Vehicle control device, program, and vehicle control method

Abstract

A vehicle control unit (110) is provided with: a sensor signal processing unit (111) that acquires sensor signals from a plurality of sensors that detect physical quantities relating to the surroundings of a vehicle, and acquires image data indicating images of the surroundings of the vehicle from one or a plurality of cameras that capture the images of the surroundings of the vehicle; a human vision calculation unit (112) that uses a human vision model calculated in advance to calculate a visual repulsion potential field that is affected by the vision of a human recognizing the image; and a potential risk prediction model processing unit (114) that, using a potential risk prediction model for predicting a potential risk from a feature amount of the surroundings of the target vehicle, calculates a physical repulsive potential field based on the surroundings of the vehicle on the basis of the physical quantity and the image, and performs prediction of the potential risk based on the physical repulsive potential field. And calculating a comprehensive repulsive force potential field obtained by correcting the physical repulsive force potential field by using the visual repulsive force potential field.

Description

Technical Field

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

[0002] In recent years, technologies related to autonomous driving of vehicles have been developed. The basic technologies for driving a vehicle are cognition, judgment, and operation. Among them, it is necessary to judge the existence of an unspecified number of traffic participants and deal with their complex combinations. Research on driving control, which determines and controls the path of the vehicle based on such judgments, is prevalent.

[0003] In vehicle path determination and control, the artificial potential field method (also called potential field method) described in Patent Document 1 is often used. The artificial potential field method basically performs control using a physical vehicle control model based on the idea that human sense of danger is reflected in vehicle behavior.

[0004] Prior art literature

[0005] Patent Literature

[0006] Patent Document 1: Japanese Patent Application Publication No. 2018-192954 Summary of the invention

[0007] Problems to be solved by the invention

[0008] However, conventional vehicle control based on the artificial potential field method is not necessarily consistent with vehicle control that passengers perceive, judge, and operate through perception, and there is a problem that passengers cannot ride safely.

[0009] Therefore, an object of one or more aspects of the present disclosure is to perform automatic driving of a vehicle in a manner that reduces the discomfort felt by passengers.

[0010] Means for solving problems

[0011] A vehicle control device according to one embodiment of the present invention is characterized in that it comprises: a sensor signal acquisition unit that acquires sensor signals from a plurality of sensors, wherein the plurality of sensors detect physical quantities related to the surrounding environment of the vehicle; an image data acquisition unit that acquires image data representing the image from one or more cameras that capture the surrounding image of the vehicle; and a physical potential field calculation unit that calculates a repulsive potential field based on the surrounding environment of the vehicle based on the physical quantity and the image, using a potential risk prediction model that has been pre-learned in order to predict potential risks based on characteristic quantities of the surrounding environment of the target vehicle. A physical repulsive potential field; a human vision calculation unit, which uses a pre-calculated human vision model to calculate the repulsive potential field affected by the vision of humans recognizing the image, namely, a visible repulsive potential field; a potential field correction unit, which calculates a comprehensive repulsive potential field obtained by correcting the physical repulsive potential field using the visible repulsive potential field; a target path generation unit, which generates a target path in such a way that the vehicle travels from the current position of the vehicle to the target location according to the gradient calculated based on the comprehensive repulsive potential field; and a vehicle control unit, which controls the vehicle in such a way that the vehicle travels along the target path.

[0012] A program of one embodiment of the present disclosure is characterized in that a computer is made to function as the following parts: a sensor signal acquisition unit that acquires sensor signals from a plurality of sensors, wherein the plurality of sensors detect physical quantities related to the surrounding environment of a vehicle; an image data acquisition unit that acquires image data representing the image from one or more cameras that capture the surrounding image of the vehicle; and a physical potential field calculation unit that uses a potential risk prediction model that has been pre-learned in order to predict potential risks based on characteristic quantities of the surrounding environment of the target vehicle, and calculates a repulsive potential field based on the surrounding environment of the vehicle based on the physical quantity and the image. The force potential field is a physical repulsive potential field; a human vision calculation unit, which uses a pre-calculated human vision model to calculate the repulsive potential field affected by the vision of humans recognizing the image, namely the visible repulsive potential field; a potential field correction unit, which calculates the comprehensive repulsive potential field obtained by correcting the physical repulsive potential field using the visible repulsive potential field; a target path generation unit, which generates a target path in such a way that the vehicle travels from the current position of the vehicle to the target location according to the gradient calculated according to the comprehensive repulsive potential field; and a vehicle control unit, which controls the vehicle in such a way that the vehicle travels along the target path.

