Driver Assistance Systems

By setting the target object status parameters and probability settings of multiple modes in the driving assistance system, the precise assessment of the collision risk of target object ahead of the vehicle is improved, and the problems of excessive or unnecessary risk avoidance control in the prior art are solved, so as to achieve More precise and necessary risk avoidance control.

CN113370975BActive Publication Date: 2024-12-03TOYOTA JIDOSHA KK
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110245154.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2020-03-10
Filing Date
2021-03-05
Publication Date
2024-12-03
Estimated Expiration
2041-03-05

AI Technical Summary

Technical Problem

When estimating the collision risk of target objects in front of a vehicle, it is assumed that the target objects move perpendicularly to the lane, resulting in excessive or unnecessary risk aversion control and inability to accurately reflect the actual situation, which may lead to unnecessary excessive or insufficient risk aversion work. .

Method used

By setting the target object state parameters of multiple modes in the driving assistance system, including the imaginary movement direction and speed, the risk value is calculated based on these parameters, and the risk field is comprehensively evaluated through probability settings to ensure the accuracy and physicality of the risk value. , and then carry out precise risk avoidance control.

Benefits of technology

It improves the accurate assessment of the collision risk of target objects in front of the vehicle, reduces the excess or unnecessary work of risk avoidance control, enhances the accuracy and necessity of control, and reduces the sense of incongruity of excessive or insufficient.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113370975B_ABST
    Figure CN113370975B_ABST
Patent Text Reader

Abstract

A driving assistance system is provided for suppressing excess or unnecessary work of risk avoidance control for reducing the risk of collision with a target object existing in front of a vehicle. The risk of collision with a target object existing in front of the vehicle is represented by a risk field representing the distribution of risk values. In the calculation of the risk field, target object state parameters of multiple modes are considered. In more detail, the probability of each target object state parameter is set based on target object information representing the moving direction and moving speed of the target object. In addition, the risk value when each target object state parameter is used is calculated as a partial risk value. The risk field is set using the risk value that is the sum of the product of the probability of the target object state parameter for multiple modes and the partial risk value. And, based on the risk field, it is determined whether to execute risk avoidance control or whether the working conditions of risk avoidance control are met.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to driving assistance control for assisting driving of a vehicle, and more particularly to risk avoidance control for reducing the risk of collision with a target object in front of the vehicle. Background Art

[0002] Patent document 1 discloses a driving assistance device. The driving assistance device detects an obstacle in front of a vehicle and estimates a potential risk existing in a blind spot of the obstacle. The potential risk is a hypothetical moving body existing in the blind spot of the obstacle. The hypothetical moving body is assumed to move perpendicularly to the lane at a predetermined speed and enter the lane. The driving assistance device calculates the risk level of the potential risk based on the estimated collision speed between the vehicle and the hypothetical moving body. In particular, the driving assistance device assumes a plurality of different vehicle positions and vehicle speeds and calculates a risk field representing the risk level in each of those plurality of vehicle positions and vehicle speeds. Furthermore, the driving assistance device sets an appropriate vehicle trajectory based on the risk field.

[0003] Prior art literature

[0004] Patent Document 1: Japanese Patent Application Publication No. 2017-206117 Summary of the invention

[0005] Technical problem to be solved by the invention

[0006] Consideration is given to "risk avoidance control" for reducing the risk of collision with a target object in front of the vehicle. In order to properly perform risk avoidance control, it is necessary to properly estimate the magnitude of the risk of collision with the target object. This is because: if the risk of collision is overestimated, the risk avoidance control will work excessively or unnecessarily.

[0007] According to the technology disclosed in the above-mentioned patent document 1, it is assumed that a hypothetical moving object existing in the blind spot of an obstacle will move perpendicularly to the lane at a predetermined speed and enter the lane. However, the actual target object is not necessarily limited to moving perpendicularly to the lane and entering the lane. When such an assumption is also applied to a target object that is not present in the blind spot but exists within a range that can be identified from the vehicle, the collision risk will be overestimated, and the possibility of risk avoidance control working excessively or unnecessarily will increase.

[0008] An object of the present invention is to provide a technology capable of suppressing redundant work or unnecessary work of risk avoidance control for reducing the risk of collision with a target object existing in front of a vehicle.

[0009] Technical solutions to solve problems

[0010] A first aspect relates to a driving assistance system for assisting driving of a vehicle.

[0011] The driving assistance system includes a storage device and a processor.

[0012] The storage device stores driving environment information indicating a driving environment of the vehicle.

[0013] The processor executes risk avoidance control including at least one of steering control and deceleration control for reducing a risk of collision with a target object existing in front of the vehicle based on the driving environment information.

[0014] The vehicle state parameters include the assumed relative position between the vehicle and the target object and the assumed speed of the vehicle.

[0015] The target object state parameters include the target object's assumed moving direction and assumed moving speed.

[0016] The risk value associated with the object is represented by a function of an estimated collision speed of the vehicle defined by the vehicle state parameters and the object defined by the object state parameters.

[0017] The risk field represents the distribution of risk values ​​of vehicle state parameters for multiple modes.

[0018] The driving environment information includes target object information, which indicates the position, moving direction, and moving speed of the target object.

[0019] The processor sets target state parameters of multiple modes.

[0020] The processor sets the probability of each of the target state parameters of the plurality of modes based on the target information.

[0021] The processor calculates a risk value using each of the target object state parameters of the plurality of patterns as a partial risk value.

[0022] The processor sets the risk field using a risk value that is a sum of products of probabilities of target object state parameters for a plurality of patterns and partial risk values.

[0023] The processor determines whether to execute risk avoidance control or whether a working condition of risk avoidance control is met based on the risk field.

[0024] The second viewpoint has the following features in addition to the first viewpoint.

[0025] The processor sets the target state parameters of the plurality of patterns so that the assumed movement directions of the plurality of patterns are distributed in a circumferential direction around the target.

[0026] The third viewpoint has the following features in addition to the second viewpoint.

[0027] The processor sets the target state parameters of the plurality of patterns so that the assumed movement directions of the plurality of patterns are evenly distributed in the circumferential direction.

[0028] The fourth viewpoint has the following features in addition to the second viewpoint or the third viewpoint.

[0029] The closer the assumed movement direction of the target object state parameter is to the movement direction of the target object represented by the target object information, the higher the probability of the target object state parameter is set by the processor.

[0030] The fifth viewpoint has the following features in addition to the second viewpoint or the third viewpoint.

[0031] The driving environment information also includes obstacle information indicating the position of an obstacle through which the target object cannot pass.

[0032] The processor sets the probability of the target object state parameter that an obstacle exists in the assumed moving direction to zero based on the obstacle information.

[0033] The sixth aspect has the following features in addition to any one of the first to fifth aspects.

[0034] The processor sets a plurality of modes of target state parameters in a manner assuming that the moving speed includes a plurality of modes.

[0035] The seventh viewpoint has the following features in addition to the sixth viewpoint.

[0036] The closer the assumed moving speed of the target object state parameter is to the moving speed of the target object represented by the target object information, the higher the probability of the target object state parameter is set by the processor.

