Vehicle control device, vehicle control method, and program

Abstract

To suppress a possibility that control of an own vehicle is inappropriately executed when a limit detection distance of a surrounding situation sensor is short.SOLUTION: A vehicle control device 14 includes an inference section 3B for inferring a limit detection distance which is a maximum value of a distance from a surrounding situation sensor 11 mounted on an own vehicle 1, detectable by the surrounding situation sensor 11. The inference section 3B infers the limit detection distance of the surrounding situation sensor 11 on the basis of sensor data of the surrounding situation sensor 11 by using a machine learning model obtained by performing learning using teacher data which is a data set of sensor data of a learning surrounding situation sensor L11 mounted on a learning vehicle L1 and a label indicating a limit detection distance of the learning surrounding situation sensor L11 when the sensor data of the learning surrounding situation sensor L11 is obtained.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a vehicle control device, a vehicle control method, and a program. [Background technology]

[0002] Patent document 1 describes a technology in which a vehicle distance warning device equipped with a rainfall amount detection means for detecting the amount of rainfall changes the settings of the vehicle distance warning and ACC according to the amount of rainfall and notifies the driver. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2004-230910 Summary of the Invention [Problem to be solved by the invention]

[0004] In the technology described in Patent Document 1, when the amount of rainfall exceeds a predetermined amount, the vehicle speed is controlled to be reduced and the distance to an object ahead at which an alarm should be issued is set to be small. However, the vehicle cannot be controlled appropriately in bad weather, such as in dense fog or in bad weather due to snowfall (dry snowfall) (specifically, when the onboard sensor cannot detect an object ahead).

[0005] In view of the above points, the present disclosure aims to provide a vehicle control device, a vehicle control method, and a program that can reduce the risk of inappropriate control of the vehicle when the limit detection distance of the surrounding condition sensor installed in the vehicle is short (specifically, when the surrounding condition sensor cannot detect a vehicle ahead of the vehicle). [Means for solving the problem]

[0006] (1) One aspect of the present disclosure is a vehicle control device that includes an inference unit that infers a limit detection distance, which is the maximum distance from a surrounding condition sensor that can be detected by the surrounding condition sensor mounted on a vehicle, and the inference unit infers the limit detection distance of the surrounding condition sensor based on the sensor data of the surrounding condition sensor by using a machine learning model obtained by learning using teacher data, which is a data set of sensor data of the learning surrounding condition sensor mounted on a training vehicle and a label indicating the limit detection distance of the learning surrounding condition sensor at the time the sensor data of the learning surrounding condition sensor was obtained.

[0007] (2) In the vehicle control device of (1), the limit detection distance of the learning surrounding conditions sensor may be the inter-vehicle distance between the learning vehicle and the preceding vehicle detected by a radar mounted on the learning vehicle when the state switches between a state in which the learning vehicle can detect a preceding vehicle based on sensor data from the learning surrounding conditions sensor and a state in which the learning vehicle cannot detect the preceding vehicle based on sensor data from the learning surrounding conditions sensor.

[0008] (3) The vehicle control device of (1) or (2) includes an upper limit speed setting unit that sets an upper limit speed of the host vehicle while driving assistance for the host vehicle is being performed, and when the limit detection distance of the surrounding condition sensor inferred by the inference unit is equal to or less than a threshold, the upper limit speed setting unit may assume that the inter-vehicle distance between the host vehicle and a vehicle preceding the host vehicle is approximately equal to the limit detection distance inferred by the inference unit, and set the upper limit speed to a speed at which the host vehicle can follow the preceding vehicle.

[0009] (4) One aspect of the present disclosure is a vehicle control method including an inference step in which a vehicle control device infers a limit detection distance, which is the maximum distance from a surrounding condition sensor that can be detected by the surrounding condition sensor mounted on the vehicle, and in the inference step, the limit detection distance of the surrounding condition sensor is inferred based on the sensor data of the surrounding condition sensor by using a machine learning model obtained by learning using teacher data, which is a data set of sensor data of the learning surrounding condition sensor mounted on a training vehicle and a label indicating the limit detection distance of the learning surrounding condition sensor at the time the sensor data of the learning surrounding condition sensor was obtained.

