Object detection system

Abstract

To provide an object detection system that reduces data volume and enables high-speed communication via wireless communication when the results of object detection by multiple external sensors are collected and integrated by a server.SOLUTION: An object detection system 100 has an information processing unit 110 and a server 130. The information processing unit 110 has a map generator 111 that generates a grid map in a polar coordinate system based on the detection result of an object by an external sensor 11, and a data extractor 112 that extracts from the grid map specific point data including the dynamic diameter and a deflection angle of a specific point whose spatial distance changes. The server 130 has a data acquisition unit 131 that acquires the specified point data from the information processing unit 110 via wireless communication, and a map restoration unit 132 that restores the grid map based on the specified point data.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] This disclosure relates to an object detection system.

Background Art

[0002] Conventionally, inventions related to object detection devices have been known (see Patent Document 1 below). The object detection device described in Patent Document 1 detects an object existing around the host vehicle using a grid map that divides the periphery of the host vehicle into a plurality of grids. This object detection device includes a first object recognition unit, a second object recognition unit, an object integration unit, and an object identification unit (Patent Document 1, paragraph 0007, claim 1, and abstract).

[0003] The first object recognition unit outputs a probing wave and inputs sensing information of a ranging sensor that measures the position of the object by acquiring a reflected wave that has been reflected by the object. Also, the first object recognition unit uses the grid map to set a predetermined number of votes for a main grid that is a grid corresponding to the detection coordinates indicating the position where the object included in the sensing information of the ranging sensor is detected. Thereby, the first object recognition unit creates a first grid map indicating the positions of the objects existing around the host vehicle.

[0004] The second object recognition unit inputs sensing information of an in-vehicle camera that images the periphery of the host vehicle. Also, the second object recognition unit uses the grid map to set a predetermined number of votes for a main grid that is a grid corresponding to the detection coordinates indicating the position where the object included in the sensing information of the in-vehicle camera is detected. Thereby, the second object recognition unit creates a second grid map indicating the positions of the objects existing around the host vehicle.

[0005] The object integration unit creates an integrated grid map by integrating the number of votes for each grid in the first grid map and the number of votes for each grid in the second grid map. The object identification unit identifies, as the grid where the object exists, the grid in which the total value of the number of votes in the integrated grid map exceeds a predetermined threshold. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2021-135191 [Overview of the project] [Problems that the invention aims to solve]

[0007] The object detection device described in Patent Document 1 above can improve the accuracy of object detection by creating an integrated grid map that takes into account the characteristics of various sensors mounted on the vehicle. However, with this conventional object detection device, blind spots may occur around the vehicle where objects cannot be detected by the range-measuring sensors and on-board cameras. If an autonomous vehicle has blind spots around it where objects cannot be detected by external sensors, its operational design area (ODD) is limited, forcing it to slow down or stop to ensure safety, thus reducing the efficiency of autonomous driving.

[0008] To reduce blind spots in external sensors and expand the ODD (Object Detection Depth) of autonomous vehicles, it is conceivable to collect and share object detection results from external sensors on a server. However, object detection results from external sensors have a large data volume, posing challenges to communication speed when transmitting data via wireless communication. This disclosure provides an object detection system that reduces data volume and enables high-speed communication via wireless communication when collecting object detection results from external sensors on a server. [Means for solving the problem]

[0009] One aspect of the present disclosure is an object detection system comprising: an information processing device for acquiring object detection results from an external sensor; and a server connected to the information processing device via wireless communication, wherein the information processing device includes: a map generation unit that generates a grid map in polar coordinates with the position of the external sensor as the origin, including occupied areas where the object is detected and blank areas where the object is not detected; a data extraction unit that extracts specific point data from the grid map, including the radial and angular deviation of specific points whose spatial distance changes, which is the distance from the origin to the occupied area in an angular range where the occupied area exists, or the distance from the origin to the outer edge of the grid map in an angular range where the occupied area does not exist; and the server includes: a data acquisition unit that acquires the specific point data from the information processing device via wireless communication; and a map restoration unit that restores the grid map based on the specific point data. [Effects of the Invention]

[0010] According to one aspect of the present disclosure, an object detection system can be provided that reduces data volume and enables high-speed communication via wireless communication when collecting and integrating object detection results from multiple external sensors using a server. [Brief explanation of the drawing]

[0011] [Figure 1] A block diagram showing Embodiment 1 of the object detection system relating to this disclosure. [Figure 2] A flowchart illustrating an example of the operation of the in-vehicle information processing system shown in Figure 1. [Figure 3] A flowchart showing an example of the operation of the roadside information processing device shown in Figure 1. [Figure 4] A flowchart illustrating an example of server operation in Figure 1. [Figure 5] Figure 2 is an explanatory diagram illustrating an example of the data extraction process. [Figure 6] Figure 2 is an explanatory diagram illustrating an example of the data extraction process. [Figure 7] Figure 2 is an explanatory diagram illustrating an example of the data extraction process. [Figure 8] Explanatory diagram of an example of a process for generating a road-side grid map in FIG. 3. [Figure 9] Explanatory diagram of an example of a process for recognizing a blind spot and a boundary in FIG. 4. [Figure 10] Explanatory diagram of an example of a process for integrating the grid map in FIG. 4. [Figure 11] Explanatory diagram of an example of a process for integrating the grid map in FIG. 4. [Figure 12] Explanatory diagram of an example of a process for integrating the grid map in FIG. 4. [Figure 13] Grid map showing a modified example of the object detection system according to Embodiment 1. [Figure 14] Grid map showing a modified example of the object detection system according to Embodiment 1. [Figure 15] Grid map showing a modified example of the object detection system according to Embodiment 1. [Figure 16] Functional block diagram showing Embodiment 2 of the object detection system according to the present disclosure.