[0037] The eighth aspect has the following features in addition to any one of the first to seventh aspects.

[0038] The target object information also indicates the type of the target object.

[0039] The processor calculates the probability based on the type of target object.

[0040] The ninth aspect has the following features in addition to any one of the first to eighth aspects.

[0041] When there are multiple targets, the processor calculates multiple risk fields for each of the multiple targets, and acquires the risk field by superimposing the multiple risk fields.

[0042] The tenth aspect has the following features in addition to any one of the first to ninth aspects.

[0043] The processor further executes emergency braking control to decelerate the vehicle at a first deceleration when a collision margin time with respect to the target object is less than a first threshold value.

[0044] The processor calculates an estimated collision speed in consideration of the operation of the emergency brake control, and calculates a risk value based on the estimated collision speed.

[0045] Effects of the Invention

[0046] According to the present invention, the risk of collision with a target object existing in front of a vehicle is represented by a risk field. In more detail, the risk field represents the distribution of risk values ​​for vehicle state parameters of multiple modes. The vehicle state parameters include the hypothetical relative position between the vehicle and the target object and the hypothetical speed of the vehicle. The target object state parameters include the assumed moving direction and the assumed moving speed of the target object. The risk value is represented by a function of the estimated collision speed of the vehicle defined by the vehicle state parameters and the target object defined by the target object state parameters. That is, the risk value is not sensory, but has a physical meaning. Therefore, the risk field (risk value) represents the risk of collision with the target object with high accuracy.

[0047] Further, according to the present invention, in the setting of the risk field, target object state parameters of multiple modes are considered. Specifically, the probability of each target object state parameter of multiple modes is set. The target object does not move randomly, but moves with a certain degree of continuity with the past movement. At least the most recent moving direction and moving speed of the target object serve as clues for estimating the future moving direction and moving speed. Therefore, the probability of each target object state parameter can be set based on the target object information representing the moving direction and moving speed of the target object. In addition, the risk value when each target object state parameter is used is calculated as a partial risk value. And, the sum of the product of the probability of the target object state parameter of all modes and the partial risk value is used as the final risk value.

[0048] In this way, in the setting of the risk field, the target state parameter is not fixed to one type, but multiple modes of target state parameters are considered. That is, multiple modes of target state parameters are reflected in the risk field. Therefore, the accuracy of the risk field representing the risk of collision with the target is further improved.

[0049] The high-precision risk field thus obtained is applied to risk avoidance control for reducing the risk of collision with the target object. The risk field represents the collision risk with high precision, and therefore, the excess work or unnecessary work of the risk avoidance control is suppressed. For example, by performing risk avoidance control based on a high-precision risk field, the excess work of the risk avoidance control can be suppressed. In addition, by determining whether the working conditions of the risk avoidance control are met based on the high-precision risk field, the unnecessary work of the risk avoidance control can be suppressed. As a result, the sense of disharmony with the excess work and unnecessary work of the risk avoidance control is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a conceptual diagram for explaining the outline of a driving assistance system according to an embodiment of the present invention.

[0051] Figure 2 It is a block diagram showing a configuration example of a vehicle and a driving assistance system according to an embodiment of the present invention.

[0052] Figure 3 It is a block diagram showing an example of driving environment information in the embodiment of the present invention.

[0053] Figure 4 This is a flowchart showing basic processing related to risk avoidance control of the driving assistance system according to the embodiment of the present invention.

[0054] Figure 5 This is a conceptual diagram for explaining the risk calculation process (step S300 ) according to the embodiment of the present invention.

[0055] Figure 6 This is a diagram showing an example of a risk field in the embodiment of the present invention.

[0056] Figure 7 This is a flowchart showing the basic risk calculation process (step S300 ) according to the embodiment of the present invention.

[0057] Figure 8 This is a flowchart showing the high-precision risk calculation process (step S300 ) according to the embodiment of the present invention.

[0058] Fig. 9 This is a conceptual diagram showing an example of target object state parameters of a plurality of patterns according to an embodiment of the present invention.

[0059] Fig.10 This is a conceptual diagram showing another example of target object state parameters of a plurality of patterns according to the embodiment of the present invention.

[0060] Fig.11It is a conceptual diagram showing a first example of the probability setting process (step S320) according to the embodiment of the present invention.

[0061] Fig.12 It is a conceptual diagram showing a second example of the probability setting process (step S320) according to the embodiment of the present invention.

[0062] Fig.13 It is a conceptual diagram showing a third example of the probability setting process (step S320) according to the embodiment of the present invention.

[0063] Fig.14 It is a conceptual diagram showing a fourth example of the probability setting process (step S320) according to the embodiment of the present invention.

[0064] Fig.15 It is a conceptual diagram showing a fourth example of the probability setting process (step S320) according to the embodiment of the present invention.

[0065] Fig.16 This is a block diagram showing information used in the driving assistance system according to the embodiment of the present invention.

[0066] Fig.17 This is a conceptual diagram for explaining superposition of a plurality of risk fields for a plurality of targets in the embodiment of the present invention.

[0067] Description of symbols

[0068] 1 Vehicle

[0069] 5 Target

[0070] 6 Obstacles

[0071] 10. Driving Assistance Systems

[0072] 20 Sensor Group

[0073] 21 Position sensor

[0074] 22 Vehicle status sensor

[0075] 23 Identification Sensor

[0076] 30 Travel device

[0077] 31 Steering device

[0078] 32 Drive unit

[0079] 33 Braking device

[0080] 100 Control device

[0081] 110 Processor

[0082] 120 Storage Device

[0083] 200 Driving environment information

[0084] 210 Vehicle location information

[0085] 220 Vehicle status information

[0086] 230 Surrounding situation information

[0087] 240 Road structure information

[0088] 250 Target information

[0089] 260 Obstacle Information

[0090] 270 Map Information

[0091] 300 Risk Information

[0092] 320 Probability Setting Information

[0093] AR Risk Determination Area

[0094] DLAT lateral distance

[0095] DLON Longitudinal distance

[0096] EDT envisions moving in the direction of

[0097] EVT Estimated Movement Speed

[0098] R Value at Risk

[0099] RF Risk Field

[0100] RP Partial Risk Value

[0101] ST Target state parameters

[0102] SV Vehicle status parameters DETAILED DESCRIPTION

[0103] Embodiments of the present invention will be described with reference to the drawings.

[0104] 1. Driving assistance systems

[0105] 1-1. Overview

[0106] Figure 1This is a conceptual diagram for explaining the outline of the driving assistance system 10 involved in the present embodiment. The driving assistance system 10 performs "driving assistance control" to assist the driving of the vehicle 1. Driving assistance control may also be included in automatic driving control. Typically, the driving assistance system 10 is mounted on the vehicle 1. Alternatively, at least a portion of the driving assistance system 10 may also be configured in an external device outside the vehicle 1 to perform driving assistance control in a remote manner. That is, the driving assistance system 10 may also be configured in a dispersed manner in the vehicle 1 and the external device.