[0010] (5) One aspect of the present disclosure is a program for causing a processor to execute an inference step of inferring a limit detection distance, which is the maximum distance from a surrounding condition sensor that can be detected by the surrounding condition sensor mounted on a vehicle, wherein in the inference step, the limit detection distance of the surrounding condition sensor is inferred based on the sensor data of the surrounding condition sensor by using a machine learning model obtained by learning using teacher data, which is a data set of sensor data of the learning surrounding condition sensor mounted on a training vehicle and a label indicating the limit detection distance of the learning surrounding condition sensor at the time the sensor data of the learning surrounding condition sensor was obtained. [Effects of the Invention]

[0011] According to the present disclosure, it is possible to reduce the risk of inappropriate control of the host vehicle being executed when the limit detection distance of the surrounding condition sensor mounted on the host vehicle is short. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a diagram illustrating an example of a host vehicle to which a vehicle control device according to a first embodiment is applied. [Figure 2] FIG. 1 is a diagram illustrating an example of a learning vehicle used to obtain a machine learning model used for inferring a limit detection distance by an inference unit. [Figure 3]10 is a flowchart for explaining an example of processing executed by the processor of the vehicle control device of the first embodiment when driving assistance (active cruise control) of the host vehicle 1 is executed in bad weather. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, embodiments of a vehicle control device, a vehicle control method, and a program according to the present disclosure will be described with reference to the drawings.

[0014] First Embodiment FIG. 1 is a diagram illustrating an example of a host vehicle 1 to which a vehicle control device 14 according to a first embodiment is applied. In the example illustrated in FIG. 1, the host vehicle 1 includes a surrounding situation sensor 11, a vehicle speed sensor 12, an HMI (Human Machine Interface) 13, a vehicle control device 14, a steering actuator 15, a braking actuator 16, and a drive actuator 17. The surrounding situation sensor 11 detects the surrounding situation (surrounding environment) of the host vehicle 1. The surrounding situation sensor 11 includes a camera that captures an image of the area ahead of the host vehicle 1. The surrounding situation sensor 11 has a function of detecting, based on an image of the area ahead of the host vehicle 1 captured by the camera, the distance from the surrounding situation sensor 11 to a preceding vehicle or the like included in the image (approximate distance between the host vehicle 1 and the preceding vehicle or the like). The surrounding situation sensor 11 transmits sensor data (e.g., an image of the area ahead of the host vehicle 1 captured by the camera, a signal indicating the distance between the host vehicle 1 and the preceding vehicle or the like) to the vehicle control device 14. The vehicle speed sensor 12 detects the speed of the host vehicle 1 and transmits a signal indicating the speed of the host vehicle 1 to the vehicle control device 14. The HMI 13 has functions such as accepting various operations by the driver of the host vehicle 1, and transmits a signal indicating the operations by the driver of the host vehicle 1 to the vehicle control device 14.

[0015] The vehicle control device 14 is configured by a microcomputer including a communication interface (I / F) 141, a memory 142, and a processor 143. The communication interface 141 has an interface circuit for connecting the vehicle control device 14 to the surrounding condition sensor 11, the vehicle speed sensor 12, and the HMI 13. The memory 142 stores programs and various data used in the processing executed by the processor 143. The processor 143 has a function as an acquisition unit 3A, a function as an inference unit 3B, a function as a control unit 3C, and a function as an upper limit speed setting unit 3D. The acquisition unit 3A acquires sensor data transmitted from the surrounding condition sensor 11. The acquisition unit 3A also acquires a signal indicating the speed of the host vehicle 1 transmitted from the vehicle speed sensor 12, a signal indicating the operation of the driver of the host vehicle 1 transmitted from the HMI 13, etc.

[0016] However, due to bad weather (e.g., rainfall, snowfall, thick fog), the accuracy of the surrounding conditions (surrounding environment) detection by the surrounding conditions sensor 11 may decrease, and the surrounding conditions sensor 11 may be unable to detect a vehicle in front of the vehicle 1, and may be unable to detect the distance from the surrounding conditions sensor 11 to the preceding vehicle, etc. Therefore, in the example shown in Figure 1, the inference unit 3B infers the limit detection distance (< the distance from the surrounding condition sensor 11 to the preceding vehicle, etc.), which is the maximum distance from the surrounding condition sensor 11 that can be detected by the surrounding condition sensor 11 during bad weather, etc.

