing load may be reduced by further limiting the selection of other vehicles 26 to those in the traveling direction of the host vehicle.

[0017] The point cloud data merge processing unit 16 performs processing to merge three-dimensional point cloud data detected by the LiDAR 18 mounted on the host vehicle with three-dimensional point cloud data acquired from another vehicle 26. This makes it possible to supplement blind spots of the host vehicle with the point cloud data of the other vehicle 26. Therefore, by using the three-dimensional point cloud data merged with the point cloud data of the other vehicle 26 to detect obstacles in the same way as point cloud data processing of the host vehicle alone, it becomes possible to detect obstacles in blind spots that hinder recognition by the ADS sensor of the host vehicle.

[0018] The noise removal unit 20 performs processing to remove outliers and reflection intensities greater than or equal to a predetermined threshold value as noise from the point cloud data.

[0019] The static object removal unit 22 performs processing to remove points corresponding to static objects such as the ground and buildings from the point cloud data.

[0020] The map DB 24 stores point cloud data from which point cloud data other than obstacles has been deleted by the noise removal unit 20 and the static object removal unit 22.

[0021] Next, specific processing performed by the automatic driving device 10 according to the basic embodiment configured as described above will be described.

[0022] First, a description will be given of the reception selection process performed by the automatic driving device 10 according to the basic embodiment. Fig. 2 is a flowchart showing an example of the flow of the reception selection process performed by the automatic driving device 10 according to the basic embodiment.

[0023] In step 100, the vehicle-to-vehicle communication unit 12 receives position information from the other vehicle 26 through vehicle-to-vehicle communication, and the process proceeds to step .

[0024] In step 102, the reception selection processing unit 14 determines whether the distance between the other vehicle 26 and the subject vehicle is equal to or less than a predetermined threshold. If the determination is negative, the process returns to step 100 and repeats the above-described process. If the determination is positive, the process proceeds to step 104.

[0025] In step 104, the reception selection processing unit 14 determines whether the position of the other vehicle 26 is in front of the host vehicle. If the determination is negative, the process returns to step 100 and repeats the above-described process, but if the determination is positive, the process proceeds to step 106. Note that step 104 may be omitted.

[0026] In step 106, the reception selection processing unit 14 receives point cloud data from the other vehicle via vehicle-to-vehicle communication, and the series of reception selection processes is terminated. That is, the reception selection processing unit 14 controls the vehicle-to-vehicle communication unit 12 to receive 3D point cloud data from other vehicles within a distance of the threshold.

[0027] Next, a description will be given of the point cloud data merge process performed by the automatic driving device 10 according to the basic embodiment. Fig. 3 is a flowchart showing an example of the flow of the point cloud data merge process performed by the automatic driving device 10 according to the basic embodiment. Note that the process in Fig. 3 starts immediately after the reception selection process ends, for example.

[0028] In step 200, the vehicle-to-vehicle communication unit 12 receives position information from the other vehicle 26 through vehicle-to-vehicle communication, and the process proceeds to step 202.

[0029] In step 202, the point cloud data merge processing unit 16 offsets the point cloud data of the other vehicle by the amount of the relative positional relationship between the subject vehicle and the other vehicle 26, merges it with the point cloud data of the subject vehicle, and proceeds to step 204. That is, since the point cloud data of the other vehicle 26 is relative position point cloud data from the center of the other vehicle position, the relative positional relationship between the subject vehicle position and the other vehicle position is offset from the center position of the subject vehicle by the differences in latitude, longitude, and altitude, and then superimposed.

[0030] In step 204, the point cloud data merge processing unit 16 determines whether or not there is another vehicle 26 selected by the reception selection processing unit 14. If the determination is affirmative, the process proceeds to step 206, and if the determination is negative, the series of point cloud data merge processing is terminated.

[0031] In step 206, the vehicle-to-vehicle communication unit 12 switches the vehicle-to-vehicle communication to another vehicle 26 selected by the reception selection processing unit 14, and the process returns to step 200 to repeat the above-described processing.

[0032] In this way, in the automatic driving device 10 according to the basic form, 3D point cloud data is shared between vehicles, so information from other vehicles 26 received within the scope of the capability and responsibility of the ADS of the own vehicle can be utilized without being affected by the obstacle recognition and discrimination level of the other vehicles 26. This makes it possible to detect obstacles in blind spots that would prevent recognition using only the ADS sensor of the own vehicle.

[0033] (Embodiment) Next, an automatic driving device according to this embodiment will be described. The automatic driving device 10 according to this embodiment has the same configuration as the automatic driving device 10 according to the basic form, and only some of the functions are different, so only the differences will be described. Note that the inter-vehicle communication unit 12 in this embodiment corresponds to the receiving unit, and the reception selection processing unit corresponds to the control unit.

[0034] In the basic form, the reception selection processing unit 14 selects another vehicle 26 within a threshold distance from the own vehicle and receives the point cloud data of the other vehicle 26, but in this embodiment, the other vehicle 26 from which to receive point cloud data is selected based on the point cloud density of the point cloud data of the own vehicle.