[0107] The driving assistance control includes a "risk avoidance control" for avoiding risks in front of the vehicle 1 in advance. More specifically, the driving assistance system 10 identifies a target object 5 that exists in front of the vehicle 1. Typically, the target object 5 is a moving object. In addition, the driving assistance system 10 performs risk avoidance control in order to reduce (avoid) the risk of collision with the target object 5 in advance. Such risk avoidance control includes at least one of steering control and deceleration control.

[0108] For example, in Figure 1 In the figure, vehicle 1 is traveling in lane 1 L1 of roadway 2. There are pedestrians in the roadside area 3 (road shoulder, roadside strip, sidewalk, etc.) adjacent to roadway 2. The pedestrians may enter roadway 2 (lane 1 L1). Therefore, pedestrians in roadside area 3 are a risk to vehicle 1. In order to reduce the risk of collision with pedestrians, driving assistance system 10 performs risk avoidance control as needed. For example, driving assistance system 10 automatically steers the vehicle away from the pedestrians. Figure 1 In FIG. 1 , the trajectory TR0 represents the trajectory of the vehicle 1 when the risk avoidance control is not executed. On the other hand, the trajectory TR1 represents the trajectory of the vehicle 1 when the risk avoidance control is executed.

[0109] The target object 5 to be subject to risk avoidance control is not limited to pedestrians present in the roadside area 3. Other examples of the target object 5 include bicycles, two-wheeled vehicles, parked vehicles, etc. present in the roadside area 3. Still other examples of the target object 5 include pedestrians, bicycles, two-wheeled vehicles, and leading vehicles present in the roadway 2. That is, the target object 5 to be subject to risk avoidance control includes at least one of pedestrians, bicycles, two-wheeled vehicles, and other vehicles present in front of the vehicle 1.

[0110] 1-2. Example of configuration

[0111] Figure 2 1 is a block diagram schematically showing a configuration example of the vehicle 1 and the driving assistance system 10 according to the present embodiment. In particular, Figure 2 A configuration example related to risk avoidance control is shown. The vehicle 1 includes a sensor group 20 and a travel device 30 .

[0112] The sensor group 20 includes a position sensor 21, a vehicle state sensor 22, and an identification sensor 23. The position sensor 21 detects the position and orientation of the vehicle 1. As the position sensor 21, a GPS (Global Positioning System) sensor can be exemplified. The vehicle state sensor 22 detects the state of the vehicle 1. As the vehicle state sensor 22, a vehicle speed sensor, a yaw rate sensor, a lateral acceleration sensor, a steering angle sensor, etc. can be exemplified. The identification sensor 23 identifies (detects) the conditions around the vehicle 1. As the identification sensor 23, a camera, a radar, a laser radar (LIDAR: Laser Imaging Detection and Ranging), etc. can be exemplified.

[0113] The travel device 30 includes a steering device 31, a drive device 32, and a brake device 33. The steering device 31 steers the wheels of the vehicle 1. For example, the steering device 31 includes an electric power steering (EPS: Electric Power Steering) device. The drive device 32 is a power source that generates a driving force. Examples of the drive device 32 include an engine, an electric motor, an in-wheel motor, etc. The brake device 33 generates a braking force.

[0114] The driving assistance system 10 includes at least a control device 100. The driving assistance system 10 may also include a sensor group 20. The driving assistance system 10 may also include a travel device 30.

[0115] The control device 100 controls the vehicle 1. Typically, the control device 100 is a microcomputer mounted on the vehicle 1. The control device 100 is also called an ECU (Electronic Control Unit). Alternatively, the control device 100 may be an information processing device outside the vehicle 1. In this case, the control device 100 communicates with the vehicle 1 and controls the vehicle 1 remotely.

[0116] The control device 100 includes a processor 110 and a storage device 120. The processor 110 executes various processes. Various information is stored in the storage device 120. As the storage device 120, a volatile memory, a non-volatile memory, etc. can be exemplified. The processor 110 executes a control program as a computer program to implement various processes based on the processor 110 (control device 100). The control program is stored in the storage device 120 or recorded in a computer-readable recording medium.

[0117] 1-3. Information acquisition and processing

[0118] The processor 110 (control device 100) executes an “information acquisition process” to acquire driving environment information 200 indicating the driving environment of the vehicle 1. The driving environment information 200 is acquired based on the detection results of the sensor group 20 mounted on the vehicle 1. The acquired driving environment information 200 is stored in the storage device 120.

[0119] Figure 3 2 is a block diagram showing an example of driving environment information 200. Driving environment information 200 includes vehicle position information 210, vehicle state information 220, surrounding situation information 230, map information 270, and the like.

[0120] The vehicle position information 210 is information indicating the position and orientation of the vehicle 1 . The processor 110 acquires the vehicle position information 210 from the detection result of the position sensor 21 .

[0121] The vehicle state information 220 is information indicating the state of the vehicle 1 . Examples of the state of the vehicle 1 include vehicle speed, yaw rate, lateral acceleration, steering angle, etc. The processor 110 acquires the vehicle state information 220 from the detection result of the vehicle state sensor 22 .

[0122] The surrounding condition information 230 is information indicating the condition around the vehicle 1. The surrounding condition information 230 includes information obtained by the recognition sensor 23. For example, the surrounding condition information 230 includes image information indicating the condition around the vehicle 1 captured by a camera. As another example, the surrounding condition information 230 includes measurement information measured by a radar or a laser radar. The surrounding condition information 230 also includes road configuration information 240, target object information 250, and obstacle information 260.

[0123] The road configuration information 240 is information related to the road configuration around the vehicle 1. The road configuration around the vehicle 1 includes demarcation lines (white lines) and road end objects. The road end objects are three-dimensional objects that indicate the end of the road. Examples of road end objects include curbstones, guardrails, walls, and central medians. The road configuration information 240 indicates at least the positions of the demarcation lines and road end objects (relative positions relative to the vehicle 1).

[0124] For example, by analyzing the image information obtained by the camera, the demarcation line can be identified and the relative position of the demarcation line can be calculated. As an image analysis method, semantic segmentation and edge detection can be exemplified. Similarly, by analyzing the image information, the road side object can be identified and the relative position of the road side object can be calculated. Alternatively, the relative position of the road side object can also be obtained from the radar measurement information.

[0125] The target object information 250 is information about the target object 5 around the vehicle 1. Examples of the target object 5 include pedestrians, bicycles, two-wheeled vehicles, other vehicles (preceding vehicles, parked vehicles), etc. The target object information 250 indicates the relative position and relative speed of the target object 5 relative to the vehicle 1. For example, by analyzing the image information obtained by the camera, the target object 5 can be identified and the relative position of the target object 5 can be calculated. In addition, the target object 5 can also be identified based on radar measurement information to obtain the relative position and relative speed of the target object 5. The target object information 250 may also include the moving direction and moving speed of the target object 5. The moving direction and moving speed of the target object 5 can be calculated by tracking the position of the target object 5. The target object information 250 may also include a historical record of the relative position, relative speed, moving direction and moving speed of the target object 5 in a certain period of time in the past. The target object information 250 may also indicate the type of the target object 5 (pedestrian, bicycle, two-wheeled vehicle, other vehicle, etc.).