[0017] FIG. 2 is a diagram showing an example of a training vehicle L1 used to obtain a machine learning model used by the inference unit 3B to infer the limit detection distance. In the example shown in FIG. 2, the training vehicle L1 is equipped with a training surroundings sensor L11 and a radar L12. The training surroundings sensor L11 detects the surroundings of the training vehicle L1. The training surroundings sensor L11 includes a training camera that captures images ahead of the training vehicle L1. The radar L12 detects the distance between the training vehicle L1 and a vehicle ahead of the training vehicle L1, for example, in bad weather.

[0018] For example, in bad weather or the like, when the distance between the learning vehicle L1 and the preceding vehicle is short, the learning surrounding condition sensor L11 can detect the preceding vehicle based on the sensor data of the learning surrounding condition sensor L11 (specifically, an image of the front of the learning vehicle L1 captured by the learning camera). However, when the distance between the learning vehicle L1 and the preceding vehicle increases, the learning surrounding condition sensor L11 cannot detect the preceding vehicle based on the sensor data of the learning surrounding condition sensor L11 (the image of the front of the learning vehicle L1 captured by the learning camera). In other words, when the distance between the learning vehicle L1 and the preceding vehicle changes from a short state to a long state, the learning surrounding condition sensor L11 switches from a state in which it can detect the preceding vehicle based on the image of the front of the learning vehicle L1 (detectable state) to a state in which it cannot detect the preceding vehicle based on the image of the front of the learning vehicle L1 (undetectable state). Furthermore, when the inter-vehicle distance changes from a long state to a short state, the state switches from the undetectable state to the detectable state. In the example shown in Figure 2, when the detection state switches between the detectable state and the undetectable state, radar L12 detects the distance between the learning vehicle L1 and the preceding vehicle, and the distance between the learning vehicle L1 and the preceding vehicle at that time is used as the limit detection distance of the learning surrounding condition sensor L11.

[0019] 1 and 2, the inference unit 3B infers the limit detection distance of the surrounding condition sensor 11 (the maximum distance from the surrounding condition sensor 11 that can be detected by the surrounding condition sensor 11) using the limit detection distance of the learning surrounding condition sensor L11. In detail, the inference unit 3B infers the limit detection distance of the surrounding condition sensor 11 based on the sensor data of the surrounding condition sensor 11 (the image in front of the host vehicle 1 taken by the camera) by using a machine learning model obtained by learning using a data set (teacher data) of an image in front of the learning vehicle L1 taken by the learning camera when the above-mentioned detectable state switches to the undetectable state and the inter-vehicle distance between the learning vehicle L1 and the preceding vehicle detected by the radar L12 at that time (the limit detection distance of the learning surrounding condition sensor L11). In other words, the inference unit 3B infers the limit detection distance of the surrounding situation sensor 11 based on the sensor data of the surrounding situation sensor 11 by using a machine learning model obtained by learning using a dataset (teacher data) of sensor data of the learning surrounding situation sensor L11 (an image of the front of the learning vehicle L1 taken by the learning camera) and a label indicating the limit detection distance of the learning surrounding situation sensor L11 at the time the sensor data of the learning surrounding situation sensor L11 was obtained.

[0020] The control unit 3C controls the steering actuator 15, the braking actuator 16, and the drive actuator 17 based on signals transmitted from the HMI 13, etc. Specifically, the control unit 3C has a function of executing driving assistance for the host vehicle 1. Driving assistance for the host vehicle 1 includes, for example, adaptive cruise control (ACC). While adaptive cruise control is being executed, the control unit 3C controls the braking actuator 16 and the drive actuator 17 based on the set speed of the host vehicle 1 and the distance between the host vehicle 1 and a preceding vehicle, which are received by the HMI 13, and the sensor data of the surrounding situation sensor 11 (an image of the area ahead of the host vehicle 1 captured by a camera). Specifically, while adaptive cruise control is being executed, the control unit 3C executes control to make the host vehicle 1 travel following the preceding vehicle while maintaining a constant distance between the host vehicle 1 and the preceding vehicle.