[0013] A vehicle control method according to one embodiment of the present invention is characterized in that sensor signals are obtained from a plurality of sensors, wherein the plurality of sensors detect physical quantities related to the surrounding environment of the vehicle, image data representing the image is obtained from one or more cameras that capture the surrounding image of the vehicle, a potential risk prediction model that has been pre-learned in order to predict potential risks based on characteristic quantities of the surrounding environment of the target vehicle is used, a repulsive potential field based on the surrounding environment of the vehicle, i.e., a physical repulsive potential field, is calculated based on the physical quantity and the image, a repulsive potential field, i.e., a visible repulsive potential field, which is affected by the vision of a human recognizing the image is calculated using a pre-calculated human visual model, and a comprehensive repulsive potential field is calculated by correcting the physical repulsive potential field using the visible repulsive potential field, so as to generate a target path for the vehicle to travel from the current position of the vehicle to the target location according to a gradient calculated according to the comprehensive repulsive potential field, so as to control the vehicle so that the vehicle travels along the target path.

[0014] Effects of the Invention

[0015] According to one or more aspects of the present disclosure, it is possible to perform automatic driving of a vehicle in a manner that reduces the discomfort felt by passengers. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a block diagram showing the main part structure of a vehicle control system installed in a vehicle that performs automatic driving.

[0017] Figure 2 This is a block diagram schematically showing the structure of the human vision calculation unit and the potential risk prediction model processing unit.

[0018] Figure 3 This is a schematic diagram showing an example of a head-on environment of a vehicle equipped with a vehicle control system.

[0019] Figure 4 (A) and (B) are block diagrams showing hardware configuration examples. DETAILED DESCRIPTION

[0020] Implementation Method 1

[0021] Figure 1 This is a block diagram showing a main part configuration of a vehicle control system 100 mounted on a vehicle that performs automatic driving.

[0022] The vehicle control system 100 includes a sensor group 101 , a vehicle control unit 110 , an actuator 102 , and a vehicle drive unit 103 .

[0023] The sensor group 101 includes a plurality of sensors that detect physical quantities related to the surrounding environment of the vehicle.

[0024] The sensor group 101 includes one or more cameras that function as an imaging unit that captures an image of the surroundings of the vehicle. Image data representing the captured image is provided to the vehicle control unit 110 .

[0025] Here, the sensor group 101 includes an environment sensor for detecting people and obstacles around the vehicle. For example, the sensor group 101 includes at least one of a high-frequency radar sensor, an ultrasonic sensor, and a LiDAR. The environment sensor provides a sensor signal indicating the detected content to the vehicle control unit 110 .

[0026] Furthermore, the sensor group 101 includes a vehicle sensor for detecting the operation state and behavior of the vehicle equipped with the vehicle control system 100. For example, the sensor group 101 includes at least one of a vehicle speed sensor, an acceleration sensor, and an angular velocity sensor. In addition, the sensor group 101 includes at least one of an accelerator position sensor, a brake stroke sensor, a brake pressure sensor, a steering angle sensor, an engine speed sensor, a brake light switch, and a turn signal switch. The vehicle sensor provides a sensor signal indicating the detected content to the vehicle control unit 110.

[0027] The sensor group 101 includes a GPS receiving device that functions as a GPS receiving unit including a GPS antenna for receiving GPS (Global Positioning System) signals. The received GPS signals are provided to the vehicle control unit 110 .

[0028] The vehicle control unit 110 functions as a vehicle control device that controls the vehicle on which the vehicle control system 100 is mounted.

[0029] The vehicle control unit 110 has a sensor signal processing unit 111, a human vision calculation unit 112, a risk prediction knowledge database (hereinafter referred to as risk prediction knowledge DB) 113, a potential risk prediction model processing unit 114, a target path generation unit 115 and a vehicle control unit 116.

[0030] The sensor signal processing unit 111 acquires various signals and data from the sensor group 101 , performs signal processing such as analog / digital conversion when necessary, and provides the processed signals and data to various units at subsequent stages.

[0031] For example, the sensor signal processing unit 111 functions as an image data acquisition unit that acquires image data from one or more cameras, and provides the image data to the human vision calculation unit 112 and the potential risk prediction model processing unit 114 .

[0032] Furthermore, the sensor signal processing unit 111 functions as a sensor signal acquisition unit that acquires sensor signals from vehicle sensors, surrounding environment sensors, and a GPS receiving device.