[0126] The obstacle information 260 is information about obstacles around the vehicle 1. Examples of obstacles include utility poles, signs, walls, buildings, etc. The obstacle information 260 indicates at least the position of the obstacle (relative position with respect to the vehicle 1).

[0127] The map information 270 indicates lane configuration, road shape, etc. The control device 100 obtains the map information 270 of the required area from the map database. The map database may be stored in a predetermined storage device mounted on the vehicle 1 or in a management server outside the vehicle 1. In the latter case, the processor 110 communicates with the management server to obtain the required map information 270.

[0128] 1-4. Vehicle driving control

[0129] The processor 110 (control device 100) performs "vehicle driving control" for controlling the driving of the vehicle 1. The vehicle driving control includes steering control for controlling the steering of the vehicle 1, acceleration control for controlling the acceleration of the vehicle 1, and deceleration control for controlling the deceleration of the vehicle 1. The processor 110 performs vehicle driving control by controlling the driving device 30. Specifically, the processor 110 performs steering control by controlling the steering device 31. In addition, the processor 110 performs acceleration control by controlling the driving device 32. In addition, the control device 100 performs deceleration control by controlling the braking device 33.

[0130] 1-5. Emergency brake control

[0131] The processor 110 (control device 100) performs "emergency braking control" to automatically decelerate the vehicle 1 in an emergency. Emergency braking control is also called AEB (Autonomous Emergency Braking) or PCS (Pre-Crash Safety System). Specifically, the processor 110 calculates the collision margin time (TTC: Time To Collision) of the vehicle 1 relative to the target object 5 in front based on the driving environment information 200. The vehicle speed of the vehicle 1 is obtained from the vehicle state information 220. The distance to the target object 5 is obtained from the target object information 250. When the calculated collision margin time is lower than the first threshold, the processor 110 performs deceleration control at the first deceleration to decelerate and stop the vehicle 1. The first deceleration is typically a predetermined value.

[0132] 1-6. Risk avoidance control

[0133] The processor 110 (control device 100) performs driving assistance control to assist the driving of the vehicle 1. The driving assistance control includes risk avoidance control. Risk avoidance control is a vehicle driving control used to reduce (avoid) the risk of collision between the vehicle 1 and the target object 5 in front in advance, including at least one of steering control and deceleration control. The processor 110 performs risk avoidance control based on the above-mentioned driving environment information 200.

[0134] The risk avoidance control according to this embodiment will be described in more detail below.

[0135] 2. Basic treatment of risk avoidance control

[0136] Figure 4 This is a flowchart showing basic processing related to risk avoidance control according to the present embodiment. Figure 4 The processing flow shown is repeatedly executed at every predetermined cycle.

[0137] 2-1. Step S100

[0138] In step S100 , the processor 110 executes the above-mentioned information acquisition process. That is, the processor 110 acquires the driving environment information 200 based on the detection result of the sensor group 20 . The driving environment information 200 is stored in the storage device 120 .

[0139] 2-2. Step S200

[0140] In step S200, the processor 110 determines whether there is a target object 5 in front of the vehicle 1 based on the target object information 250. In other words, the processor 110 determines whether the target object 5 is recognized in the area in front of the vehicle 1. Typically, the target object 5 is a moving object. The target object 5 includes at least one of a pedestrian, a bicycle, a two-wheeled vehicle, and other vehicles in front of the vehicle 1. In the case where the target object 5 in front of the vehicle 1 is recognized (step S200; yes), the process proceeds to step S300. In other cases (step S200; no), the process returns to step S100.

[0141] 2-3. Step S300 (Risk Calculation Process)

[0142] In step S300, the processor 110 estimates the risk of collision with the target object 5. According to the present embodiment, the risk of collision with the target object 5 is represented by a "risk field RF" described below. That is, in step S300, the processor 110 sets (calculates) the risk field RF related to the target object 5. This process is hereinafter referred to as "risk calculation process".

[0143] Figure 5 This is a conceptual diagram for explaining the risk calculation process involved in this embodiment. The vehicle coordinate system (X, Y) is a relative coordinate system fixed to the vehicle 1, and changes with the movement of the vehicle 1. The X direction is the front direction (travel direction) of the vehicle 1. The Y direction is the lateral direction of the vehicle 1. The X direction and the Y direction are orthogonal to each other. In the following description, the longitudinal distance means the distance in the X direction, and the lateral distance means the distance in the Y direction.

[0144] First, the processor 110 sets a risk determination area AR in front of the vehicle 1. The risk determination area AR is set to cover at least the area between the vehicle 1 and the target object 5. The lateral width of the risk determination area AR is set to be consistent with the lateral width of the roadway 2, for example. The relative position of the target object 5 relative to the vehicle 1 is obtained from the target object information 250. The position of the roadway 2 is obtained from the road configuration information 240 or the map information 270.

[0145] The processor 110 sets a plurality of virtual vehicle positions PV in the risk determination area AR. i The subscript i takes an integer value of 1 to 2 or more. Each vehicle position PVi is converted into a virtual relative position between the vehicle 1 and the target object 5 by using the position of the target object 5 represented by the target object information 250. The virtual assumed position is determined by the longitudinal distance DLON between the vehicle 1 and the target object 5. i and lateral distance DLAT i Further, the processor 110 calculates the vehicle position PV i Set multiple imaginary speeds V jThe subscript j takes an integer value greater than 1 or 2. The hypothetical relative position [DLAT i ,DLON i ] and the imaginary speed V j The combination of ij ”.

[0146] In addition, the processor 110 sets a “target object state parameter ST” indicating the movement state of the target object 5 . The target object state parameter ST includes an assumed movement direction EDT and an assumed movement speed EVT of the target object 5 .

[0147] Furthermore, the processor 110 estimates the vehicle state parameter SV in a quantitative manner. ij The collision risk between the vehicle 1 defined by the target state parameter ST and the target object 5 defined by the target state parameter ST. According to this embodiment, the collision risk between the vehicle 1 and the target object 5 is determined by the “risk value R ij "This quantitative value is expressed.

[0148] More specifically, the processor 110 determines the vehicle state parameter SV ij The vehicle 1 defined by the target state parameter ST will collide with the target 5 defined by the target state parameter ST, and the estimated collision speed U between the vehicle 1 and the target 5 is calculated. ij The estimated collision speed U when it is determined that no collision occurs ij =0. In determining whether a collision has occurred, the operation of the emergency brake control described above may also be considered. The risk value R associated with the target object 5 ij The estimated collision speed U ij For example, the risk value R ij is the estimated collision speed U ij As another example, the risk value R ij It can also be the estimated collision speed U ij The square of . In any case, the risk value R ij Based on the estimated collision speed U ij As the collision speed U ij Higher, risk value R ij Increase.