[0021] For example, in bad weather or the like, there may be a case where the limit detection distance of the surrounding situation sensor 11 is shorter than the inter-vehicle distance between the host vehicle 1 received by the HMI 13 and the preceding vehicle. When adaptive cruise control is executed in such a case, although the preceding vehicle exists, the preceding vehicle is not detected by the surrounding situation sensor 11, the inter-vehicle distance between the host vehicle 1 and the preceding vehicle becomes inappropriate, or the driver of the host vehicle 1 desires to continue the adaptive cruise control, but there is a possibility that the adaptive cruise control may be released by, for example, the control unit 3C or the like. Therefore, in the example shown in FIGS. 1 and 2, the upper speed setting unit 3D sets the upper speed of the host vehicle 1 during the execution of the driving support (specifically, active cruise control) of the host vehicle 1. Specifically, when the limit detection distance of the surrounding situation sensor 11 inferred by the inference unit 3B is less than or equal to the threshold value (specifically, when the preceding vehicle cannot be detected based on the sensor data of the surrounding situation sensor 11), the upper speed setting unit 3D assumes that the inter-vehicle distance between the host vehicle 1 and the preceding vehicle is approximately equal to the limit detection distance inferred by the inference unit 3B, and sets the speed at which the host vehicle 1 can safely follow the preceding vehicle as the upper speed of the host vehicle 1 described above. As a result, when the learning surrounding situation sensor L11 cannot detect the preceding vehicle during the execution of the active cruise control, the control unit 3C causes the host vehicle 1 to travel at the upper speed of the host vehicle 1 set by the upper speed setting unit 3D (<the set speed of the host vehicle 1 received by the HMI 13) instead of the set speed of the host vehicle 1 received by the HMI 13. Therefore, even in bad weather or the like, the execution of the active cruise control can be continued safely (specifically, without the host vehicle 1 approaching the preceding vehicle excessively).

[0022] FIG. 3 is a flowchart for explaining an example of the processing executed by the processor 143 of the vehicle control device 14 according to the first embodiment when the driving support (active cruise control) of the host vehicle 1 is executed in bad weather. 3, in step S10, the acquisition unit 3A acquires sensor data (images of the area ahead of the vehicle 1 captured by a camera) from the surrounding condition sensor 11. The acquisition unit 3A also acquires a signal indicating the speed of the vehicle 1 detected by the vehicle speed sensor 12, a signal indicating an operation by the driver of the vehicle 1 received by the HMI 13 (the set speed of the vehicle 1 while active cruise control is being executed and the distance between the vehicle 1 and the preceding vehicle), and the like. In step S11, the inference unit 3B uses a machine learning model obtained by learning using a dataset (teacher data) of sensor data from the learning surrounding condition sensor L11 (an image in front of the learning vehicle L1 taken by a learning camera) and a label indicating the limit detection distance of the learning surrounding condition sensor L11 at the time the sensor data from the learning surrounding condition sensor L11 was obtained, to infer the limit detection distance of the surrounding condition sensor 11 based on the sensor data from the surrounding condition sensor 11 (an image in front of the vehicle 1 taken by a camera). In step S12, for example, the upper limit speed setting unit 3D determines whether the limit detection distance of the surrounding condition sensor 11 inferred in step S11 is shorter than the inter-vehicle distance between the host vehicle 1 and the preceding vehicle during active cruise control execution, which was acquired in step S10. If the answer is YES, the process proceeds to step S13, and if the answer is NO, the process proceeds to step S16.

[0023] In step S13, upper limit speed setting unit 3D assumes that the inter-vehicle distance between host vehicle 1 and the preceding vehicle is approximately equal to the limit detection distance of surrounding condition sensor 11 inferred in step S11, and sets a speed at which host vehicle 1 can safely follow the preceding vehicle as the above-mentioned upper limit speed of host vehicle 1. In detail, if the speed of host vehicle 1 acquired in step S10 (the set speed of host vehicle 1 while active cruise control is being executed) is higher than the upper limit speed of host vehicle 1, upper limit speed setting unit 3D changes the set speed of host vehicle 1 while active cruise control is being executed to the upper limit speed of host vehicle 1 (reduces it to the upper limit speed of host vehicle 1).

[0024] In step S14, for example, the control unit 3C determines whether the speed of the host vehicle 1 acquired in step S10 (the speed of the host vehicle 1 detected by the vehicle speed sensor 12) is higher than the upper limit speed set in step S13 for the host vehicle 1. If YES, the process proceeds to step S15, and if NO, the process proceeds to step S16. In step S15, the control unit 3C decelerates the host vehicle 1 until the speed of the host vehicle 1 detected by the vehicle speed sensor 12 becomes equal to the upper limit speed of the host vehicle 1 set in step S13. In step S16, the control unit 3C continues to execute the adaptive cruise control.