Mode for Carrying Out the Invention

[0012] Hereinafter, embodiments of the object detection system according to the present disclosure will be described with reference to the drawings.

[0013] [Embodiment 1] FIG. 1 is a functional block diagram showing Embodiment 1 of the object detection system according to the present disclosure. The object detection system 100 of the present embodiment includes, for example, a plurality of information processing devices 110 and 120 that acquire detection results of an object by a plurality of external sensors 11 and 20, and a server 130 that is connected to at least one of the information processing devices 110 via wireless communication.

[0014] In the example shown in FIG. 1, among the plurality of external sensors 11 and 20, at least one is, for example, the external sensor 11 mounted on the vehicle 10. Further, among the plurality of information processing devices 110 and 120, at least one is an in-vehicle information processing device 110 mounted on the vehicle 10 together with the external sensor 11 and acquiring the detection result of an object around the vehicle 10 by the external sensor 11. In FIG. 1, for the sake of illustration, only one in-vehicle information processing device 110 mounted on one vehicle 10 is shown among the in-vehicle information processing devices 110 mounted on each of the plurality of vehicles 10.

[0015] Each vehicle 10 on which the in-vehicle information processing device 110 is mounted is, for example, an autonomous vehicle capable of autonomous driving under the control of the in-vehicle information processing device 110 or another electronic control unit (ECU). Further, each vehicle 10 may be, for example, a manually driven vehicle or a semi-autonomous vehicle equipped with an advanced driver assistance system (ADAS). Each vehicle 10 is equipped with, for example, an external sensor 11, a vehicle sensor 12, a wireless communication device 13, a GNSS receiver 14, and various actuators 15.

[0016] The external sensor 11 includes, for example, at least one of a monocular camera, a stereo camera, a LiDAR (laser radar), or a millimeter-wave radar. The vehicle sensor 12 includes, for example, a speed sensor, an acceleration sensor, a steering angle sensor, an accelerator pedal sensor, a brake pedal sensor, a shift position sensor, a torque sensor, and the like. The wireless communication device 13 is communicably connected to a wireless base station WBS or the wireless communication device 13 of another vehicle 10 via a wireless communication line WCL, for example.

[0017] The GNSS receiver 14 receives radio waves from positioning satellites of a global navigation satellite system (GNSS), for example, and acquires the position information of the vehicle 10. The actuator 15 automatically operates an accelerator, a brake, a steering, a transmission, etc. based on a control signal output from the in-vehicle information processing device 110 or another ECU, for example, to autonomously drive the vehicle 10 or to assist the driving of the driver of the vehicle 10.

[0018] The in-vehicle information processing unit 110 is an in-vehicle ECU composed of, for example, one or more microcontrollers equipped with a central processing unit (CPU), memory, input / output unit, and timer. The in-vehicle information processing unit 110 includes, for example, a vehicle-side map generation unit 111 and a data extraction unit 112. Each of these parts of the in-vehicle information processing unit 110 represents a function of the in-vehicle information processing unit 110, which is realized, for example, by the CPU executing a program stored in memory. The operation of each of these parts of the in-vehicle information processing unit 110 will be described later with reference to Figure 2.

[0019] Furthermore, in the example shown in Figure 1, at least one of the multiple external sensors 11, 20 is, for example, a roadside sensor 20 installed around the road on which the vehicle 10 travels. Also, at least one of the multiple information processing devices 110, 120 is a roadside information processing device 120 that is wiredly connected to the roadside sensor 20 and the server 130. Note that in Figure 1, for illustrative purposes, only one roadside information processing device 120 connected to one of the multiple roadside information processing devices 120 connected to each of the multiple roadside sensors 20 is shown.

[0020] The roadside sensor 20 includes, for example, at least one of a stereo camera, a monocular camera, LiDAR, millimeter-wave radar, an ultrasonic sensor, or an infrared sensor, and detects objects around the roadside sensor 20, including a vehicle 10 traveling on the road. The roadside information processing device 120 acquires the object detection results from the roadside sensor 20 by wired communication, for example. The roadside information processing device 120 does not necessarily have to be directly connected to the roadside sensor 20, and may be connected to the roadside sensor 20 via a network N such as the Internet, for example.

[0021] The roadside information processing device 120 is, for example, a computer equipped with a CPU, memory, input / output unit, and timer, and is connected to the server 130 via a network N such as the Internet. The roadside information processing device 120 includes, for example, a roadside map generation unit 121 and a map transmission unit 122. Each of these parts of the roadside information processing device 120 represents a function of the roadside information processing device 120 that is realized, for example, by executing a program stored in memory using the CPU. The operation of each of these parts of the roadside information processing device 120 will be described later with reference to Figure 3.

[0022] Server 130 is, for example, a computer equipped with a CPU, memory, input / output unit, and timer, and is a network server connected to a network N such as the Internet. Server 130 is, for example, connected to roadside sensors 20 and roadside information processing device 120 via a wired communication line for information communication, and connected to in-vehicle information processing device 110 via a wireless communication line WCL for information communication.

[0023] The server 130 includes, for example, a data acquisition unit 131 and a map restoration unit 132. The server 130 may also further include, for example, a map integration unit 133, and a blind spot recognition unit 134 and a boundary recognition unit 135. Each of these parts of the server 130 represents a function of the server 130, which is realized, for example, by executing a program stored in memory using the CPU. The operation of each of these parts of the server 130 will be described later with reference to Figure 4.