[0149] Thus, the risk value R associated with the target object 5 is ij The vehicle state parameter SV ij The processor 110 sets the vehicle state parameter SV of a plurality of modes in the risk determination area AR. ij Furthermore, the processor 110 determines the vehicle state parameters SV of the multiple modes. ij Calculate the risk value R ijAs a result, the vehicle state parameter SV for multiple modes is obtained. ij The risk value R ij The distribution of vehicle state parameters SV for multiple modes ij The risk value R ij The distribution of is the risk field RF.

[0150] Figure 6 An example of a risk field RF is shown. The risk field RF represents the risk value R in a three-dimensional space defined by the longitudinal distance DLON, the lateral distance DLAT, and the speed V. ij The position in three-dimensional space is equivalent to the vehicle state parameter SV ij The risk value R is calculated for a large number of positions in three-dimensional space. ij .in particular, Figure 6 The black dots in the figure represent the situation where the vehicle 1 collides with the target object 5. Each black dot has a risk value R greater than 0. ij .

[0151] Figure 7 This is a flowchart showing the basic risk calculation process (step S300 ) described above.

[0152] In step S331, the processor 110 sets the target object state parameter ST. In addition, by setting the target object state parameter ST in a variable manner according to the situation, the accuracy of the risk calculation process can be further improved. The high-precision risk calculation process will be described in detail later.

[0153] In step S332, the processor 110 sets the vehicle state parameter SV ij Specifically, the processor 110 selects the vehicle state parameters SV of the multiple modes one by one. ij .

[0154] In step S333, the processor 110 determines the vehicle state parameter SV ij Whether the vehicle 1 defined by the target state parameter ST will collide with the target 5 defined by the target state parameter ST. In step S333, it is assumed that the vehicle 1 is moving at a speed V j The processor 110 performs the constant speed motion based on the vehicle state parameter SV ij and the target object state parameter ST, it is determined whether the collision between the vehicle 1 and the target object 5 occurs. If the collision between the vehicle 1 and the target object 5 occurs (step 333; Yes), the process proceeds to step S334. On the other hand, if the collision between the vehicle 1 and the target object 5 does not occur (step 333; No), the process proceeds to step S336.

[0155] In step S334, the processor 110 assumes that the above-mentioned emergency brake control is in operation and determines whether the vehicle 1 will collide with the target object 5. The control delay of the emergency brake control (the delay time from the detection of the target object to the operation of the brake) and the first deceleration based on the emergency brake control are predetermined parameters that depend on the performance of the emergency brake control. Brake performance information (not shown) representing the predetermined parameters is pre-stored in the storage device 120. The processor 110 determines whether the vehicle 1 will collide with the target object 5 based on the vehicle state parameter SV ij , the target object state parameter ST and the brake performance information, and determine whether the collision between the vehicle 1 and the target object 5 will occur. In the case where the collision between the vehicle 1 and the target object 5 occurs (step 334; yes), the process proceeds to step S335. On the other hand, in the case where the collision between the vehicle 1 and the target object 5 does not occur (step 334; no), the process proceeds to step S336.

[0156] In step S335, the processor 110 calculates the estimated collision speed U between the vehicle 1 and the target object 5 based on the result of step S334. ij Furthermore, the processor 110 estimates the collision speed U ij , calculate the risk value R ij With the estimated collision speed U ij Higher, risk value R ij Increase. For example, the risk value R ij is the estimated collision speed U ij As another example, the risk value R ij It can also be the estimated collision speed U ij Then, the process proceeds to step S337.

[0157] In step S336, since the collision speed U ij is zero, so the processor 110 sets the risk value R ij Set to zero. Then, the process proceeds to step S337.

[0158] In step S337, the processor 110 determines the vehicle state parameter SV for all modes. ij The risk value R ij The calculation of the vehicle state parameter SV for any mode has not yet been calculated. ij The risk value R ij In the case of (step S337; No), the process returns to step S332. Steps S332 to S337 are repeated until the vehicle state parameter SV for all modes is ij The risk value R ij The calculation of is completed. As a result, the risk field RF is obtained (refer to Figure 6 ), the risk field RF represents the vehicle state parameter SV for all modesij The risk value R ij distribution.

[0159] In this way, the processor 110 executes the risk calculation process to set (calculate) the risk field RF. The risk information 300 indicating the calculated risk field RF is stored in the storage device 120 (see Figure 2 ). Then, the process proceeds to the next step S400.

[0160] 2-4. Step S400

[0161] In step S400 , the processor 110 determines whether a working condition of risk avoidance control is met.

[0162] For example, the working conditions of the risk aversion control include "the trajectory of the vehicle 1 without executing the risk aversion control has a risk value R ij is the point above the operating threshold Rth".

[0163] Figure 1 The trajectory TR0 in represents the trajectory of vehicle 1 when risk avoidance control is not performed. For example, vehicle 1 is assumed to be traveling parallel to lane 2 (lane 1 L1) at the current speed. The current speed of vehicle 1 is obtained from vehicle state information 220. The position of lane 2 (lane 1 L1) is obtained from road configuration information 240 or map information 270. Therefore, processor 110 is able to estimate trajectory TR0 based on driving environment information 200. Furthermore, processor 110 determines whether the above-mentioned working conditions are met based on trajectory TR0 and risk field RF represented by risk information 300.

[0164] The working condition of the risk avoidance control may also include "the margin time T to the target object 5 is lower than the second threshold value Tth". The margin time T refers to the time required for the vehicle 1 to travel from the current position to the position closest to the target object 5. Typically, the timing when the vehicle 1 is closest to the target object 5 is the timing when the vehicle 1 passes the side of the target object 5. The current speed of the vehicle 1 is obtained from the vehicle state information 220. The relative position of the target object 5 relative to the vehicle 1 is obtained from the target object information 250. The processor 110 can calculate the margin time T based on the trajectory TR0, the current speed of the vehicle 1, and the relative position of the target object 5. In addition, the processor 110 determines whether the working condition is met by comparing the margin time T with the second threshold value Tth. In addition, the second threshold value Tth is larger than the first threshold value related to the above-mentioned emergency braking control.

[0165] The operating condition of the risk avoidance control may also include that the speed of the vehicle 1 is above a certain speed.

[0166] If the operating condition of the risk avoidance control is satisfied (step S400; Yes), the process proceeds to step S500. On the other hand, if the operating condition of the risk avoidance control is not satisfied (step S400; No), the process proceeds to step S600.

[0167] 2-5. Step S500

[0168] In step S500, the processor 110 performs risk avoidance control, that is, makes the risk avoidance control work. Risk avoidance control is a vehicle driving control used to reduce the risk of collision with the target object 5, including at least one of steering control and deceleration control. In this embodiment, the processor 110 performs risk avoidance control to reduce the risk of collision with the target object 5 based on the risk field RF represented by the risk information 300.