[0025] In a host vehicle 1 to which the vehicle control device 14 of the first embodiment is applied, the risk of inappropriate control of the host vehicle 1 being performed can be reduced when the limit detection distance of the surrounding condition sensor 11 is short (for example, when the surrounding condition sensor 11 cannot detect a vehicle ahead of the host vehicle 1 due to bad weather, etc.).

[0026] Second Embodiment The host vehicle 1 to which the vehicle control device 14 of the second embodiment is applied is configured similarly to the host vehicle 1 to which the vehicle control device 14 of the first embodiment shown in Fig. 1 is applied. As described above, the training vehicle L1 used to obtain the machine learning model used to infer the limit detection distance by the inference unit 3B of the vehicle control device 14 of the first embodiment is equipped with a radar L12. On the other hand, the training vehicle L1 used to obtain the machine learning model used to infer the limit detection distance by the inference unit 3B of the vehicle control device 14 of the second embodiment is not equipped with a radar.

[0027] In one example of the second embodiment, the limit detection distance of the learning surrounding condition sensor L11 is the inter-vehicle distance between the learning vehicle L1 and the preceding vehicle calculated from the sensor data of the learning surrounding condition sensor L11 before (preferably just before) the state switches from a state in which the vehicle preceding the learning vehicle L1 can be detected (detectable state) based on the sensor data of the learning surrounding condition sensor L11 (an image of the front of the learning vehicle L1 taken by the learning camera) to a state in which the vehicle preceding the learning vehicle L1 cannot be detected (undetectable state) based on the sensor data of the learning surrounding condition sensor L11. In another example of the second embodiment, the limit detection distance of the learning surrounding condition sensor L11 is the inter-vehicle distance between the learning vehicle L1 and the preceding vehicle calculated from the sensor data of the learning surrounding condition sensor L11 after (preferably immediately after) the state switches from one in which the learning vehicle L1 cannot detect the preceding vehicle (undetectable state) based on the sensor data of the learning surrounding condition sensor L11 to one in which the learning vehicle L1 can detect the preceding vehicle (detectable state) based on the sensor data of the learning surrounding condition sensor L11. In other words, in the second embodiment, the inter-vehicle distance between the learning vehicle L1 and the preceding vehicle calculated based on an image of the front of the learning vehicle L1 taken by the learning camera before or after the detection state switches between the detectable state and the undetectable state (for example, an image including a base point sign of the inter-vehicle distance confirmation section) is used as the limit detection distance of the learning surrounding condition sensor L11. In the second embodiment, the inference unit 3B uses a machine learning model obtained by learning using a data set (teacher data) of an image of the front of the training vehicle L1 taken by the training camera when the detection state switches between a detectable state and an undetectable state, and the inter-vehicle distance between the training vehicle L1 and the preceding vehicle (the limit detection distance of the training surrounding condition sensor L11) calculated based on an image of the front of the training vehicle L1 taken by the training camera before or after the detection state switches between a detectable state and an undetectable state, to infer the limit detection distance of the surrounding condition sensor 11 based on the sensor data of the surrounding condition sensor 11 (an image of the front of the host vehicle 1 taken by the camera).

[0028] Even in a host vehicle 1 to which the vehicle control device 14 of the second embodiment is applied, the risk of inappropriate control of the host vehicle 1 being performed can be reduced when the limit detection distance of the surrounding condition sensor 11 is short (for example, when the surrounding condition sensor 11 cannot detect a vehicle ahead of the host vehicle 1 due to bad weather, etc.).

[0029] Third Embodiment The host vehicle 1 to which the vehicle control device 14 of the third embodiment is applied is configured similarly to the host vehicle 1 to which the vehicle control device 14 of the first embodiment shown in Fig. 1 is applied, except for the points described below. The learning vehicle L1 used to obtain a machine learning model used to infer the limit detection distance by the inference unit 3B of the vehicle control device 14 of the third embodiment is configured similarly to the learning vehicle L1 shown in Fig. 2, except for the points described below.