[0035] Specifically, in the autonomous driving device 10 according to this embodiment, the reception selection processing unit 14 projects three-dimensional point cloud data detected by the vehicle itself onto a two-dimensional plane to calculate a point cloud density. Based on the position information of the other vehicle 26 received by the vehicle-to-vehicle communication unit 12, the reception selection processing unit 14 controls the vehicle-to-vehicle communication unit 12 to receive three-dimensional point cloud data detected by the other vehicle from the other vehicle in an area where the point cloud density is equal to or less than a predetermined threshold. Specifically, as shown in FIG. 4 , the three-dimensional point cloud data of the vehicle itself is projected onto a two-dimensional ground projection plane to calculate the point cloud density (number of points / unit volume) for each grid. The unit then calculates an angle at which an area where the point cloud density is equal to or less than a predetermined threshold exists, and acquires only the three-dimensional point cloud data of the other vehicle 26 located at the corresponding angle. Furthermore, the data acquisition history from the other vehicle 26 in each area is also included in the acquisition determination conditions so that point cloud data is acquired from only one other vehicle 26 in each area of ​​the point cloud density angle.

[0036] The unit volume used to calculate the point cloud density is preferably a volume space in the shape of a radiating fan-shaped Baumkuchen slice with the vehicle at the center, as shown in Figure 4, but a cubic grid arrangement type may also be used to facilitate calculation processing.

[0037] Next, specific processing performed by the automatic driving device 10 according to this embodiment will be described. As with the basic form, the automatic driving device 10 according to this embodiment also performs a reception selection process and a point cloud data merge process, but since the point cloud data merge process is the same as in the basic form, only the reception selection process will be described. Figure 5 is a flowchart showing an example of the flow of the reception selection process of the automatic driving device 10 according to this embodiment.

[0038] In step 300 , the reception selection processing unit 14 projects the three-dimensional point cloud data of the vehicle onto a two-dimensional plane, and then the process proceeds to step 302 .

[0039] In step 302 , the reception selection processing unit 14 calculates the point cloud density for each grid, and then the process proceeds to step 304 .

[0040] In step 304, the reception selection processing unit 14 calculates the angle at which the point cloud density is equal to or less than a predetermined threshold, and proceeds to step 306. That is, if the point cloud density is equal to or less than the threshold, the angle at which the point cloud density is equal to or less than the threshold is a blind spot from the view point of the vehicle, and therefore calculates the angle at which the blind spot occurs.

[0041] In step 306, the reception selection processing unit 14 receives the point cloud data from the other vehicle via vehicle-to-vehicle communication, and the process proceeds to step 308. That is, the reception selection processing unit 14 controls the vehicle-to-vehicle communication unit 12 to receive the three-dimensional point cloud data from the other vehicle.

[0042] In step 308, the reception selection processing unit 14 focuses on one other vehicle and proceeds to step 310.

[0043] In step 310, the reception selection processing unit 14 determines whether the position of the other vehicle 26 is within the point cloud rough angle. If the determination is negative, the process proceeds to step 312, and if the determination is affirmative, the process proceeds to step 314.

[0044] In step 312, the reception selection processing unit 14 changes the other vehicle 26 of interest, and returns to step 310 to repeat the above-described processing.

[0045] On the other hand, in step 314, the reception selection processing unit 14 determines whether there is no history of data acquisition from another vehicle 26 at the target point cloud rough angle. This determination is made by determining whether there is another vehicle 26 from which point cloud data has been acquired in each area of ​​the point cloud rough angle. If the determination is negative, the process proceeds to step 312, and if positive, the process proceeds to step 316.

[0046] In step 316, the reception selection processing unit 14 receives the point cloud data from the other vehicle 26 via vehicle-to-vehicle communication, and the process proceeds to step 318. That is, the reception selection processing unit 14 controls the vehicle-to-vehicle communication unit 12 to receive the three-dimensional point cloud data from the other vehicle.

[0047] In step 318, the reception selection processing unit 14 determines whether there is another vehicle 26. If the determination is affirmative, the process proceeds to step 312, and if the determination is negative, the series of reception selection processes ends.

[0048] By performing the reception selection process in this manner, it is possible to reduce the performance requirements of devices necessary for communication and calculation processing.

[0049] In the above embodiment, the position information and point cloud data of the other vehicle 26 are acquired through vehicle-to-vehicle communication, but the communication is not limited to vehicle-to-vehicle communication. For example, an ITS (Intelligent Transport Systems) on-board device capable of general-purpose transmission and reception, such as V2I (Vehicle-to-Roadside-Infrastructure), V2P (Vehicle-to-Pedestrian), or V2N (Vehicle-to-Network), may be applied.

[0050] Furthermore, the processing performed by the automated driving device 10 in each of the above embodiments may be software processing performed by executing a program, or may be hardware processing such as a GPU (Graphics Processing Unit), an ASIC (Application Specific Integrated Circuit), or an FPGA (Field-Programmable Gate Array). Alternatively, the processing may be a combination of both software and hardware. Furthermore, if the processing is software, the program may be stored in various storage media and distributed.

[0051] Furthermore, the present invention is not limited to the above, and it goes without saying that various modifications can be made without departing from the spirit of the present invention. [Explanation of symbols]

[0052] 10 Automatic driving device 12 Vehicle-to-vehicle communication unit (communication unit) 14 Reception selection processing unit (control unit)

Claims

[Claim 1] a receiving unit that receives position information of other vehicles and three-dimensional point cloud data detected by the other vehicles; a control unit that projects three-dimensional point cloud data detected by the vehicle onto a two-dimensional plane to calculate a point cloud density, and controls the receiving unit to receive three-dimensional point cloud data detected by another vehicle that is present in an area where the point cloud density is equal to or less than a predetermined threshold, based on the position information received by the receiving unit; An automatic driving device equipped with

Images