[0024] The operation of the object detection system 100 of this embodiment will be described below with reference to Figures 2 to 4. Figure 2 is a flowchart showing an example of the operation of the in-vehicle information processing device 110 in Figure 1. Figure 3 is a flowchart showing an example of the operation of the roadside information processing device 120 in Figure 1. Figure 4 is a flowchart showing an example of the operation of the server 130 in Figure 1. Figures 5 to 7 are explanatory diagrams illustrating examples of the process P11 for generating the vehicle-side grid map in Figure 2 and the data extraction process P12, respectively.

[0025] Each in-vehicle information processing device 110 constituting the object detection system 100 of this embodiment, for example, when the processing flow PF1 shown in Figure 2 is started, first executes a process P11 that generates a vehicle-side grid map. In this process P11, the vehicle-side map generation unit 111 of the in-vehicle information processing device 110 generates a grid map GM1 in polar coordinates with the position of the external sensor 11 as the origin O1, based on the object detection result of the external sensor 11, for example, as shown in Figures 5 to 7.

[0026] More specifically, the vehicle-side map generation unit 111 obtains the distance and direction from the external sensor 11 to objects around vehicle 10, such as other vehicles V1, V2, detected by the external sensor 11. Then, as shown in Figures 5 to 7, the vehicle-side map generation unit 111 generates a vehicle-side grid map GM1 that includes occupied areas OS1 occupied by objects detected by the external sensor 11 and blank areas VS1 where no objects were detected by the external sensor 11. The outer edge OE1 of the grid map GM1 corresponds, for example, to the outer edge of the detection range of the external sensor 11.

[0027] Next, the in-vehicle information processing device 110 performs a data extraction process P12, for example, as shown in Figure 2. In this process P12, the data extraction unit 112 extracts specific point data, including the radial and angular deviations of a specific point P1 where the spatial distance SD changes, from the grid map GM1 shown in Figures 5 to 7. Here, the spatial distance SD is, for example, the distance from the origin O1 to the occupied area OS1 in the angular range where the occupied area OS1 exists, and the distance from the occupied area OS1 to the outer edge OE1 of the grid map GM1 in the angular range where the occupied area OS1 does not exist.

[0028] More specifically, the data extraction unit 112 scans the spatial distance SD with the angle decreasing direction Dd or the angle increasing direction Di around the origin O1 of the grid map GM1 as the scanning direction, for example, as shown in Figure 5 or Figure 6. The data extraction unit 112 then extracts points where the spatial distance SD changes before and after the scanning direction as specific points P1, and generates specific point data including the radial and angular deviation of specific points P1 and the scanning direction.

[0029] Alternatively, the data extraction unit 112 may scan the spatial distance SD using the angle increasing direction Di and the angle decreasing direction Dd of the grid map GM1 as scanning directions, for example, as shown in Figure 7. In this case as well, the data extraction unit 112 can extract as specific points P1 the points where the spatial distance SD changes before and after each scanning direction of the angle increasing direction Di and the angle decreasing direction Dd.

[0030] Next, the in-vehicle information processing device 110 executes, for example, the specific point data transmission process P13 shown in Figure 2. In this process P13, the data extraction unit 112 transmits, for example, the specific point data extracted from the grid map GM1 in the previous process P12 to the server 130 via, for example, the vehicle's wireless communication device 13, the wireless communication line WCL, the wireless base station WBS, and the network N. After that, the in-vehicle information processing device 110 terminates, for example, the processing flow PF1 shown in Figure 2 and repeats at a predetermined cycle.

[0031] Figure 8 is an explanatory diagram illustrating an example of the process P21 for generating the roadside grid map shown in Figure 3. Each roadside information processing device 120 constituting the object detection system 100 of this embodiment, for example, when the processing flow PF2 shown in Figure 3 is started, first executes the process P21 for generating the roadside grid map. In this process P21, the roadside map generation unit 121 of the roadside information processing device 120 generates a roadside grid map GM2 in polar coordinates with the position of the roadside sensor 20 as the origin O2, based on the object detection result by the roadside sensor 20, for example, as shown in Figure 8.

[0032] More specifically, the roadside map generation unit 121 obtains the distance and direction from the roadside sensor 20 to objects around vehicle 10, such as other vehicles V3 and V4, detected by the roadside sensor 20. Then, as shown in Figure 8, the roadside map generation unit 121 generates a roadside grid map GM2 that includes occupied areas OS2 occupied by objects detected by the roadside sensor 20 and blank areas VS2 where no objects were detected by the roadside sensor 20.

[0033] Next, the roadside information processing device 120 executes, for example, the map transmission process P22 shown in Figure 3. In this process P22, the map transmission unit 122 of the roadside information processing device 120 transmits, for example, the roadside grid map GM2 generated by the roadside map generation unit 121 in the previous process P21, to the server 130 via wired communication. That is, the communication line between the roadside information processing device 120 and the server 130 is, for example, a wired communication line that does not include a wireless communication line WCL. After the completion of this process P22, the roadside information processing device 120 terminates, for example, the processing flow PF2 shown in Figure 3, and repeats at a predetermined cycle.

[0034] The roadside information processing device 120 may be connected to the server 130 via a wireless communication line WCL, similar to the in-vehicle information processing device 110. In this case, the roadside information processing device 120 may include a data extraction unit similar to the data extraction unit 112 of the in-vehicle information processing device 110. In this case, the data extraction unit of the roadside information processing device 120 may extract a specific point P2 from the roadside grid map GM2 that is the same as a specific point P1 in the vehicle-side grid map GM1, and transmit specific point data, including the radial and angular deviation of the specific point P2, to the server 130 via wireless communication.

[0035] Furthermore, when the server 130, which constitutes the object detection system 100 of this embodiment, starts the processing flow PF3 shown in Figure 4, for example, it first executes a process P31 to acquire specific point data. In this process P31, the data acquisition unit 131 of the server 130 acquires specific point data from the in-vehicle information processing device 110 via wireless communication, for example.