[0033] The sensor signal processing unit 111 provides the sensor signal from the vehicle sensor to the potential risk prediction model processing unit 114. Furthermore, the sensor signal processing unit 111 provides the sensor signal from the surrounding environment sensor and the sensor signal from the GPS receiving device, that is, the GPS signal, to the target path generation unit 115.

[0034] The human vision calculation unit 112 uses a human vision model obtained by mathematical modeling in advance to calculate a visibility repulsive potential field that makes the recognition of the image represented by the image data as a repulsive potential energy affected by human vision. The calculated visibility repulsive potential field is provided to the potential risk prediction model processing unit 114.

[0035] For example, physiological or psychophysical human vision models include: lateral inhibition models in which optic nerve cells inhibit surrounding cells when stimulated by brightness, motion vision models such as optical flow or pseudo-motion, and color vision models, some of which are mathematically modeled. The human vision calculation unit 112 calculates the psychological potential field as the visibility repulsive potential field based on these mathematical models.

[0036] Figure 2 It is a block diagram schematically showing the structure of the human vision calculation unit 112 and the potential risk prediction model processing unit 114.

[0037] The human vision calculation unit 112 includes a lateral inhibition model processing unit 112 a and a motion vision model processing unit 112 b .

[0038] The lateral inhibition model processing unit 112 a calculates a psychological potential field as a visibility repulsion potential field using the lateral inhibition model as a visual model.

[0039] For example, the lateral inhibition model processing unit 112a uses a lateral inhibition model obtained by mathematically modeling the lateral inhibition of humans to calculate, as at least a part of the visibility repulsive potential field, a psychological repulsive potential field that affects the detection of obstacles due to the lateral inhibition of humans.

[0040] The lateral inhibition model applies inhibition from the peripheral part to the central part when homogeneous stimulation intensity is applied to the central part and the peripheral part of the receptive field, and when there is a difference in the stimulation intensity between them, the difference is enhanced by lateral inhibition, thereby affecting the contour and contrast enhancement, etc. Therefore, it can be considered that lateral inhibition affects the driver's ability to detect obstacles.

[0041] Here, first, from the viewpoint of physiological engineering, the lateral inhibition model shown in the following document 1 is used as a reference.

[0042] Reference 1: Katsuhiko Fujii, Akira Matsuoka, Ryuya Morita, "Analysis of visual illusion caused by lateral inhibition", Medical Electronics and Bioengineering, Vol. 5, No. 2, pp. 25-34, 1967

[0043] As object extraction, this side suppression model can be expressed by the following formula (1).

[0044] [Formula 1]

[0045]

[0046] Here, i(ξ, η) is the stimulus pattern, p(x, y) is the intensity of neural activity, and w(ξ-x, η-y) is the binding function.

[0047] In addition, x, y, ξ, and η are coordinate values ​​on the retina, x and y are reference coordinates of p, and ξ and η are coordinates of the obstacle.

[0048] The receptive fields of retinal nerve cells are concentrically structured, and when the coupling function is regarded as a spatial characteristic, it can be approximated by a DOG (Difference of Gaussian) function as shown in the following formula (2).

[0049] [Formula 2]

[0050]

[0051] Here, K1 is a coefficient representing the strength of excitatory binding, K2 is a coefficient representing the strength of inhibitory binding, σ1 is the variance of excitatory binding, and σ2 is the variance of inhibitory binding.

[0052] As mentioned above, when the output p(x, y) is assumed to be the psychological potential field Up for the shape of the object, the psychological potential field U p It is calculated by the following formula (3).

[0053] [Formula 3]

[0054]

[0055] In addition, i(ξ,η) in equation (3) is a stimulus pattern, and when it is treated as a higher-order color vision mechanism as color vision and treated as a binary image, the influence is regarded as an obstacle shape. Therefore, it is regarded as a definite integral and the following equation (4) is obtained.

[0056] [Formula 4]

[0057]

[0058] As described above, the lateral inhibition model processing unit 112a calculates the psychological potential field Up using equation (4).

[0059] The motion visual model processing unit 112 b calculates the motion visual repulsive potential field as the visibility repulsive potential field using the motion visual model as the visual model.

[0060] For example, the motion vision model processing unit 112b uses a motion vision model obtained by mathematically modeling human motion vision to calculate the motion vision repulsive potential field that affects the detection of obstacles due to human motion vision as at least a part of the visibility repulsive potential field.