[0169] For example, the processor 110 generates a target trajectory TRt of the vehicle 1 based on the risk field RF. The target trajectory TRt includes a target position and a target speed of the vehicle 1 in the roadway 2. The processor 110 sets the target trajectory TRt with reference to the risk field RF so that the risk value R on the target trajectory TRt is ij For example, the processor 110 sets the target trajectory TRt so that the vehicle 1 moves away from the target object 5 (see Figure 1 The current position of the vehicle 1 is obtained from the vehicle position information 210. The vehicle speed is obtained from the vehicle state information 220. The position of the roadway 2 is obtained from the road configuration information 240 or the map information 270. Therefore, the processor 110 can generate a target trajectory TRt based on the driving environment information 200 and the risk information 300 (risk field RF).

[0170] The processor 110 may also generate the target trajectory TRt by taking into account the smoothness of the vehicle behavior together with the risk field RF. In this case, the processor 110 sets the target trajectory TRt that takes into account both safety and the smoothness of the vehicle behavior by using an evaluation function that includes a risk component and a vehicle behavior component. For details of the target trajectory TRt that also takes into account the smoothness of the vehicle behavior, please refer to the above-mentioned patent document 1 (Japanese Patent Publication No. 2017-206117).

[0171] The processor 110 performs at least one of steering control and deceleration control so that the vehicle 1 follows the target trajectory TRt. The steering control and deceleration control are performed based on the driving environment information 200. Specifically, the processor 110 calculates the target steering angle and target deceleration required to make the vehicle 1 follow the target trajectory TRt. The processor 110 controls the steering device 31 according to the target steering angle. In addition, the processor 110 controls the braking device 33 according to the target deceleration.

[0172] 2-6. Step S600

[0173] In step S600, the processor 110 does not execute the risk avoidance control. That is, the processor 110 does not cause the risk avoidance control to work. If the risk avoidance control is already being executed, the processor 110 stops the risk avoidance control.

[0174] 3. High-precision risk calculation process (step S300)

[0175] In order to properly perform risk avoidance control, it is necessary to properly estimate the magnitude of the risk of collision with the target object 5 (ie, the risk value R ij This is because, if the collision risk is overestimated, the risk avoidance control will operate excessively or unnecessarily.

[0176] According to the technology disclosed in the above-mentioned patent document 1 (Japanese Patent Publication No. 2017-206117), it is assumed that a hypothetical moving body existing in the blind spot of the obstacle moves perpendicularly to the roadway 2 at a predetermined speed and enters the roadway 2. This means that the target object state parameters ST (assumed moving direction EDT, assumed moving speed EVT) are the same regardless of the situation. However, the actual target object 5 is not necessarily limited to moving perpendicularly to the roadway 2 and entering the roadway 2. The assumption in patent document 1 can be said to be too pessimistic for the actual target object 5. When such an assumption is also applied to the actual target object 5, the collision risk will be overestimated, and the possibility of risk avoidance control working excessively or unnecessarily becomes higher. The occupants of the vehicle 1 (typically the driver) may feel a sense of disobedience for the excessive and unnecessary work of the risk avoidance control.

[0177] Therefore, the present embodiment provides a technique for estimating the collision risk with higher accuracy in order to suppress the redundant or unnecessary work of the risk avoidance control. Figure 8 This is a flowchart showing the high-precision risk calculation process (step S300 ) according to the present embodiment.

[0178] 3-1. Step S310

[0179] In step S310, the processor 110 sets the target state parameter ST. In this embodiment, in order to consider various movement states of the target 5, the target state parameter ST (the assumed movement direction EDT, the assumed movement speed EVT) is not fixed to one. Instead, the processor 110 sets a plurality of modes of the target state parameter ST. k The subscript k takes a value of an integer greater than or equal to 1 or 2.

[0180] Fig. 9is a target state parameter ST representing multiple modes k A conceptual diagram of an example of . Fig. 9 In the example shown, eight target state parameters ST are set. k (k=1 to 8). More specifically, eight patterns of assumed moving directions EDT are set. k (k=1 to 8). The assumed moving speed EVT is set to the current moving speed of the object 5 indicated by the object information 250. The 8 patterns of assumed moving directions EDT k The target object 5 is distributed in the circumferential direction. It is preferred that there are 8 patterns of assumed movement directions EDT. k In other words, the assumed movement direction EDT of 8 modes is preferably k The pixels are distributed over the entire circumference at intervals of 45 degrees. Thus, various movement directions of the target object 5 can be fully considered with limited computing resources.

[0181] Fig.10 is a target state parameter ST representing multiple modes k Conceptual diagram of other examples. Fig.10 In the example shown, the target state parameter ST k EDT is the assumed moving direction of multiple modes m and multiple modes of envisioned moving speed EVT n The subscript m takes a value that is an integer greater than 1 to 2. The subscript n takes a value that is an integer greater than 1 to 2. For example, Fig.10 As shown, the EDT is assumed to move in 8 modes m (m = 1 to 8) and the assumed moving speed EVT of the three modes n (n = 1 to 3) to obtain 24 modes of target state parameters ST k (k=1 to 24) 8 modes of assumed moving directions EDT m With the above Fig. 9 The same is true for the case of . The three modes of assumed moving speeds EVT1 to EVT3 are set to 0.5 times, 1.0 times, and 1.5 times the current moving speed of the target object 5, respectively. Thus, various moving states of the target object 5 can be fully considered with limited computing resources.

[0182] Each target state parameter ST k The target vector represents the movement of the target 5. The position of the front end of the target vector is the candidate of the movement destination of the target 5. The target state parameter ST is set kThis is equivalent to setting an object vector indicating the movement of the object 5 and setting a candidate destination of the object 5. Setting multiple patterns of object state parameters ST k This is equivalent to setting a plurality of patterns of target object vectors and setting a plurality of patterns of destination candidates.

[0183] 3-2. Step S320 (Probability Setting Process)

[0184] In step S320, the processor 110 executes to set each target state parameter ST k The probability P k "Probability setting process". Set each target state parameter ST k The probability P k The probability P of the target object 5 moving to each candidate destination is set k In addition, the target state parameters ST of multiple modes k The probability of each k The total is 1.0.

[0185] The target object 5 does not move randomly, but moves with a certain degree of continuity with past movements. At least the recent moving direction and moving speed of the target object 5 serve as clues for estimating the future moving direction and moving speed. Therefore, the processor 110 sets each target state parameter ST based on at least the target object information 250. k The probability P k Various examples of the probability setting process are described below.

[0186] 3-2-1. Case 1

[0187] Fig.11 : is a conceptual diagram showing a first example of the probability setting process. Time t represents the current time. Time t-3 to t-1 represent past time. Time t+1 represents future time. Target state parameter ST k This corresponds to the movement state (the assumed movement direction EDT, the assumed movement speed EVT) of the target object 5 between the current time t and the future time t+1.

[0188] The target object 5 is, for example, a pedestrian. The target object 5 has been moving parallel to the direction of travel (X direction) of the vehicle 1 for a certain period of time in the past. Such a target object 5 has a high probability of maintaining the same moving direction. Therefore, the probability P1 of the target state parameter ST1 closest to the moving direction of the target object 5 is set to the highest.