[0036] More specifically, in process P13 shown in Figure 2, specific point data is transmitted from each in-vehicle information processing device 110, for example, via the wireless communication device 13, the wireless communication line WCL, the wireless base station WBS, and the network N. In process P31 shown in Figure 4, the data acquisition unit 131 of the server 130 acquires the specific point data transmitted from the in-vehicle information processing device 110 via wireless communication, for example, via the network N.

[0037] Here, the specific point data acquired by the data acquisition unit 131 includes, for example, the radial and angular deviation of each specific point P1 shown in Figure 5, Figure 6, or Figure 7. Also, as shown in Figure 5 or Figure 6, if the scanning direction of the spatial distance SD by the data extraction unit 112 is in the angle increasing direction Di or the angle decreasing direction Dd, the specific point data acquired by the data acquisition unit 131 includes the scanning direction of the spatial distance SD by the data extraction unit 112. Note that, as shown in Figure 7, if the scanning direction of the spatial distance SD by the data extraction unit 112 is in both the angle increasing direction Di and the angle decreasing direction Dd, the specific point data acquired by the data acquisition unit 131 does not need to include the scanning direction of the spatial distance SD by the data extraction unit 112.

[0038] After the completion of process P31, which acquires specific point data as shown in Figure 4, the server 130 executes process P32, for example, to restore the vehicle-side grid map GM1. In this process P32, the map restoration unit 132 generates a restored grid map RGM1 (see Figure 9) by restoring the vehicle-side grid map GM1 based on the specific point data acquired in the previous process P31.

[0039] More specifically, in the example shown in Figure 5 or Figure 6, the specific point data extracted by the data extraction unit 112 of the in-vehicle information processing device 110 includes the radial and angular deviation of each specific point P1, and the scanning direction of the spatial distance SD (angle decreasing direction Dd or angle increasing direction Di). In this case, the map restoration unit 132 plots, for example, multiple specific points P1 included in the specific point data on a restored grid map RGM1 in polar coordinates.

[0040] Furthermore, the map restoration unit 132 sequentially restores the occupied area OS1 and the blank area VS1 in the scanning direction of the spatial distance SD by the data extraction unit 112, from one specific point P1 to the next specific point P1 of the restored grid map RGM1. For example, the map restoration unit 132 starts from any specific point P1 plotted on the outer edge OE1 of the restored grid map RGM1 and restores the blank area VS1 in the scanning direction to the next specific point P1.

[0041] If the radius of the next specified point P1 is shorter than the radius of the specified point P1 on the outer edge OE1, the map restoration unit 132 restores the occupied area OS1 in the scanning direction from that specified point P1 to the next specified point P1. If the radius of the next specified point P1 is shorter than the radius of the specified point P1 on the outer edge OE1, the map restoration unit 132 restores the occupied area OS1 in the scanning direction from that specified point P1 to the next specified point.

[0042] On the other hand, if the next specific point P1 on the occupied area OS1 is a specific point P1 on the outer edge OE1, the map restoration unit 132 restores the blank area VS1 in the scanning direction from that specific point P1 to the next specific point P1. By performing the restoration process described above at a 360-degree angle in the scanning direction from the first specific point P1, the map restoration unit 132 can generate a restored grid map RGM1 by restoring the vehicle-side grid map GM1 shown in Figure 5 or Figure 6 based on the specific point data.

[0043] Furthermore, in the example shown in Figure 7, the specific point data includes at least the radius vector and angle of each specific point P1. Therefore, the map restoration unit 132 restores the occupied area OS1 between specific points P1 on the outer edge OE1 whose radius vector is shorter than that of a specific point P1 on the outer edge OE1 and whose radius vectors are equal to those of two adjacent points P1 in the angular direction. The map restoration unit 132 also restores blank areas VS1 between specific points P1 on the outer edge OE1 that are outside the angular range of the restored occupied area OS1. As a result, the map restoration unit 132 can generate a restored grid map RGM1 by restoring the vehicle-side grid map GM1 shown in Figure 7 based on the specific point data.

[0044] After the completion of process P32, which restores the vehicle-side grid map GM1 shown in Figure 4, the server 130 executes process P33, for example, to acquire the road-side grid map GM2. In this process P33, the data acquisition unit 131 of the server 130 receives the road-side grid map GM2, which was transmitted from the map transmission unit 122 of the roadside information processing device 120 in process P22 shown in Figure 3, via wired communication that does not include the wireless communication line WCL.

[0045] Figure 9 is an explanatory diagram illustrating an example of the blind spot recognition process P34 and boundary recognition process P35 shown in Figure 4. For example, after the completion of the process P33 for acquiring the road-side grid map GM2 shown in Figure 4, the server 130 executes the blind spot recognition process P34. In this process P34, the blind spot recognition unit 134 of the server 130 recognizes, for example, as shown in Figure 9, any area within the angular range of the occupied area OS1 of the restored grid map RGM1 that is further from the origin O1 than the occupied area OS1 as a blind spot area BS1. Here, the area outside the outer edge OE1 of the grid map GM1 can also be considered a blind spot area BS1.

[0046] Furthermore, the server 130 recognizes blind spots in the roadside grid map GM2 obtained from the roadside information processing device 120, for example, in the same way as the blind spot section BS1 of the restored grid map RGM1. Subsequently, the server 130 executes a boundary recognition process P35, for example, as shown in Figure 4. In this process P35, the boundary recognition unit 135 of the server 130 recognizes the boundary BL1 between the blind spot section BS1 and the blank section VS1 of the restored grid map RGM1, for example, as shown in Figure 9.