[0030] As described above, in the host vehicle 1 to which the vehicle control device 14 of the first embodiment is applied, the surrounding condition sensor 11 includes a camera that captures images ahead of the host vehicle 1. Also, in the first embodiment, the learning surrounding condition sensor L11 includes a learning camera that captures images ahead of the learning vehicle L1. On the other hand, in the host vehicle 1 to which the vehicle control device 14 of the third embodiment is applied, the surrounding condition sensor 11 includes a LiDAR (Light Detection And Ranging) that detects the distance between the host vehicle 1 and a vehicle preceding the host vehicle 1. Also, in the third embodiment, the learning surrounding condition sensor L11 includes a learning LiDAR that detects the distance between the learning vehicle L1 and a vehicle preceding the learning vehicle L1.

[0031] As described above, in the first embodiment, an image of the front of the host vehicle 1 taken by a camera is used as the sensor data of the surrounding condition sensor 11, and an image of the front of the learning vehicle L1 taken by a learning camera is used as the sensor data of the learning surrounding condition sensor L11. On the other hand, in the third embodiment, sensor data of a LiDAR (e.g., reflected light intensity, detected point cloud, etc.) is used as the sensor data of the surrounding condition sensor 11, and sensor data of a learning LiDAR (e.g., reflected light intensity, detected point cloud, etc.) is used as the sensor data of the learning surrounding condition sensor L11. Also, in the third embodiment, a state in which the learning vehicle L1 cannot detect a preceding vehicle based on the sensor data of the learning surrounding condition sensor L11 includes a state in which the inter-vehicle distance between the learning vehicle L1 and the preceding vehicle calculated based on the sensor data of the learning LiDAR differs by a predetermined threshold or more from the inter-vehicle distance between the learning vehicle L1 and the preceding vehicle calculated based on the sensor data of the radar L12.

[0032] <Fourth embodiment> The host vehicle 1 to which the vehicle control device 14 of the fourth embodiment is applied is configured similarly to the host vehicle 1 to which the vehicle control device 14 of the third embodiment described above is applied. The training vehicle L1 used to obtain a machine learning model used to infer the limit detection distance by the inference unit 3B of the vehicle control device 14 of the third embodiment is equipped with a radar L12. On the other hand, the training vehicle L1 used to obtain a machine learning model used to infer the limit detection distance by the inference unit 3B of the vehicle control device 14 of the fourth embodiment is not equipped with a radar.

[0033] In one example of the fourth embodiment, the limit detection distance of the learning surrounding condition sensor L11 is the inter-vehicle distance between the learning vehicle L1 and the preceding vehicle calculated from the sensor data of the learning surrounding condition sensor L11 before (preferably just before) the state switches from a state in which the learning vehicle L1 can detect a preceding vehicle (detectable state) based on the sensor data of the learning surrounding condition sensor L11 (learning LiDAR) to a state in which the learning vehicle L1 cannot detect a preceding vehicle (undetectable state) based on the sensor data of the learning surrounding condition sensor L11. In another example of the fourth embodiment, the limit detection distance of the learning surrounding condition sensor L11 is the inter-vehicle distance between the learning vehicle L1 and the preceding vehicle calculated from the sensor data of the learning surrounding condition sensor L11 after (preferably immediately after) the state switches from one in which the learning vehicle L1 cannot detect a preceding vehicle (undetectable state) based on the sensor data of the learning surrounding condition sensor L11 (learning LiDAR) to one in which the learning vehicle L1 can detect a preceding vehicle (detectable state) based on the sensor data of the learning surrounding condition sensor L11. In other words, in the fourth embodiment, the inter-vehicle distance between the learning vehicle L1 and the preceding vehicle calculated based on the sensor data of the learning surrounding condition sensor L11 (learning LiDAR) before or after the detection state switches between the detectable state and the undetectable state is used as the limit detection distance of the learning surrounding condition sensor L11. In the fourth embodiment, the inference unit 3B uses a machine learning model obtained by learning using a data set (teacher data) of sensor data from the learning surrounding condition sensor L11 (learning LiDAR) when the detectable state switches between a detectable state and an undetectable state, and the inter-vehicle distance between the learning vehicle L1 and the preceding vehicle (the limit detection distance of the learning surrounding condition sensor L11) calculated based on the sensor data of the learning surrounding condition sensor L11 (learning LiDAR) before or after the detectable state switches between a detectable state and an undetectable state, to infer the limit detection distance of the surrounding condition sensor 11 based on the sensor data of the surrounding condition sensor 11 (LiDAR).