[0061] In addition, motion vision is calculated by providing optical flow such as Lucas-Kanada method such as optical flow, apparent motion and guided motion as a detection model for moving objects, or providing it as a motion vision model such as Reichardt model and gradient detection model as a computational approach.

[0062] In this case, when the pixel of the target object is assumed to be pixel I (x, y, t), the pixel after Δt seconds is calculated by the following formula (5).

[0063] [Formula 5]

[0064] I(x,y,t)=I(x+dx,y+dy,t+Δt) (5)

[0065] Formula (5) is Taylor expanded and divided by dt to obtain the pixel movement speed. The constraint equation is calculated using the following formulas (6) and (7).

[0066] [Formula 6]

[0067]

[0068] [Formula 7]

[0069] f x ′u′+f y ′v′=0 (7)

[0070] As mentioned above, the motion visual repulsive potential field U m It is calculated by the following formula (8).

[0071] [Formula 8]

[0072] U m =m(f x ′u′+f y ′v′)l (8)

[0073] Here, m is the mass of the moving object, and l is the moving distance.

[0074] The apparent motion and the guided motion can also be calculated based on the same idea as above.

[0075] return Figure 1 , the risk prediction knowledge DB 113 stores potential risk prediction models.

[0076] As described in the above-mentioned Patent Document 1, the potential risk prediction model is a learned model that introduces the repulsive potential field composed of road boundaries, stationary obstacles, and protrusions from blind spots caused by the obstacles, and the gravitational potential field of the vehicle trajectory (path) to the target into the control model and learns the idea that human sense of danger is expressed in operation. The potential risk prediction model can be generated using a known technology.

[0077] The potential risk prediction model processing unit 114 uses the visibility repulsive potential field calculated by the human vision calculation unit 112 to correct the physical repulsive potential field, thereby calculating the comprehensive repulsive potential field. The physical repulsive potential field is the repulsive potential field calculated using the potential risk prediction model stored in the risk prediction knowledge DB 113.

[0078] like Figure 2 As shown, the potential risk prediction model processing unit 114 includes a physical potential field calculation unit 114a and a potential field correction unit 114b.

[0079] The physical potential field calculation unit 114a uses the potential risk prediction model stored in the risk prediction knowledge DB 113 to calculate the repulsive potential field, i.e., the physical repulsive potential field, obtained based on the physical position relationship between a vehicle equipped with the vehicle control system 100 and objects such as people or obstacles, according to the sensor signal from the sensor signal processing unit 111.

[0080] Here, the physical potential field calculation unit 114a uses a potential risk prediction model that has been pre-learned to predict potential risks based on characteristic quantities of the surrounding environment of the target vehicle, and calculates the repulsive potential field, i.e., the physical repulsive potential field, based on the surrounding environment of the vehicle equipped with the vehicle control system 100 based on the physical quantities represented by the sensor signals and the images represented by the image data.

[0081] For example, here, Figure 3 As shown, it is assumed that a vehicle 150 equipped with the vehicle control system 100 is traveling in the direction of arrow D on a straight road without lane boundaries and with walls on both sides. Furthermore, it is assumed that there is a parked vehicle 151 on the left shoulder of the road relative to the vehicle 150, and there are no other traffic participants when the vehicle 150 passes the right side thereof.

[0082] Under such circumstances, it is assumed that the associated repulsive potential field targets both the road boundary and the obstacle repulsion, the control model of the vehicle 150 targets the lateral control, and the vehicle 150 performs uniform linear motion.

[0083] First, according to the above definition, the road boundary repulsive potential field function U w (x, y) is represented by the following formula (9).

[0084] [Formula 9]

[0085]

[0086] Here, y wc is the y coordinate of the center of the road, w w is the weight coefficient, σ w is the variance.

[0087] In addition, the obstacle repulsive force potential field function U o (x, y) is represented by the following formula (10).

[0088] [Formula 10]

[0089]

[0090] Here, x or is the x coordinate of the rear of the parked vehicle 151, x of is the x coordinate of the front of the parked vehicle 151, y o is the y coordinate of the center of the parking vehicle 151 in the width direction, w o is the weight coefficient, σ ox , σ oy is the variance.

[0091] The above equation (9) is a one-dimensional Gaussian function when the length of the wall is set to infinity, and equation (10) becomes a two-dimensional Gaussian function.