[0047] Furthermore, the boundary recognition unit 135 may recognize, for example, the boundary BL1 between the occupied area OS1 and the blank area VS1 of the restored grid map RGM1, and the boundary BL1 on the outer edge OE1 of the restored grid map RGM1. The boundary recognition unit 135 also recognizes boundaries in the roadside grid map GM2 acquired from the roadside information processing device 120, for example, in the same way as the boundary BL1 of the restored grid map RGM1.

[0048] The boundary BL1 of the restored grid map RGM1 is represented, for example, by the distance and angle from the origin O1, i.e., the radius vector and angle of the intersection of each line or arc that constitutes the boundary BL1. Similarly, the boundary of the road-side grid map GM2 is represented by the distance and angle from the origin O1 of the road-side grid map GM2, i.e., the radius vector and angle of the intersection of each line or arc that constitutes the boundary.

[0049] Figures 10 to 12 are explanatory diagrams illustrating an example of the process P36 for integrating the grid maps in Figure 4. For example, after the completion of the boundary recognition process P35 shown in Figure 4, the server 130 executes the grid map integration process P36. In this process P36, the map integration unit 133 of the server 130 integrates, for example, the restored grid map RGM1 on the vehicle side restored by the map restoration unit 132 and the road side grid map GM2 acquired by the data acquisition unit 131.

[0050] Here, the map integration unit 133 integrates the restored grid map RGM1 and the road-side grid map GM2 based, for example, on the coordinates of the origins O1 and O2 of the vehicle-side restored grid map RGM1 and the road-side grid map GM2, and the directions of the reference angles θ1 and θ2, respectively. More specifically, the map integration unit 133 calculates the direction of the reference angle θ1 of the vehicle-side restored grid map RGM1 based, for example, on the attitude of the external sensor 11 included in the specific point data acquired from the in-vehicle information processing device 110. The map integration unit 133 also calculates the direction of the reference angle θ2 of the road-side grid map GM2 based, for example, on the attitude of the road-side sensor 20 transmitted from the map transmission unit 122 of the road-side information processing device 120.

[0051] Furthermore, the map integration unit 133 converts, for example, the polar coordinate system restored grid map RGM1 shown in Figure 9 into the Cartesian coordinate system vehicle-side grid map CGM1 shown in Figure 10, based on the coordinates of the origin O1 and the direction of the reference angle θ1. Similarly, the map integration unit 133 converts, for example, the polar coordinate system road-side grid map GM2 shown in Figure 8 into a Cartesian coordinate system road-side grid map, based on the coordinates of the origin O2 and the direction of the reference angle θ2.

[0052] The map integration unit 133 may, for example, convert the polar coordinate system restored grid map RGM1 to a Cartesian coordinate system vehicle-side grid map CGM1 based on a direction such as north (N), as shown in Figure 11. In this case, the map integration unit 133 also converts the polar coordinate system road-side grid map GM2 to a Cartesian coordinate system road-side grid map based on the same direction. Furthermore, the map integration unit 133 integrates the converted Cartesian coordinate system vehicle-side grid map CGM1 and road-side grid map to generate a Cartesian coordinate system integrated grid map UGM as shown in Figure 12.

[0053] As a result, the map integration unit 133 can recognize a blind spot area BS1 in the vehicle-side grid map CGM1, which is a blank area in the road-side grid map, as a compensated blank area CVS in the integrated grid map UGM. Similarly, the map integration unit 133 can recognize a blind spot area in the road-side grid map, which is a blank area VS1 in the vehicle-side grid map CGM1, as a compensated blank area CVS in the integrated grid map UGM. Consequently, in the integrated grid map UGM, the blind spot area BS around the vehicle 10 decreases, and the blank area VS and compensated blank area CVS around the vehicle 10 expand.

[0054] Subsequently, the server 130 executes a process P37 that shares the integrated grid map UGM shown in Figure 4. In this process P37, the map integration unit 133 of the server 130 extracts, for example, the boundaries between the blank area VS and the compensated blank area CVS, and the occupied area OS and the blind spot area BS from the generated integrated grid map UGM. Furthermore, the map integration unit 133 transmits the extracted boundary information to the in-vehicle information processing device 110 of each vehicle 10 via wireless communication. After that, the server 130 terminates the processing flow PF3 shown in Figure 4, for example, and repeats it at a predetermined cycle.

[0055] The operation of the object detection system 100 of this embodiment will be described below.

[0056] As described above, the object detection system 100 of this embodiment includes an information processing device 110 that acquires the object detection results from the external sensor 11, and a server 130 that is connected to the information processing device 110 via wireless communication. The information processing device 110 includes a map generation unit 111 and a data extraction unit 112. Based on the object detection results from the external sensor 11, the map generation unit 111 generates a grid map GM1 in polar coordinates, with the position of the external sensor 11 as the origin O1, and including occupied areas OS1 where objects are detected and blank areas VS1 where no objects are detected. The data extraction unit 112 extracts specific point data from the grid map GM1, including the radial movement and angular deviation of a specific point P1 where the spatial distance SD changes. Here, the spatial distance SD is the distance from the origin O1 to the occupied area OS1 in the angular range in which the occupied area OS1 exists, or the distance from the origin O1 to the outer edge OE1 of the grid map GM1 in the angular range in which the occupied area OS1 does not exist. The server 130 includes a data acquisition unit 131 that acquires specific point data from the information processing device 110 via wireless communication, and a map restoration unit 132 that restores the grid map GM1 based on the specific point data.