[0034] As described above, embodiments of the vehicle control device, vehicle control method, and program of the present disclosure have been described with reference to the drawings. However, the vehicle control device, vehicle control method, and program of the present disclosure are not limited to the above-described embodiments and may be modified as appropriate without departing from the spirit and scope of the present disclosure. The configurations of the above-described embodiments may be combined as appropriate. In the above-described embodiments, the processing performed in the vehicle control device 14 has been described as software processing performed by executing a program. However, the processing performed in the vehicle control device 14 may be processing performed by hardware. Alternatively, the processing performed in the vehicle control device 14 may be processing that combines both software and hardware. Furthermore, the program stored in the memory 142 of the vehicle control device 14 (the program that realizes the functions of the processor 143 of the vehicle control device 14) may be recorded on a computer-readable storage medium such as a semiconductor memory, a magnetic recording medium, an optical recording medium, or the like, and provided, distributed, etc. [Explanation of symbols]

[0035] 1... host vehicle, 11... surrounding condition sensor, 12... vehicle speed sensor, 13... HMI, 14... vehicle control device, 141... communication interface, 142... memory, 143... processor, 3A... acquisition unit, 3B... inference unit, 3C... control unit, 3D... upper limit speed setting unit, 15... steering actuator, 16... braking actuator, 17... drive actuator, L1... learning vehicle, L11... learning surrounding condition sensor, L12... radar

Claims

1. an inference unit that infers a limit detection distance that is a maximum value of a distance from a surrounding condition sensor that can be detected by the surrounding condition sensor mounted on the host vehicle; The vehicle control device, wherein the inference unit infers the limit detection distance of the surrounding condition sensor based on the sensor data of the surrounding condition sensor by using a machine learning model obtained by learning using teacher data, which is a data set of sensor data of the learning surrounding condition sensor mounted on the learning vehicle and a label indicating the limit detection distance of the learning surrounding condition sensor at the time the sensor data of the learning surrounding condition sensor was obtained.

2. 2. The vehicle control device according to claim 1, wherein the limit detection distance of the learning surrounding condition sensor is the inter-vehicle distance between the learning vehicle and the preceding vehicle detected by a radar mounted on the learning vehicle when a state in which the learning vehicle can detect a preceding vehicle based on sensor data from the learning surrounding condition sensor switches to a state in which the learning vehicle cannot detect the preceding vehicle based on sensor data from the learning surrounding condition sensor.

3. an upper limit speed setting unit that sets an upper limit speed of the host vehicle while driving assistance is being performed on the host vehicle; 2. The vehicle control device according to claim 1, wherein when the limit detection distance of the surrounding condition sensor inferred by the inference unit is equal to or less than a threshold, the upper limit speed setting unit assumes that the inter-vehicle distance between the subject vehicle and a vehicle preceding the subject vehicle is approximately equal to the limit detection distance inferred by the inference unit, and sets the upper limit speed to a speed at which the subject vehicle can follow the preceding vehicle.

4. the vehicle control device includes an inference step of inferring a limit detection distance that is a maximum value of a distance from a surrounding condition sensor that can be detected by the surrounding condition sensor mounted on the host vehicle, In the inference step, the limit detection distance of the surrounding environment sensor is inferred based on the sensor data of the surrounding environment sensor by using a machine learning model obtained by learning using teacher data, which is a data set of sensor data of the learning surrounding environment sensor mounted on the learning vehicle and a label indicating the limit detection distance of the learning surrounding environment sensor at the time the sensor data of the learning surrounding environment sensor was obtained.

5. A program for causing a processor to execute an inference step of inferring a limit detection distance, which is a maximum value of a distance from a surrounding condition sensor that can be detected by the surrounding condition sensor mounted on the host vehicle, In the inference step, the limit detection distance of the surrounding condition sensor is inferred based on the sensor data of the surrounding condition sensor by using a machine learning model obtained by learning using teacher data, which is a data set of sensor data of a learning surrounding condition sensor mounted on a learning vehicle and a label indicating the limit detection distance of the learning surrounding condition sensor at the time the sensor data of the learning surrounding condition sensor was obtained.

Images