[0092] The above repulsive potential field function U of the road boundary w (x, y) calculated road boundary repulsive potential field U w and the repulsive potential field function U through the obstacle o (x, y) calculated obstacle repulsion potential field U o is provided to the potential field correction unit 114b.

[0093] The potential field correction unit 114 b uses the visibility repulsive potential field calculated by the human vision calculation unit 112 to correct the physical repulsive potential field calculated by the physical potential field calculation unit 114 a , thereby calculating a comprehensive repulsive potential field.

[0094] For example, the potential field correction unit 114b calculates the comprehensive repulsive potential field U by the following formula (11): a .

[0095] [Formula 11]

[0096] U a =αU p +U m +U w +W o (11)

[0097] Here, since the above formula (4) is a coordinate system on the retina, α is a magnification factor for matching the real world with other items. The magnification factor may be calculated in advance by experiments or the like.

[0098] Here, the potential field correction unit 114 b calculates the comprehensive repulsive potential field by adding the visible repulsive potential field to the physical repulsive potential field. However, the comprehensive repulsive potential field may be calculated by multiplying the visible repulsive potential field and the physical repulsive potential field.

[0099] return Figure 1 The target path generation unit 115 generates a driving path, i.e., a target path, of a vehicle equipped with the vehicle control system 100 based on the GPS signal from the sensor signal processing unit 111 and the comprehensive repulsive potential field from the potential risk prediction model processing unit 114 .

[0100] For example, the target path generation unit 115 generates a target path in such a manner that the vehicle travels from the current position indicated by the GPS signal to the target location set by the driver of the vehicle according to the gradient calculated based on the comprehensive repulsive potential field. The generated target path is provided to the vehicle control unit 116. The generation of the target path using the potential field itself may be performed using a known method such as the method described in the above-mentioned patent document 1.

[0101] Specifically, the target path generation unit 115 uses the potential field method to calculate the gradient of the potential field as the force acting on the vehicle. The target path generation unit 115 determines a target yaw rate and a target speed with a low overall repulsive potential field on the trajectory of the vehicle several seconds in the future.

[0102] The vehicle control unit 116 controls the vehicle on which the vehicle control system 100 is mounted so that the vehicle travels along the target route generated by the target route generation unit 115 .

[0103] For example, the vehicle control unit 116 generates a control signal for controlling the vehicle so that the vehicle travels on the target route generated by the target route generation unit 115 , and supplies the control signal to the actuator 102 .

[0104] Specifically, the vehicle control unit 116 converts the target yaw rate determined by the target path generation unit 115 into a target steering angle. In addition, the vehicle control unit 116 converts the target speed determined by the target path generation unit 115 into a torque. Furthermore, the vehicle control unit 116 generates a control signal indicating the converted target steering angle and torque.

[0105] The actuator 102 operates the vehicle drive unit 103 , which is a mechanism for driving the vehicle, such as an engine, an accelerator, a brake, and a steering wheel, according to a control signal from the vehicle control unit 116 .

[0106] The vehicle drive unit 103 is a mechanism for driving the vehicle, such as an engine, an accelerator, a brake, and a steering wheel.

[0107] The sensor signal processing unit 111, the human vision calculation unit 112, the potential risk prediction model processing unit 114, the target path generation unit 115 and the vehicle control unit 116 described above may be partly or entirely Figure 4 As shown in (A), it can be composed of a memory 10 and a processor 11 such as a CPU (Central Processing Unit) that executes a program stored in the memory 10. Such a program can be provided via a network, or can be provided by recording in a recording medium. That is, such a program can also be provided as a program product, for example.

[0108] In addition, a part or all of the sensor signal processing unit 111, the human vision calculation unit 112, the potential risk prediction model processing unit 114, the target path generation unit 115, and the vehicle control unit 116 may be used. Figure 4 As shown in (B), the processing circuit 12 can also be composed of a single circuit, a composite circuit, a processor that operates according to a program, a parallel processor that operates according to a program, an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

[0109] As described above, the sensor signal processing unit 111 , the human vision calculation unit 112 , the potential risk prediction model processing unit 114 , the target path generation unit 115 , and the vehicle control unit 116 can be realized by a processing circuit network.

[0110] The risk prediction knowledge DB 113 can be realized by a storage device such as a HDD (Hard Disc Drive), an SDD (Solid State Drive), or a nonvolatile memory.