[0057] With this configuration, the information processing device 110 generates a grid map GM1 in polar coordinates based on the detection results of the external sensor 11 using the map generation unit 111, and extracts specific point data, including the radial movement and declination angle of a specific point P1, using the data extraction unit 112. The data capacity of this specific point data is significantly reduced compared to the data capacity of the grid map GM1. Furthermore, the server 130 acquires the specific point data from the information processing device 110 via wireless communication using the data acquisition unit 131, and restores the grid map GM1 based on the specific point data using the map restoration unit 132. In this way, by the server 130 acquiring the specific point data from the information processing device 110 and restoring the grid map GM1, the amount of data transmitted via wireless communication is significantly reduced compared to when the grid map GM1 is acquired from the in-vehicle information processing device 110. As a result, for example, high-speed communication using 4G communication, which is a common wireless communication method for vehicles, can be enabled.

[0058] Furthermore, in the object detection system 100 of this embodiment, the information processing device 110 is mounted on the vehicle 10 together with the external sensor 11 and is an on-board information processing device 110 that acquires the detection results of objects around the vehicle 10 by the external sensor 11. The object detection system 100 also further includes a roadside information processing device 120 that is wiredly connected to a roadside sensor 20 installed around the road on which the vehicle 10 travels and to a server 130 and acquires the detection results of objects by the roadside sensor 20. The roadside information processing device 120 has a roadside map generation unit 121 and a map transmission unit 122. Based on the object detection results by the roadside sensor 20, the roadside map generation unit 121 generates a roadside grid map GM2 in polar coordinates, with the position of the roadside sensor 20 as the origin, and including occupied areas OS2 where objects were detected and blank areas VS2 where no objects were detected. The map transmission unit 122 transmits the roadside grid map GM2 to the server 130 via wired communication. Server 130 further includes a map integration unit 133 that integrates the vehicle-side grid map RGM1 restored by the map restoration unit 132 and the road-side grid map GM2 acquired by the data acquisition unit 131.

[0059] With this configuration, the on-board information processing unit 110 mounted on each vehicle 10 can transmit specific point data extracted from the grid map GM1 based on the object detection results from the external sensors 11 of each vehicle 10 to the server 130 via wireless communication. Furthermore, the roadside information processing unit 120, which is wired to each roadside sensor 20, can transmit the roadside grid map GM2 based on the surrounding object detection results from the roadside sensors 20 to the server 130 via wired communication. In addition, the server 130 can integrate the vehicle-side grid map GM1 reconstructed from the specific point data with the roadside grid map GM2 obtained from the roadside information processing unit 120, thereby reducing the blind spot area BS around the vehicle 10. As a result, the operational design domain (ODD) of the autonomous vehicle can be expanded, preventing unnecessary deceleration and stopping of the vehicle 10 and improving the efficiency of autonomous driving.

[0060] Furthermore, in the object detection system 100 of this embodiment, the server 130 further includes a blind spot recognition unit 134 and a boundary recognition unit 135. The blind spot recognition unit 134 recognizes a blind spot area BS1 as an area that is further from the origin O1 than the occupied area OS1 within the angular range of the occupied area OS1. The boundary recognition unit 135 recognizes the boundary BL1 between the blind spot area BS1 and the empty area VS1.

[0061] With this configuration, the server 130 transmits boundary BL1 information of the grid map GM1 to each in-vehicle information processing unit 110 via wireless communication, and the in-vehicle information processing unit 110 can use boundary BL1 to reconstruct the grid map GM1, including the blind spot area BS1 and the blank area VS1. Therefore, compared to transmitting the grid map GM1 from the server 130 to the in-vehicle information processing unit 110, the data volume can be reduced, and high-speed communication becomes possible in wireless communication where communication capacity is limited.

[0062] Furthermore, in the object detection system 100 of this embodiment, the boundary BL1 is represented by the distance and angle from the origins O1 and O2 of the vehicle-side grid map GM1 and the road-side grid map GM2, respectively.

[0063] With this configuration, the server 130 transmits boundary BL1 information of the grid map GM1 to each in-vehicle information processing unit 110 via wireless communication, and the in-vehicle information processing unit 110 can use boundary BL1 to reconstruct the grid map GM1, including the blind spot area BS1 and the blank area VS1. Therefore, compared to transmitting the grid map GM1 from the server 130 to the in-vehicle information processing unit 110, the data volume can be reduced, and high-speed communication becomes possible in wireless communication where communication capacity is limited.

[0064] Furthermore, in the object detection system 100 of this embodiment, the data extraction unit 112 scans the spatial distance SD with the angle increasing direction Di or the angle decreasing direction Dd around the origin O1 of the grid map GM1 as the scanning direction. The data extraction unit 112 then extracts points where the spatial distance SD changes before and after the scanning direction as specific points P1, and generates specific point data including the scanning direction. With this configuration, the server 130 that has acquired the specific point data can generate a restored grid map RGM1 by restoring the grid map GM1 based on the radial and angular deviation of the specific points P1 included in the specific point data, and the scanning direction.

[0065] Furthermore, in the object detection system 100 of this embodiment, the data extraction unit 112 scans the spatial distance SD with the angle increasing direction Di and angle decreasing direction Dd of the grid map GM1 as the scanning directions. The data extraction unit 112 then extracts points where the spatial distance SD changes before and after the scanning direction as specific points P1. With this configuration, the server 130 that has acquired the specific point data can generate a restored grid map RGM1 by restoring the grid map GM1 based only on the radial movement and angular deviation of the specific point P1 included in the specific point data.

[0066] Furthermore, in the object detection system 100 of this embodiment, the map integration unit 133 integrates the vehicle-side grid map GM1 and the road-side grid map GM2 based on the coordinates of their respective origins O1 and O2 and the directions of their reference angles θ1 and θ2. With this configuration, an integrated grid map UGM is generated by integrating the grid maps GM1 and GM2 based on the object detection results from multiple different external sensors 11 and 20, thereby reducing the blind spot area BS around the vehicle 10.