[0189] The first example can be generalized as follows. The target state parameter ST kThe closer the assumed moving direction EDT is to the moving direction of the target 5 indicated by the target information 250, the greater the target state parameter ST k The probability P k On the contrary, as the target state parameter ST k The assumed moving direction EDT is away from the moving direction of the target 5 represented by the target information 250. k The probability P k By taking into account the moving direction of the target object 5 represented by the target object information 250, the target object state parameter ST can be set with high accuracy. k The probability P k .

[0190] 3-2-2. Case 2

[0191] Fig.12 : is a conceptual diagram showing the second example of the probability setting process. The descriptions repeated in the first example are appropriately omitted. The target object 5 has been moving at a constant speed for a certain period of time in the past. The probability that the target object 5 maintains the same moving speed is high. Therefore, the target state parameter ST that is closest to the moving speed of the target object 5 is 12 The probability P 12 is set to highest.

[0192] The second example can be generalized as follows. The target state parameter ST k The closer the assumed moving speed EVT is to the moving speed of the target object 5 indicated by the target object information 250, the greater the target object state parameter ST k The probability P k Conversely, as the assumed moving speed EVT of the target object state parameter STk moves away from the moving speed of the target object 5 represented by the target object information 250, the target object state parameter ST k The probability P k By taking into account the moving speed of the target object 5 represented by the target object information 250, the target object state parameter ST can be set with good accuracy. k The probability P k .

[0193] 3-2-3.Example 3

[0194] Fig.13 2 is a conceptual diagram showing the third example of the probability setting process. The descriptions repeated in the first example are omitted as appropriate. Fig.13 In the example shown, the target object 5 is a bicycle. Bicycles are more likely to move straight than pedestrians. It is very unlikely that a bicycle will suddenly turn at a right angle. Therefore, the target object state parameter ST that approaches the moving direction of the target object 5 represented by the target object information 250 isk The probability P k Compared with the case where the target object 5 is a pedestrian (refer to Fig.11 ) and is set higher.

[0195] The third example can be generalized as follows. The object information 250 indicates the type of the object 5 (pedestrian, bicycle, two-wheeled vehicle, other vehicle, etc.). The processor 110 sets each object state parameter ST according to the type of the object 5. k The probability P k By also taking the type of the target object 5 into consideration in this way, the accuracy of the probability setting process is further improved.

[0196] 3-2-4.Case 4

[0197] Fig.14 : is a conceptual diagram showing the fourth example of the probability setting process. The descriptions repeated in the first example are appropriately omitted. When there is an obstacle 6 such as a telephone pole or a signboard in the moving direction, the target object 5 is likely to move forward while avoiding the obstacle 6. Therefore, the probability P2 of the target object state parameter ST2 such as avoiding the obstacle 6 and moving forward is set high. On the other hand, the probability P1 of the target object state parameter ST1 such as colliding with the obstacle 6 is set to zero.

[0198] exist Fig.15 In the example shown, the obstacle 6 is a wall-like structure. The target object 5 does not move toward the obstacle 6. Therefore, the probabilities P6 to P8 of the target object state parameters ST6 to ST8 that the target object 5 moves toward the obstacle 6 are set to zero.

[0199] The fourth example can be generalized as follows. Obstacle information 260 indicates the position of obstacle 6 through which target object 5 cannot pass. Based on obstacle information 260, processor 110 sets target object state parameter ST k The probability P k In addition, the processor 110 sets other target state parameters ST based on the target information 250. k The probability P k By taking the obstacle 6 into consideration in this way, the accuracy of the probability setting process is further improved.

[0200] 3-2-5.Example 5

[0201] As long as there is no contradiction, a plurality of the first to fourth examples described above may be combined.

[0202] 3-2-6. Probability setting information

[0203] As described above, the processor 110 dynamically sets the target state parameter ST according to the “situation S”. k The probability P k The condition S at least includes the moving direction and moving speed of the target object 5 obtained from the target object information 250. The condition S may also include the past movement history of the target object 5. The condition S may also include the condition around the target object 5 (for example, whether there is an obstacle 6).

[0204] Target state parameter ST k The probability P k It is represented by the function of the condition S (P k =f(S)). For example, the function f is generated in advance by machine learning (including deep learning). Specifically, the function f is generated in advance by learning the movement of the target object 5 in many actual environments. As another example, the function f can also be implemented by a Kalman filter. As another example, the function f can also be set manually.

[0205] Fig.16 The probability setting information 320 shown in FIG. k The probability setting information 320 is generated in advance and stored in the storage device 120. The situation S is obtained from the driving environment information 200 (target information 250, obstacle information 260, etc.). The processor 110 sets the probability P corresponding to the situation S based on the driving environment information 200 and the probability setting information 320. k .

[0206] 3-3. Step S330

[0207] In step S330, the processor 110 uses each target state parameter ST k , calculate the risk value R ij . Use each target state parameter ST k The calculated risk value R ij In the following it is referred to as “partial value at risk RP ijk "Partial Risk Value RP ijk The calculation method of the risk value R is the same as that described in Section 2-3 above. ij The calculation method is the same (refer to Figure 7 ). In step S331, each target state parameter ST k . Partial Value at Risk RP ijk It can also be said that the vehicle state parameter SV ij and the target state parameter ST k function.

[0208] 3-4. Step S340

[0209] Through the steps S310 to S330 described above, according to each target state parameter ST k Get the probability P k and the fractional value at risk RP ijk In step S340, the processor 110 calculates the target state parameters ST for all modes. k Partial Value at Risk RP ijk Integrate and calculate the final risk value R ij More specifically, as expressed by the following equation (1), for the target state parameter ST of all modes: k The probability P k Partial Value at Risk (RP) ijk The sum of the products is taken as the final risk value R ij To figure out.

[0210]

[0211] The risk field RF represents the risk value R provided by formula (1) ij The risk information 300 indicating the calculated risk field RF is stored in the storage device 120 (refer to Fig.16 The risk field RF indicated by the risk information 300 is applied to the process associated with the risk avoidance control (steps S400 and S500).

[0212] 4. Effects and Discussions

[0213] As described above, according to the present embodiment, the risk of collision with the target object 5 existing in front of the vehicle 1 is represented by the risk field RF. More specifically, the risk field RF represents the vehicle state parameter SV for a plurality of modes. ij The risk value R ij The distribution of vehicle state parameter SV ij The virtual relative position between the vehicle 1 and the target 5 [DLAT i ,DLON i ] and the imaginary speed V of vehicle 1 j The target object state parameter ST includes the assumed moving direction EDT and the assumed moving speed EVT of the target object 5. Risk value R ij By the vehicle state parameter SV ij The estimated collision speed U between the vehicle 1 defined by the target object 5 and the target object state parameter ST is ij The function representation of .

[0214] Thus, in the risk field RF (risk value R ij ) setting, the vehicle state parameter SV ijThe risk value R is evaluated comprehensively with the target state parameter ST. ij is based on the estimated collision velocity U ij In other words, the risk value R ij It is not a perceptual one, but a physical one. Therefore, the risk field RF (risk value Rij) can be said to represent the risk of collision with the target object 5 with high accuracy.