[0111] Description of symbols

[0112] 100: vehicle control system; 101: sensor group; 102: actuator; 103: vehicle drive unit; 110: vehicle control unit; 111: sensor signal processing unit; 112: human vision calculation unit; 112a: side inhibition model processing unit; 112b: motion vision model processing unit; 113: risk prediction knowledge DB; 114: potential risk prediction model processing unit; 114a: physical potential field calculation unit; 114b: potential field correction unit; 115: target path generation unit; 116: vehicle control unit.

Claims

1. A vehicle control device, characterized in that: The vehicle control device comprises: a sensor signal acquisition unit that acquires sensor signals from a plurality of sensors, wherein the plurality of sensors detect physical quantities related to the surrounding environment of the vehicle; an image data acquisition unit that acquires image data representing the image from one or more cameras that capture an image of the surroundings of the vehicle; A physical potential field calculation unit that uses a potential risk prediction model that has been previously learned in order to predict potential risks based on feature quantities of the surrounding environment of the target vehicle, and calculates a repulsive potential field based on the surrounding environment of the vehicle, that is, a physical repulsive potential field, based on the physical quantity and the image; a human vision calculation unit that uses a pre-calculated human vision model to calculate a repulsive potential field affected by vision of a human recognizing the image, that is, a visibility repulsive potential field; a potential field correction unit, which calculates a comprehensive repulsive potential field obtained by correcting the physical repulsive potential field using the visibility repulsive potential field; a target path generating unit for generating a target path in such a manner that the vehicle travels from the current position of the vehicle to a target location according to a gradient calculated based on the comprehensive repulsive potential field; and A vehicle control unit controls the vehicle so that the vehicle travels along the target route.

2. The vehicle control device according to claim 1, characterized in that: The human vision calculation unit calculates, as at least a part of the visibility repulsive potential field, a psychological repulsive potential field that affects the detection of obstacles due to the lateral inhibition of humans, using a lateral inhibition model obtained by mathematically modeling the lateral inhibition of humans.

3. The vehicle control device according to claim 1, characterized in that: The human vision calculation unit uses a motion vision model obtained by mathematically modeling human motion vision to calculate a motion vision repulsive potential field that affects the discovery of obstacles due to human motion vision as at least a part of the visibility repulsive potential field.

4. The vehicle control device according to any one of claims 1 to 3, characterized in that: The potential field correction section corrects the physical repulsive potential field by adding or multiplying the visibility repulsive potential field and the physical repulsive potential field.

5. A program, characterized in that The program causes the computer to function as: a sensor signal acquisition unit that acquires sensor signals from a plurality of sensors, wherein the plurality of sensors detect physical quantities related to the surrounding environment of the vehicle; an image data acquisition unit that acquires image data representing the image from one or more cameras that capture an image of the surroundings of the vehicle; A physical potential field calculation unit that uses a potential risk prediction model that has been previously learned in order to predict potential risks based on feature quantities of the surrounding environment of the target vehicle, and calculates a repulsive potential field based on the surrounding environment of the vehicle, that is, a physical repulsive potential field, based on the physical quantity and the image; a human vision calculation unit that uses a pre-calculated human vision model to calculate a repulsive potential field affected by vision of a human recognizing the image, that is, a visibility repulsive potential field; a potential field correction unit, which calculates a comprehensive repulsive potential field obtained by correcting the physical repulsive potential field using the visibility repulsive potential field; a target path generating unit for generating a target path in such a manner that the vehicle travels from the current position of the vehicle to a target location according to a gradient calculated based on the comprehensive repulsive potential field; and A vehicle control unit controls the vehicle so that the vehicle travels along the target route.

6. A vehicle control method, characterized in that: obtaining sensor signals from a plurality of sensors, wherein the plurality of sensors detect physical quantities related to the surrounding environment of the vehicle, acquiring image data representing the image from one or more cameras capturing an image of the surroundings of the vehicle, Using a potential risk prediction model that has been previously learned in order to predict potential risks based on feature quantities of the surrounding environment of the target vehicle, a repulsive potential field based on the surrounding environment of the vehicle, i.e., a physical repulsive potential field, is calculated based on the physical quantity and the image, Using a pre-calculated human vision model, calculating a repulsive potential field affected by the vision of a human recognizing the image, namely, a visibility repulsive potential field, calculating a comprehensive repulsive potential field obtained by correcting the physical repulsive potential field using the visual repulsive potential field, generating a target path in such a way that the vehicle travels from the current position of the vehicle to the target location according to the gradient calculated based on the comprehensive repulsive potential field, The vehicle is controlled in such a way that the vehicle travels along the target path.

Images