[0067] Furthermore, in the object detection system 100 of this embodiment, the map integration unit 133 calculates the direction of the reference angle θ1 of the vehicle-side grid map GM1 based on the orientation of the external sensor 11 included in the specific point data. The map integration unit 133 also calculates the direction of the reference angle θ2 of the road-side grid map GM2 based on the orientation of the roadside sensor 20 transmitted from the map transmission unit 122. With this configuration, grid maps GM1 and GM2 based on the object detection results from multiple different external sensors 11 and 20 can be integrated to reduce the blind spot area BS around the vehicle 10.

[0068] Furthermore, in the object detection system 100 of this embodiment, the map restoration unit 132 plots a plurality of specific points P1 included in the specific point data acquired by the data acquisition unit 131 on a restored grid map RGM1 in polar coordinates, and sequentially restores the occupied area OS and blank area VS in the scanning direction from one specific point P1 to the next specific point P1 of the restored grid map RGM1. With this configuration, the restored grid map RGM1 can be generated based on the radial movement and declination angle of the specific point P1, and the scanning direction.

[0069] As described above, according to this embodiment, when the server 130 collects and integrates the object detection results from multiple external sensors 11, 20, it is possible to provide an object detection system 100 that reduces data capacity and enables high-speed communication via wireless communication. Note that the configuration of the object detection system 100 in this embodiment is not limited to the configuration shown in Figures 1 to 12.

[0070] Figures 13 to 15 show grid maps GM1a, GM1b, and UGM illustrating modified versions of the object detection system 100 of Embodiment 1. In these modified versions, the vehicle 10 is, for example, a truck or trailer, and external sensors 11 are mounted on the front and rear ends of the vehicle 10, respectively. In these modified versions, the vehicle-side map generation unit 111 of the in-vehicle information processing device 110 acquires the detection results of objects such as vehicles V1 and V2 from the external sensors 11 at the front and rear ends of the vehicle 10, respectively.

[0071] Furthermore, as shown in Figures 13 and 14, the vehicle-side map generation unit 111 generates polar coordinate grid maps GM1a and GM1b, respectively, with the positions of the external sensors 11 as origins O1a and O1b, and including occupied areas OS1a and OS1b where objects are detected, and blank areas VS1a and VS1b where no objects are detected. The data extraction unit 112 of the in-vehicle information processing device 110 also extracts specific point data from each grid map GM1a and GM1b, including the radial and angular deviations of specific points P1a and P1b where the spatial distance SD changes.

[0072] The data acquisition unit 131 of the server 130 acquires specific point data for each grid map GM1a and GM1b from the in-vehicle information processing device 110 via wireless communication. Furthermore, the map restoration unit 132 of the server 130 restores each grid map GM1a and GM1b based on the specific point data for each grid map GM1a and GM1b. Then, the map integration unit 133 of the server 130 integrates the grid maps GM1a and GM1b restored by the map restoration unit 132 to generate an integrated grid map UGM as shown in Figure 15.

[0073] In this modified example, in the integrated grid map UGM, a portion of the blind spot area BS based on the detection result of one external sensor 11 becomes a blank area CVS compensated by the blank area VS based on the detection result of the other external sensor 11. Therefore, in this modified example as well, the object detection system 100 can be provided by collecting and integrating the object detection results of multiple external sensors 11 using the server 130, reducing the blind spots of the external sensors 11, reducing data capacity, and enabling high-speed communication via wireless communication.

[0074] [Embodiment 2] Figure 16 is a functional block diagram of a server 130 showing Embodiment 2 of the object detection system according to the present disclosure. The object detection system 100 of this embodiment differs from the object detection system 100 of Embodiment 1 described above in that it does not have the roadside information processing device 120 shown in Figure 1, and the server 130 has a roadside sensor data acquisition unit 136 and a roadside map generation unit 137 instead of a blind spot recognition unit 134 and a boundary recognition unit 135. Other aspects of the object detection system 100 of this embodiment are the same as those of the object detection system 100 of Embodiment 1 described above, so the same parts are denoted by the same reference numerals and their description is omitted.

[0075] In the object detection system 100 of this embodiment, the information processing device 110 is mounted on the vehicle 10 together with the external sensor 11 and is an in-vehicle information processing device 110 that acquires the detection results of objects around the vehicle 10 by the external sensor 11. The server 130 has a data acquisition unit 131, a map restoration unit 132, a map integration unit 133, a roadside sensor data acquisition unit 136, and a roadside map generation unit 137. The roadside sensor data acquisition unit 136 acquires the detection results of objects from roadside sensors 20 that are installed around the road on which the vehicle 10 travels and are wired to the server 130. As shown in Figure 8, the roadside map generation unit 137 generates a roadside grid map GM2 in polar coordinates, with the position of the roadside sensor 20 as the origin O2, and including occupied areas OS2 where objects were detected and blank areas VS2 where no objects were detected. The map integration unit 133 integrates the restored vehicle-side grid map RGM1 and the roadside grid map GM2.

[0076] With this configuration, the object detection system 100 of this embodiment can achieve the same effects as the object detection system 100 of Embodiment 1 described above.