[0215] Furthermore, according to the present embodiment, in setting the risk field RF, the target state parameters ST of multiple modes are considered. k Specifically, multiple target state parameters ST are set. k The probability of each k The target object 5 does not move randomly, but moves with a certain degree of continuity with past movements. At least the recent moving direction and moving speed of the target object 5 serve as clues for estimating the future moving direction and moving speed. Therefore, each target state parameter ST can be set based on the target object information 250 indicating the moving direction and moving speed of the target object 5. k The probability P k In addition, the state parameters ST of each target object are used k The risk value R ij Used as part of the risk value RP ijk And, for all modes of target state parameters ST k The probability P k and the fractional value at risk RP ijk The sum of the products is taken as the final risk value R ij to use.

[0216] Thus, in setting the risk field RF, the target object state parameter ST is not fixed to one type, but multiple modes of target object state parameters ST are considered. k That is, the target state parameters ST of multiple modes k This is reflected in the risk field RF. Therefore, the accuracy of the risk field RF indicating the risk of collision with the target object 5 is further improved.

[0217] The high-precision risk field RF thus obtained is applied to risk avoidance control for reducing the risk of collision with the target object 5. Since the risk field RF represents the risk of collision with the target object 5 with high precision, the excess work or unnecessary work of the risk avoidance control is suppressed. For example, by performing risk avoidance control based on the high-precision risk field RF, the excess work of the risk avoidance control can be suppressed. In addition, by determining whether the working conditions of the risk avoidance control are met based on the high-precision risk field RF, the unnecessary work of the risk avoidance control can be suppressed. As a result, the sense of disharmony with the excess work and unnecessary work of the risk avoidance control is reduced.

[0218] As described above, in calculating the risk field RF, the vehicle state parameter SV ij The target object state parameter ST is comprehensively evaluated. Therefore, it is easy to combine the target object state parameter ST of multiple modes. k In addition, the risk value R ij It is a quantitative value, so it is easy to calculate the risk value R ij The method involved in this embodiment can be said to have a high affinity with the risk field RF.

[0219] In addition, the risk value R ij is a quantitative value, so it is also possible to superimpose multiple risk fields RF for multiple targets 5. As an example, Fig.17 As shown, consider the situation where there are two different target objects 5-1 and 5-2 in front of the vehicle 1. The processor 110 calculates the risk fields RF-1 and RF-2 for the target objects 5-1 and 5-2, respectively. And the processor 110 obtains a final risk field RF by superimposing the risk fields RF-1 and RF-2. For example, the risk values ​​Rij of the risk fields RF-1 and RF-2 are added. Alternatively, the maximum risk value Rij in the risk fields RF-1 and RF-2 is selected. By superimposing the risk fields RF for multiple target objects 5, the collision risk can be more accurately grasped, and risk avoidance control can be performed more appropriately.

[0220] According to the present embodiment, the target trajectory TRt of the vehicle 1 is generated based on the risk field RF. Since the target object state parameters ST of multiple modes are already reflected in the risk field RF, k Therefore, when generating the target trajectory TRt, it is not necessary to consider the target state parameters ST of multiple modes individually. k This helps to reduce the computational load required for calculating the target trajectory TRt.

[0221] As a comparative example, consider the case where the target trajectory TRt is generated without using the risk field RF.k Consider each of them separately and generate multiple types of target trajectories TRt k Furthermore, using the evaluation function, multiple types of target trajectories TRt are calculated. k Then, based on the evaluation values, multiple types of target trajectories TRt are selected. k In the comparative example, however, the calculation load increases. In particular, in the case of Fig.17 When there are a plurality of targets 5 as shown in FIG. 1 , the target state parameter ST associated with each target 5 is k The number of combinations becomes enormous. In addition, the weighting criteria of the plurality of targets 5 also become complicated. When there are a plurality of targets 5, it is not realistic to generate the target trajectory TRt by the method according to the comparative example.

[0222] On the other hand, according to the present embodiment, the target state parameters ST k is reflected in one risk field RF. Even when there are multiple targets 5, multiple risk fields RF for multiple targets 5 are integrated into one risk field RF. And, based on one risk field RF, the target trajectory TRt is calculated only once. Therefore, the calculation load is reduced. The more the number of targets 5 increases, the more significant the effect of reducing the calculation load is.

Claims

1. A driving assistance system that assists the driving of a vehicle and has: a storage device storing driving environment information, the driving environment information indicating the driving environment of the vehicle, including target object information indicating the position, moving direction, and moving speed of a target object existing in front of the vehicle, and obstacle information indicating the position of an obstacle through which the target object cannot pass, the target object information and the obstacle information being acquired by a recognition sensor provided in the vehicle for recognizing a condition around the vehicle; and a processor that performs risk avoidance control based on the driving environment information, the risk avoidance control including at least one of steering control and deceleration control for reducing the risk of collision with the target object, The vehicle state parameters include the hypothetical relative position between the vehicle and the target object and the hypothetical speed of the vehicle. The target object state parameters include the assumed moving direction and assumed moving speed of the target object. a risk value associated with the object represented by a function of an estimated collision speed of the vehicle defined by the vehicle state parameters and the object defined by the object state parameters, The risk field represents the distribution of the risk values ​​of the vehicle state parameters for a plurality of modes, the processor, Setting multiple modes of target state parameters, Based on the target object information, the probability of each of the target object state parameters of the multiple modes is set, calculating the risk value when using each of the target object state parameters of the plurality of patterns as a partial risk value, setting the risk field using the risk value which is the sum of the products of the probabilities of the target object state parameters for the plurality of patterns and the partial risk values, Based on the risk field, determining whether to execute the risk avoidance control or whether the working conditions of the risk avoidance control are met, The processor sets the target state parameters of the plurality of modes so that the assumed movement directions of the plurality of modes are distributed in a circumferential direction with the target as the center, The processor sets the probability of the target object state parameter to be higher as the assumed movement direction of the target object state parameter is closer to the movement direction of the target object represented by the target object information, The processor sets the probability of the target object state parameter that the obstacle exists in the assumed moving direction to zero based on the obstacle information, When there are a plurality of the target objects, the processor calculates a plurality of risk fields for each of the plurality of target objects, and acquires the risk field by superimposing the plurality of risk fields.

2. The driving assistance system according to claim 1, The processor sets the target state parameters in a plurality of modes so that the assumed moving speed includes a plurality of modes.

3. The driving assistance system according to claim 2, The processor sets the probability of the target object state parameter to be higher as the assumed moving speed of the target object state parameter is closer to the moving speed of the target object represented by the target object information.

4. The driving assistance system according to claim 1 or 2, The target object information also indicates the type of the target object. The processor further calculates the probability based on the type of the target object.

5. The driving assistance system according to claim 1 or 2, The processor further executes emergency braking control to decelerate the vehicle at a first deceleration when a collision margin time with respect to the target object is less than a first threshold value, The processor calculates the estimated collision speed in consideration of the operation of the emergency brake control, and calculates the risk value based on the estimated collision speed.