[0077] While embodiments of the object detection system relating to this disclosure have been described in detail above using drawings, the specific configuration is not limited to these embodiments, and any design changes, etc., that do not depart from the gist of this disclosure are also included in this disclosure. [Explanation of Symbols]

[0078] 10 Vehicle, 11 External sensor, 20 Roadside sensor (external sensor), 100 Object detection system, 110 In-vehicle information processing unit (information processing unit), 111 Vehicle-side map generation unit (map generation unit), 112 Data extraction unit, 120 Roadside information processing unit (information processing unit), 121 Roadside map generation unit (map generation unit), 122 Map transmission unit, 130 Server, 131 Data acquisition unit, 132 Map restoration unit, 133 Map integration unit, 134 Blind spot recognition unit, 135 Boundary recognition unit, 136 Roadside sensor data acquisition unit, 137 Roadside map generation unit, BL1 Boundary, Dd Angle decrease direction (scanning direction), Di Angle increase direction (scanning direction), GM1 Grid map, GM2 Roadside grid map (grid map), O1 Origin, O2 Origin, OS Occupied area, OS1 Occupied area, OS2 Occupied area, SD Spatial distance, UGM integrated grid map (grid map), V1 vehicle (object), V2 vehicle (object), V3 vehicle (object), V4 vehicle (object), VS blank area, VS1 blank area, VS2 blank area, P1 specific point, P1a specific point, P1b specific point, P2 specific point, θ1 reference angle, θ2 reference angle.

Claims

1. An object detection system comprising an information processing device that acquires the detection results of an object by an external sensor, and a server connected to the information processing device via wireless communication, The information processing device includes: a map generation unit that generates a grid map in polar coordinates, with the position of the external sensor as the origin, based on the object detection result by the external sensor, and including occupied areas where the object is detected and blank areas where the object is not detected; and a data extraction unit that extracts specific point data from the grid map, including the radial and angular deviation of specific points where the spatial distance changes, which is the distance from the origin to the occupied area in the angular range where the occupied area exists, or the distance from the origin to the outer edge of the grid map in the angular range where the occupied area does not exist. The server includes a data acquisition unit that acquires the specific point data from the information processing device via wireless communication, and a map restoration unit that restores the grid map based on the specific point data. The object detection system is characterized in that the data extraction unit scans the spatial distance with the scanning direction being the direction of increasing or decreasing angle around the origin of the grid map, extracts points where the spatial distance changes before and after the scanning direction as specific points, and generates specific point data including the scanning direction.

2. The object detection system according to claim 1, characterized in that the map restoration unit plots a plurality of specific points included in the specific point data acquired by the data acquisition unit onto a restored grid map in polar coordinates, and sequentially restores the occupied area and the blank area in the scanning direction from one specific point to the next specific point on the restored grid map.

3. The aforementioned information processing device is an in-vehicle information processing device that is mounted on the vehicle together with the external sensor and acquires the detection results of objects around the vehicle by the external sensor. The object detection system further comprises roadside sensors installed around the road on which the vehicle travels and a roadside information processing device that is wired to the server and acquires the object detection results from the roadside sensors. The roadside information processing device includes a roadside map generation unit that generates a roadside grid map in polar coordinates, based on the object detection results by the roadside sensor, with the position of the roadside sensor as the origin, and including occupied areas where the object was detected and blank areas where the object was not detected; and a map transmission unit that transmits the roadside grid map to the server via wired communication. The object detection system according to claim 1, further comprising a map integration unit that integrates the vehicle-side grid map restored by the map restoration unit with the road-side grid map acquired by the data acquisition unit.

4. The aforementioned information processing device is an in-vehicle information processing device that is mounted on the vehicle together with the external sensor and acquires the detection results of objects around the vehicle by the external sensor. The object detection system according to claim 1, further comprising: a roadside sensor data acquisition unit that acquires object detection results from roadside sensors installed around the road on which the vehicle travels and connected to the server by wire; a roadside map generation unit that generates a roadside grid map in polar coordinates, with the position of the roadside sensor as the origin, and including occupied areas where the object is detected and blank areas where the object is not detected, based on the object detection results of the roadside sensor; and a map integration unit that integrates the grid map on the vehicle side and the roadside grid map.

5. The object detection system according to claim 3 or 4, further comprising: a server; a blind spot recognition unit that recognizes a section further from the origin than the occupied section within the angular range of the occupied section as a blind spot section; and a boundary recognition unit that recognizes the boundary between the blind spot section and the blank section.

6. The object detection system according to claim 5, characterized in that the boundary is represented by the distance and angle from the origin of the vehicle-side grid map and the road-side grid map, respectively.

7. The object detection system according to claim 3, characterized in that the map integration unit integrates the grid map on the vehicle side and the grid map on the road side based on the coordinates of the origin and the direction of the reference angle of the respective grid maps on the vehicle side and the road side.

8. The object detection system according to claim 7, characterized in that the map integration unit calculates the direction of the reference angle of the vehicle-side grid map based on the orientation of the external sensor included in the specific point data, and calculates the direction of the reference angle of the road-side grid map based on the orientation of the road-side sensor transmitted from the map transmission unit.

9. An object detection system comprising an information processing device that acquires object detection results from external sensors, and a server connected to the information processing device via wireless communication, The information processing device includes: a map generation unit that generates a grid map in polar coordinates, with the position of the external sensor as the origin, based on the object detection result by the external sensor, and including occupied areas where the object is detected and blank areas where the object is not detected; and a data extraction unit that extracts specific point data from the grid map, including the radial and angular deviation of specific points where the spatial distance changes, which is the distance from the origin to the occupied area in the angular range where the occupied area exists, or the distance from the origin to the outer edge of the grid map in the angular range where the occupied area does not exist. The server includes a data acquisition unit that acquires the specific point data from the information processing device via wireless communication, and a map restoration unit that restores the grid map based on the specific point data. The object detection system is characterized in that the data extraction unit scans the spatial distance with the angle increasing direction and angle decreasing direction of the grid map as the scanning directions, and extracts points where the spatial distance changes before and after the scanning direction as the specific points.

Images